ra
Center for Earth and Planetary Studies, National Air and Space Museum, MRC 315, Washington, DC 20560, USA
a r t i c l e i n f o
Article history:
Received 2 December 2011
Revised 26 May 2012
Accepted 12 June 2012
Available online 25 June 2012
Keywords:
Mars
Mars, Surface
Radar observations
planets and the Moon, can be modeled as the sum of two separate
components: a ??quasispecular?? echo from large (super-wave-
length-scale) planar surfaces and a ??diffuse?? echo associated with
high-angle backscatter off small (wavelength-scale) objects such
of the planet and thus is well suited to large-scale or full-disk radar
imaging. Also unlike the quasispecular echo, the diffuse echo is
partially depolarized. The depolarized echo is particularly favor-
able for high-resolution imaging because of its freedom from
quasispecular glare and its utility for rendering regions of en-
hanced small-scale roughness as high-contrast radar brightness
features. Furthermore, the fractional depolarization of the echo
(expressed as a circular polarization ratio) can be a useful diagnos-
tic for inferring textural qualities of the surface roughness, espe-
cially when studied in the context of terrestrial radar polarimetry
and accompanying ground truth.
? Corresponding author. Address: 1902 Calle Cacique, San Juan, PR 00911, USA.
Fax: +1 787 878 1861.
E-mail addresses: harmon@naic.edu (J.K. Harmon), nolan@naic.edu (M.C. Nolan),
campbellb@si.edu (B.A. Campbell).
Icarus 220 (2012) 990?1030
Contents lists available at
Icar
.e l1 Fax: +1 202 786 2566.platy-ridged ?ows similar to those in Cerberus.
 2012 Elsevier Inc. All rights reserved.
1. Introduction
Radar echoes from Mars, like echoes from other terrestrial
as rocks or rough lava structures (Simpson et al., 1992). Unlike
the quasispecular echo, which peaks strongly around the sub-Earth
point, the diffuse echo can be traced over much of the visible disk0019-1035/$ - see front matter  2012 Elsevier Inc. A
http://dx.doi.org/10.1016/j.icarus.2012.06.030a b s t r a c t
We present Earth-based radar images of Mars obtained with the upgraded Arecibo S-band (k = 12.6 cm)
radar during the 2005?2012 oppositions. The imaging was done using the same long-code delay-Doppler
technique as for the earlier (pre-upgrade) imaging but at a much higher resolution (3 km) and, for some
regions, a more favorable sub-Earth latitude. This has enabled us to make a more detailed and complete
mapping of depolarized radar re?ectivity (a proxy for small-scale surface roughness) over the major vol-
canic provinces of Tharsis, Elysium, and Amazonis. We ?nd that vast portions of these regions are covered
by radar-bright lava ?ows exhibiting circular polarization ratios close to unity, a characteristic that is
uncommon for terrestrial lavas and that is a likely indicator of multiple scattering from extremely blocky
or otherwise highly disrupted ?ow surfaces. All of the major volcanoes have radar-bright features on their
shields, although the brightness distribution on Olympus Mons is very patchy and the summit plateau of
Pavonis Mons is entirely radar-dark. The older minor shields (paterae and tholi) are largely or entirely
radar-dark, which is consistent with mantling by dust or pyroclastic material. Other prominent radar-
dark features include: the ??fan-shaped deposits??, possibly glacial, associated with the three major Tharsis
Montes shields; various units of the Medusae Fossae Formation; a region south and west of Biblis Patera
where ??Stealth?? deposits appear to obscure Tharsis ?ows; and a number of ??dark-halo craters?? with
radar-absorbing ejecta blankets deposited atop surrounding bright ?ows. Several major bright features
in Tharsis are associated with off-shield lava ?ows; these include the Olympus Mons basal plains, volca-
nic ?elds east and south of Pavonis Mons, the Daedalia Planum ?ows south of Arsia Mons, and a broad
expanse of ?ows extending east from the Tharsis Montes to Echus Chasma. The radar-bright lava plains
in Elysium are concentrated mainly in Cerberus and include the ?uvio-volcanic channels of Athabasca
Valles, Grjot? Valles, and Marte Valles, as well as an enigmatic region at the southern tip of the Cerberus
basin. Some of the Cerberus bright features correspond to the distinctive ??platy-ridged?? ?ows identi?ed
in orbiter images. The radar-bright terrain in Amazonis Planitia comprises two distinct but contiguous
sections: a northern section formed of lavas and sediments debouched from Marte Valles and a southern
section whose volcanics may derive, in part, from local sources. This South Amazonis region shows
perhaps the most complex radar-bright structure on Mars and includes features that correspond toaNational Astronomy and Ionosphere Center, Arecibo Observatory, HC3 Box 53995, Arecibo, PR 00612, USA
bArecibo radar imagery of Mars: The majo
John K. Harmon a,?, Michael C. Nolan a, Diana I. Husm
journal homepage: wwwll rights reserved.volcanic provinces
nn a, Bruce A. Campbell b,1
SciVerse ScienceDirect
us
sevier .com/ locate/ icarus
2. Observations and data analysis
The radar imagery presented in this paper is based on observa-
tions made with the S-band (2380 MHz; k = 12.6 cm) radar on the
Arecibo Observatory?s 305-m telescope. The observations were
made during four successive Mars oppositions in 2005, 2008,
2010, and 2012. In Table 1 we list dates and relevant parameters
for the observations used to make the images presented here.
The 2005 opposition was particularly favorable for this study be-
cause of the small Earth?Mars distance and because the southerly
sub-Earth latitudes minimized N/S-ambiguity and Doppler-
equator effects (see below). Hence, most of the imagery in this pa-
per is based on the 2005 data. We supplemented this with data
from the subsequent three oppositions in order to patch the
imagery in regions corrupted by ambiguity foldover and Doppler-
equator effects or to image selected regions at better incidence
angles than in 2005.
All observations were made using the same transmission mode.
We transmitted a long (effectively non-repeating) pseudonoise
rus 220 (2012) 990?1030 991Arecibo observations during 1980?1982 (Harmon et al., 1982;
Harmon and Ostro, 1985) showed that Mars gives strong diffuse/
depolarized radar echoes and that these echoes come mainly from
the major northern volcanic provinces Tharsis and Elysium, with
some additional enhanced echoes from Amazonis. These were
simple CW (continuous wave) observations that yielded dual-
polarization Doppler spectra but no images. Progress in radar
imaging of Mars had always been thwarted by the fact that the
planet?s rapid rotation produces highly ??overspread?? echoes
unsuitable for standard (repeating-code) delay-Doppler mapping.
In fact, the ?rst radar images of Mars were made in 1988 using
bistatic Goldstone-VLA (Very Large Array) synthesis imaging rather
than delay-Doppler mapping (Muhleman et al., 1991, 1995). These
full-disk X-band (k = 3.5 cm) images mapped the bright depolar-
ized features in Tharsis, Amazonis, and eastern Elysium at 170-
km resolution. They also revealed an equatorial dark feature
(dubbed ??Stealth??) stretching from Tharsis to Elysium as well as
a very bright feature at the south pole that was interpreted to be
enhanced volume backscatter from the residual polar ice cap. The
?rst delay-Doppler radar images of the planet were ?nally made
at Arecibo in 1990 (Harmon et al., 1992b, 1999) using a novel
??long-code?? method (Harmon, 2002) that circumvents the over-
spreading problem and that was adapted from the ??coded-long-
pulse?? method developed for incoherent radar observations of
the overspread ionosphere (Sulzer, 1986; see also Hagfors and Kof-
man, 1991). These S-band (k = 12.6 cm) observations mapped the
Mars depolarized echo at a higher (50-km) resolution and pro-
duced the ?rst complete radar images of the interesting Elysium?
Cerberus region. They also mapped some moderately bright
features in the Chryse/Xanthe channel region (Harmon, 1997). Are-
cibo long-code imaging was continued through the following
(1992) Mars opposition, after which the telescope was taken out
of service for a major upgrading. By the time the upgraded tele-
scope was recommissioned (in 1998), Mars had moved south of
Arecibo?s pointing window, where it would remain through the
2003 opposition.
We commenced post-upgrade radar imaging of Mars when the
planet returned to our view with the 2005 opposition. We contin-
ued using the same basic long-code method as for the pre-upgrade
imaging, but with modi?cations designed to improve image qual-
ity and scienti?c return. First, we took advantage of the enhanced
system sensitivity to improve the mapping resolution by better
than an order of magnitude. We also modi?ed our transmission
encoding scheme to eliminate a subtle code-sidelobe leakage prob-
lem, thus enabling us to make our ?rst useful polarization-ratio
images.
In this paper we present and discuss post-upgrade Arecibo ra-
dar imagery of Mars, concentrating speci?cally on the major volca-
nic regions of Tharsis, Elysium, and Amazonis. Results for
interesting non-volcanic regions such as Chryse/Xanthe and the
polar ice caps will be deferred to a later paper; for a brief summary
of some results for these other regions see Harmon and Nolan
(2007). As with the pre-upgrade imagery, our emphasis has been
on using depolarized images to map and identify radar brightness
features. Using depolarized radar re?ectivity as a proxy for small-
scale surface roughness, we have studied the images to gain some
new insights into the surface characteristics and geomorphology of
the volcanic regions. To assist in this work we have expanded our
comparisons with other Mars surface studies, taking advantage of
the proliferation of results from the post-Viking generation of
Mars-orbiting spacecraft. We expect the new radar imagery will
be a useful input to new Mars geologic mapping projects (e.g.,
Tanaka et al., 2011). It should also be useful for landing-site hazard
J.K. Harmon et al. / Icaassessment and as a survey of the sorts of features that one would
expect to see with any future Mars-orbiting imaging radar
(Campbell et al., 2004).binary phase code with a 10-ls ??baud?? (phase-?ip interval or syn-
thesized pulse width). To avoid spurious code sidelobes from
transmission phase errors, we used a 4-tap, 240?1 length shift-
register code instead of the 2-tap, 239?1 length code used for the
pre-upgrade observations (see Harmon, 2002). Following normal
practice, we transmitted a circularly polarized wave and received
in both (orthogonal) circular polarizations. In this paper we adopt
the standard terms ??OC?? and ??SC?? to denote the received polariza-
tion senses that are, respectively, the opposite of, or the same as,
the transmitted sense. We will use these terms interchangeably
with the terms ??polarized?? (for OC) and ??depolarized?? (for SC).
The Arecibo S-band radar is a two-klystron system nominally
transmitting 800?950 kW. However, a klystron failure forced us
to run in single-klystron mode in 2005, which limited the trans-
mitter power to an average of 430 kW during that opposition.
We continued to suffer transmitter problems through the 2008
and 2010 oppositions, which resulted in reduced transmitter
power and loss of some planned observations. For the 2012 obser-
vations we were back to full-power operation, which helped to
compensate for the large planet distance during that opposition.
We performed the long-code decoding and delay-Doppler
analysis using the same general procedure described elsewhere
(Harmon et al., 1992b, 1999; Harmon, 2002). The received signal
was complex-sampled once per baud and then multiplied by a
suitably lagged replica of the code. This lagged-product time series
Table 1
Mars observing dates and parameters.
Date Times (UT) Long. (E) Lat. () D (AU) P (kW)
2005 October 3 06:38?08:12 199?222 10.99 0.517 440
2005 October 6 05:47?08:10 160?194 11.14 0.507 444
2005 October 10 05:21?07:57 117?155 11.43 0.494 435
2005 October 14 06:30?06:54 98?104 11.83 0.484 466
2005 October 15 05:51?06:14 80?85 11.94 0.482 465
2005 November 14 02:22?04:56 124?161 17.02 0.484 414
2005 November 17 02:08?04:27 94?128 17.50 0.493 429
2005 November 23 01:37?04:03 33?68 18.34 0.515 420
2008 January 6 02:13?04:27 146?179 0.90 0.624 650
2008 January 12 01:43?04:03 86?120 1.69 0.650 644
2010 January 26 03:55?06:12 131?164 15.49 0.664 630
2010 January 28 04:47?06:04 126?145 15.22 0.664 561
2010 January 29 04:31?06:10 113?138 15.09 0.664 550
2012 February 14 05:27?07:15 131?157 22.80 0.719 887
2012 February 15 05:22?07:26 121?151 22.78 0.715 822
2012 February 16 05:17?07:14 111?139 22.76 0.711 797
2012 February 17 05:11?06:57 100?126 22.73 0.707 820Notes: ??Long.?? denotes the range of sub-Earth Mars longitudes during the observing
session, D is the Mars distance, and P is transmitter power.
bright features that do not correspond to primary volcanic ?ows
and where the small-scale roughness may take the form of rock-
rus 2was boxcar smoothed and decimated by a factor of four to reduce
the Nyquist bandwidth to 25 kHz (thrice Mars?s Doppler band-
width). The spectral (Doppler) analysis was done by applying an
8192-length FFT (Fast Fourier Transform) to the 40-ls samples of
the decimated lagged-product time series, giving spectra with a
frequency resolution of (0.32768 s)1 = 3.05 Hz or an east?west
resolution of about 3 km at the sub-Earth point. We repeated this
process for a total of 2400 delay gates or lags, shifting the lag on
one-baud steps. This gave a 24-ms delay window that spanned
the 22.6-ms delay depth of Mars. A single realization (or ??look??)
of the delay-Doppler array was computed for every 0.32768 s of
receive time. Successive looks were then summed in groups of
30 to give a delay-Doppler ??snapshot?? of the Mars echo every
9.83 s, which is about the time it takes for Mars to rotate by
3 km at the equator. After subtracting a noise baseline, each snap-
shot array was mapped from delay-Doppler space to planetary (lat-
itude?longitude) coordinates and then successive snapshot images
were summed to give an average planet image for each observing
??run?? (transmit-receive cycle). These run-averaged images (up to
10 per day) were then summed over a single day or (in most cases)
multiple days to produce the ?nal images shown in this paper. All
image summing (intra-day, multi-day, and multi-year) used
weighted averages to maximize the signal-to-noise ratio at each
point in the ?nal image. This compensated for varying transmitter
power, planet distance, and system noise level. Also included in the
weighting was a factor accounting for the falloff in radar re?ectiv-
ity with increasing incidence angle (see Section 3).
The images were normalized to express each pixel value as a
dimensionless ??re?ectivity?? or ??speci?c cross section?? r0 (radar
cross section per unit surface area), using the long-code calibration
procedures described in Harmon (2002). The images are displayed
as orthographic grayscale maps, with brighter shading represent-
ing higher re?ectivity. The image resolution is about 3 km in the
east?west (longitude) direction (set by the Doppler resolution)
by about 1.5/sin(g  g0) km in the north?south (latitude) direction,
where 1.5 km is the delay depth corresponding to the synthesized
10-ls pulse and g and g0 are the pixel latitude and sub-Earth
latitude, respectively. Based on this, we quote 3 km as a reasonable
estimate of the average image resolution.
Because of the high resolution and Mars?s extreme topography,
care was taken to ensure the accuracy of the delay-Doppler map-
ping. First, for each day we estimated the delay to the planet center
(relative to the observing ephemeris). This was done by ?tting an
echo template to the leading edge of the sharp OC specular echo
from some short stretch of data in the middle of the day?s observ-
ing session and then tying this to the corresponding planet radius
(at the sub-Earth point) from the Mars Orbiter Laser Altimeter
(MOLA) data base (Smith et al., 2003). Then, for each point in the
planet (lat.?long.) map we used the corresponding MOLA radius
and this planet-center delay to determine the echo location in
the delay-Doppler array. In this way we not only corrected for
ephemeris delay errors and inter-day drifts, but also corrected
the mapping of each pixel for planet topography and ?gure.
The image analysis procedure described above was applied sep-
arately to the SC and OC polarization components of the received
echo. Since our speci?c interest is in full-disk imaging of the diffuse
(as opposed to quasispecular) component of the echo, we display
only SC images in this paper. There are three main reasons for this:
(1) the SC echo is free of the confusing effects of quasispecular
glare in the sub-Earth region; (2) since the degree of depolarization
of the diffuse echo tends to be higher for radar-bright features, the
SC images show higher feature contrast; (3) the background noise
level for the SC images is signi?cantly lower than for the OC
992 J.K. Harmon et al. / Icaimages, because of the much lower contribution from ??self-clutter??
noise. This self-clutter is an additive echo self-noise imposed by
the long-code method as the price paid for eliminating thestrewn surfaces such as sedimentary debris ?elds (Baron et al.,
1998; Campbell, 2001).
??Radar-dark?? features can also be of considerable interest.
These have typical re?ectivities of 14 to 20 dB, although some
of the darkest r030 < 19 dB
 
features are actually consistent with
zero detectable echo when ambiguity foldover from the southernoverspread echo aliasing affecting standard delay-Doppler map-
ping (Harmon, 2002). For the 2005 observations the ratio of clutter
noise to system (thermal) noise ranged over 0.30?1.24 for the SC
echo and 0.97?4.88 for the OC echo. Despite the higher clutter
noise, the OC images were still usable for making images of the cir-
cular polarization ratio.
3. Imagery: explanatory introduction and overview
In Sections 4?6 we present and discuss the depolarized (SC)
imagery over the Tharsis, Elysium, and Amazonis regions, respec-
tively. In this section we provide some introductory background
material to assist the reader in interpreting the images.
All of the images are displayed as orthographic grayscale maps
of radar re?ectivity (speci?c cross section) r0, with brighter shad-
ing representing higher re?ectivity. The grayscale for a given image
is represented by a linear stretch between zero re?ectivity and, in
most cases, the maximum re?ectivity measured over that image. In
some cases the grayscale maximum has been truncated at some
value below the maximum re?ectivity in order to reduce the
dynamic range and bring out some of the fainter features. Both
the maximum re?ectivity and, if different, the maximum grayscale
value are quoted in the caption of each ?gure.
Table 2 lists measured SC re?ectivities for selected radar fea-
tures, both bright and dark. Included in the table is the mean inci-
dence angle h for each re?ectivity measurement as well as, for
comparison purposes, a re?ectivity r030 adjusted to a common
h = 30 incidence angle assuming a r0(h) / cosnh ??scattering law??
with n = 3/2. This particular n value was selected as it is close to
the observed incidence-angle falloff for some selected bright fea-
tures in this data set and is the typical value measured for lunar
depolarized echoes and terrestrial lava ?ows. This same cos3/2h
dependence was assumed for the incidence-angle weighting factor
in the multi-run image summations. In Table 2 we also list the
re?ectivities as logarithmic values using r030 ?dB?  10log10 r030
 
,
following a convention adopted by the terrestrial remote sensing
literature.
The measured re?ectivities of bright features in this paper are in
close agreement with re?ectivities measured for the same features
in the pre-upgrade imagery (Harmon et al., 1999). The brightest
features in each major region have r030 higher than 8 or 9 dB,
with the very brightest features on the planet showing r030 ?
4:5 dB. Such high SC re?ectivities require surfaces of extreme
wavelength-scale (decimeter) roughness. For comparison, rough
Hawaiian a?a lava ?ows have SC re?ectivities of around 11 to
12 dB (Harmon et al., 1999; Campbell, 2002). As will be shown
in Section 7, the highest Mars SC re?ectivities and circular polari-
zation ratios are consistent with lava ?ows with extremely blocky
surfaces. In addition to these very bright features, we see many
??moderately bright?? features with re?ectivities of around 10 to
13 dB. Many of these moderate features have been identi?ed
with lava ?ows which for some reason (smoother surface texture,
dust mantling, lower ?lling factor?) give weaker depolarized back-
scatter than the brightest ?ows. However, we also see moderately
20 (2012) 990?1030cratered highlands is accounted for (see below). A radar-dark
feature is interpreted to be one lacking in scatterers on or near
the surface. By ??near the surface?? we mean anywhere within the
Table 2
Depolarized radar re?ectivities of selected features.
No. Long. (E) Lat. () r0 h () r030 r030 (dB) Description
3-F 227.6 16.4 0.209 30.3 0.210 6.8 Olympus Mons bright shield ?ows
3 225.2 18.3 0.148 37.0 0.167 7.8 Olympus Mons bright shield ?ows
3-G 223.2 21.6 0.017 40.0 0.020 16.9 Olympus M. lower NW shield
3-H 221.4 14.3 0.160 33.7 0.170 7.7 Olympus M. scarp draping ?ows
3-J 224.9 12.0 0.126 31.4 0.129 8.9 Olympus M. plains member ?ows
3-N 230.3 22.3 0.122 40.8 0.149 8.3 Olympus M. scarp draping ?ows
3-R 221.2 22.4 0.008 40.9 0.010 20.1 Olympus M. glacial/slide deposit
4-L 234.7 17.9 0.072 37.2 0.082 10.9 Sulci Gordii aureole unit
4-P 215.1 26.7 0.027 45.3 0.037 14.3 Lycus Sulci aureole unit
4-Q 217.9 24.4 0.013 43.0 0.016 17.8 Eolian deposit
5-C 252.3 34.5 0.061 47.3 0.088 10.5 Alba Patera sheet ?ows
5-D 262.2 41.0 0.060 53.9 0.109 9.6 Alba Patera lava apron
7-A 255.7 10.2 0.220 24.6 0.204 6.9 Ascraeus Mons upper shield
7-C 254.1 3.1 0.200 22.7 0.182 7.4 Ascraeus south vent ?eld
7-D 252.6 0.3 0.205 20.6 0.182 7.4 Pavonis east lava apron
7-E 255.5 19.7 0.106 37.4 0.121 9.2 Ascraeus north fan ?ow ?eld
7-G 246.4 1.6 0.141 28.0 0.137 8.6 Pavonis Mons bright shield collar
7-H 244.9 4.9 0.015 30.2 0.015 18.1 Pavonis Mons fan-shaped deposit
7-J 235.5 5.0 0.156 23.6 0.143 8.4 Pavonis south volcanic ?eld
7-K 258.8 3.3 0.027 17.3 0.024 16.3 Noctis Fossae
7-L 235.5 2.6 0.020 26.9 0.019 17.1 Biblis Patera
7-M 238.5 2.7 0.035 27.3 0.034 14.7 Uranius Patera
7-P 237.4 13.0 0.032 35.7 0.035 14.5 Ulysses Fossae
7-Q 240.2 6.0 0.154 30.2 0.155 8.1 NW Tharsis ??bench?? ?ows
7-R 244.0 12.9 0.119 36.5 0.133 8.8 NW Tharsis ??anvil?? ?ows
7-V 242.6 18.1 0.025 40.7 0.030 15.2 Jovis Tholus
7-W 241.4 22.2 0.096 44.1 0.128 8.9 Ceraunius Fossae Formation
7-Y 256.4 25.3 0.022 40.8 0.027 15.7 Ceraunius Fossae
9a-K 252.3 10.7 0.016 29.5 0.016 18.1 Ascraeus Mons fan-shaped deposit
9a-M 253.4 8.3 0.033 27.3 0.031 15.0 Ascraeus M. south rift apron core
9a-R 252.0 4.4 0.265 24.0 0.245 6.1 Bright Ascraeus south apron ?ows
9b-K 247.9 1.4 0.028 22.5 0.026 15.9 Pavonis Mons summit plateau
9b-R 245.3 1.7 0.047 25.4 0.044 13.5 Pavonis Mons south rift apron
9c-V 239.7 8.9 0.140 33.4 0.148 8.3 Arsia Mons caldera
9c-Z 238.0 12.2 0.113 35.8 0.124 9.0 Arsia Mons south rift apron core
11-A 239.3 7.2 0.286 27.9 0.277 5.6 Arsia Mons shield collar (N. slope)
11-A 238.5 7.3 0.366 27.7 0.354 4.5 Arsia M. shield collar (NW spot)
11-B 233.8 6.2 0.040 29.7 0.039 14.0 Arsia Mons fan-shaped deposit
11-E 236.0 11.7 0.200 34.9 0.217 6.6 Arsia M. s. rift apron bright ?ows
11-F 238.2 15.9 0.120 39.1 0.141 8.5 Daedalia Planum At6 ?ows
11-G 237.9 18.6 0.153 41.6 0.191 7.2 Daedalia Planum At5 ?ows
11-P 222.8 22.4 0.214 12.3 0.179 7.5 western Daedalia ?ows
11-W 223.5 28.6 0.147 16.5 0.126 9.0 Daedalia Planum bright ?ow ?nger
11-Z 221.6 35.2 0.098 21.3 0.088 10.6 Daedalia Planum S. ?ow terminus
14-A 268.2 4.5 0.038 32.1 0.039 14.1 Fortuna Fossae
14-C 276.0 2.5 0.033 31.2 0.033 14.8 Echus Plateau
14-C 278.6 3.5 0.022 32.0 0.022 16.5 Echus Plateau (dark E. edge)
14-D 279.6 2.7 0.088 31.5 0.090 10.5 Echus Chasma
14 273.1 8.0 0.080 34.8 0.087 10.6 E. Tharsis ?ows (south of 14-J)
14 272.5 10.0 0.050 36.3 0.056 12.5 E. Tharsis ?ows (north of 14-J)
14-F 259.5 13.4 0.134 38.5 0.156 8.1 Flows NE of Ascraeus Mons
19-A 148.4 24.4 0.138 42.0 0.173 7.6 Elysium Mons (shield east ?ank)
19-D 146.0 21.9 0.023 40.6 0.028 15.5 Dark-halo crater
19-F 158.4 19.2 0.097 35.8 0.107 9.7 ??ESE band?? ?ows
19-H 151.0 31.4 0.076 47.5 0.110 9.6 Hecates Tholus (shield east ?ank)
21-C 154.3 6.6 0.147 26.5 0.140 8.5 Athabasca Valles (south channel)
21-E 149.9 7.7 0.163 28.5 0.159 8.0 Cerberus Palus (NW part)
21-N 151.9 1.3 0.010 23.5 0.009 20.3 Zephyria Planum (MFF)
22-B 165.4 15.5 0.102 32.0 0.105 9.8 Grjot? Valles
22-D 168.2 12.0 0.145 29.0 0.143 8.5 Grjot? Valles trough
22-J 171.2 7.4 0.115 25.4 0.108 9.7 Cerberus Fossae ?ows
22-R 163.1 12.2 0.034 29.4 0.034 14.7 Tartarus Montes
22-X 166.2 7.7 0.022 25.7 0.020 16.9 Zunil Crater
23-A 163.4 1.5 0.124 19.6 0.110 9.6 South Cerberus
23-B 165.2 1.2 0.067 21.1 0.060 12.2 Darker ?ows north of S. Cerberus
23-C 166.0 3.7 0.136 22.8 0.124 9.1 Central Cerberus ?ows
23-S 171.4 2.0 0.182 21.5 0.164 7.9 Marte Valles source ?ows
25 181.8 11.5 0.175 36.4 0.195 7.1 Marte Valles
26 194.9 33.3 0.098 50.8 0.158 8.0 Northern North Amazonis ?ows
26-D 209.2 29.3 0.053 46.5 0.075 11.2 Outer North Amazonis ?ows/?uvia
26-P 182.1 2.2 0.017 27.0 0.017 17.8 Lucus Planum (MFF)
29-A 204.9 19.9 0.122 38.3 0.142 8.5 South Amazonis ??wedge?? ?ows
29-C 200.2 19.0 0.086 37.8 0.099 10.0 Modi?ed volcanic plains
29-D 200.6 15.4 0.134 34.9 0.145 8.4 Localized bright S. Amazonis ?ows
(continued on next page)
J.K. Harmon et al. / Icarus 220 (2012) 990?1030 993
characteristic attenuation depth of any overlying mantling mate-
rial such as eolian sand or dust, pyroclastic airfall, crater ejecta,
or regolith. A typical attenuation depth for a dry terrestrial rock
powder is about 10 wavelengths (Campbell and Ulrichs, 1969), or
about a meter at our wavelength, although greater penetration
depths might be possible under extremely dessicated martian con-
ditions (Simpson et al., 1992). In this paper we will show several
examples of radar-dark terrain where absorbing mantles appear
to be masking underlying lava ?ows that would otherwise be
radar-bright.
Delay-Doppler mapping has an inherent north?south mapping
and Elysium regions can be seen on the right and left sides of the
image, respectively, and the Amazonis region can be seen in the
middle. The only major bright feature not covered in this panorama
is the Daedalia Planum region south of the Doppler equator in ex-
treme southern Tharsis. The ambiguity foldover from this feature,
including the aforementioned smearing, can be seen west of the
bright Arsia Mons feature in the south-central part of the image.
No major smeared ambiguity-foldover features are discernable in
either Elysium or Amazonis. In Fig. 2 we show an image from a sin-
gle day (November 17, 2005) covering the entire Tharsis region,
including southern Tharsis and Daedalia Planum. Note, again, the
Table 2 (continued)
No. Long. (E) Lat. () r0 h () r030 r030 (dB) Description
29-G 208.0 23.2 0.017 41.1 0.021 16.8 Tooting Crater
29-K 200.6 7.7 0.012 28.7 0.012 19.3 Eumenides Dorsum (MFF)
29-N 212.7 6.4 0.020 27.1 0.019 17.2 Gordii Dorsum (MFF)
29-S 210.0 17.6 0.109 36.2 0.121 9.2 Aureole runoff unit
Notes: ??No.?? gives the ?gure number and letter label (if there is one) for the feature. The coordinates give the center of a rectangular region over which the feature re?ectivity
was averaged. r0 is the depolarized (SC) re?ectivity, h is the corresponding mean incidence angle, r030 is the re?ectivity adjusted to a common incidence angle of 30, and r030
(dB) is the re?ectivity in logarithmic format.
994 J.K. Harmon et al. / Icarus 220 (2012) 990?1030ambiguity about the so-called ??Doppler equator?? (an E/W-trending
line intersecting the sub-Earth point and orthogonal to the appar-
ent spin axis). Proximity to the Doppler equator also results in im-
age degradation from delay projection effects and tangency of the
delay and Doppler annuli. Fortunately, the sub-Earth geometry in
2005 mitigated much of the potential confusion from N/S-ambigu-
ity foldover. Most of the interesting radar-bright features were
concentrated in the volcanic plains north of the Doppler equator,
and these features suffered little noticeable confusion from the
much darker, radar-bland cratered highlands south of the Doppler
equator. Furthermore, the Doppler equator was suf?ciently far
from the true equator that much of the ambiguity foldover was
heavily smeared out by changing delay-Doppler geometry over
the image summation period. Some of these points are illustrated
in the panorama image in Fig. 1. This image, which is a montage
of smaller regional images, shows all of the major radar-bright vol-
canic features north of the 2005 Doppler equator (the horizontal
dark bands near the bottoms of the image panels). The TharsisFig. 1. Depolarized (SC) radar image panorama of Mars covering most of the volcanic reg
Tharsis (center right and right). This is a montage of four image panels spliced at 190E, 2
south of the Doppler equator. The Doppler equator appears as a radar-dark band at 10S
image. The gray-scale maximum is set to a re?ectivity of 0.30.smeared foldover from the Daedalia Planum bright feature and
mapping degradation near the Doppler equator, as well as smeared
foldover of northern features into the southern hemisphere. Note
also the lack of crisp radar-bright features south of the Doppler
equator, except in Daedalia Planum. This shows the lack of distinc-
tive radar-bright features in the heavily-cratered southern
highlands. In Section 4.5 we will present additional imagery from
the 2010 and 2012 oppositions that resolves most of the mapping
corruption and ambiguity in the equatorial region around Arsia
Mons and northern Daedalia Planum. Finally, note that even if
the radar image of a particular region shows no discernable foldo-
ver because the ambiguity region is in the cratered highlands, there
will still be some small contribution to the measured re?ectivity
from the ambiguity region. From pre-upgrade delay-Doppler and
CW observations we have long known that the typical depolarized
re?ectivity of the cratered highlands is about 1% or so (Harmon and
Ostro, 1985; Harmon et al., 1999). This is con?rmed by our Novem-
ber 28, 2005 long-code observations of Terra Meridiani (meanions treated in this paper: Elysium (left), Amazonis (center left), Olympus Mars and
20E, and 260E longitudes. Not covered is the radar-bright Daedalia Planum region
latitude on the left side of the image and at the very bottom on the right side of the
member ?ows; K, dark-halo crater. The gray-scale maximum is set to re?ectivity of
0.284 (5.47 dB), which is 85% of the maximum re?ectivity measured over this
rus 2Fig. 2. Radar image of the Tharsis region, including Olympus Mons. This is from
observations made on November 17, 2005. Note that the only major bright feature
south of the Doppler equator is Daedalia Planum. N/S-ambiguity foldover of bright
features north of the Doppler equator appear as blurred features at the bottom of
the image, and foldover from Daedalia Planum accounts for the blurred feature west
of Arsia Mons. The linear bright features at center right are layover glints from the
north walls of the Valles Marineris channels.
J.K. Harmon et al. / Icasub-Earth point: 10E, 19S), from which we estimated an average
value of r030 ? 0:0115?19:4 dB? for the depolarized re?ectivity in
the cratered highlands. Since we have not corrected the re?ectivi-
ties in Table 2 for ambiguity foldover, the reader should keep in
mind that a even a perfect radar absorber will still have a quoted
re?ectivity of around 19 to 20 dB.
In the following three sections we survey and discuss the SC ra-
dar images over Tharsis (Section 4), Elysium (Section 5), and Amaz-
onis (Section 6). We follow this with an analysis of the circular
polarization ratio measurements and a discussion of their implica-
tions for the volcanic texture over these regions (Section 7). In the
last section before the conclusion we discuss the so-called ??dark-
halo craters?? (Section 8), examples of which have been found in
all three volcanic provinces.
Most of the radar features discussed in Sections 4?6 are identi-
?ed by letter labels (A?Z) on the accompanying image ?gures. The
corresponding feature callout in the text will consist of the ?gure
number followed by a hyphen and the feature letter label; for
example, feature A in Fig. 3 is denoted by ??3-A??. Although the same
letter may be reused for different features on other ?gures, any
given feature is always denoted by the same letter when labeled
in multiple ?gures.
At various places on the following presentation we will provide
spacecraft-derived images for context and comparison purposes.
These will include shaded-relief and vertical roughness maps from
the MOLA (Mars Orbiter Laser Altimeter) instrument on Mars Glo-
bal Surveyor, and infrared (IR) images from the THEMIS (Thermal
Emission Spectrometer) instrument on Mars Odyssey (Christensen
et al., 2004). The MOLA vertical roughness maps, which give sur-
face roughness on 0.6?9.2 km horizontal scales, are particularly
useful as they show not only the major shield volcanoes and im-
pact craters but also some of the lava plains units and debris
deposits that are not always readily apparent in shaded relief
maps.Fig. 3. Radar image of Olympus Mons. This is the sum of imagery from October 10,
November 14, and November 17, 2005. The labeled features are: A and B, Olympus
Mons perimeter scarp; C, Pangboche Crater; D, Karzok Crater; E, Olympus Mons
caldera; F, brightest region of the summit plateau, where there is a high
concentration of tube-fed ?ows; G, radar-dark ??soft?? terrain on the lower northwest
shield slopes; H and N, Olympus Mons scarp-draping ?ows; J, Olympus Mons plains20 (2012) 990?1030 9954. Tharsis
The general Tharsis region, in which we include Olympus Mons,
Alba Patera, and the volcanic plains stretching east to Lunae Pla-
num, presents a complex array of radar-bright and radar-dark fea-
tures. Early dual-polarization CW measurements of radar Doppler
spectra showed Tharsis to be the site of unusually strong depolar-
ized echoes, which were attributed to scattering off very rough
volcanic surfaces (Harmon et al., 1982; Harmon and Ostro, 1985;
Thompson and Moore, 1989; Moore and Thompson, 1991). Subse-
quent Goldstone-VLA synthesis imaging (Muhleman et al., 1991)
and Arecibo delay-Doppler imaging (Harmon et al., 1999) identi-
?ed radar-bright features with known volcanic constructs and lava
?elds and gave some indication of the radar complexity of the
region.
In this section we take a detailed look at the depolarized radar
features from the various subregions covered by Fig. 2, beginning
with the enormous shields Olympus Mons and Alba Patera, contin-
uing with the major Tharsis Montes volcanoes (Ascraeus, Pavonis,
and Arsia Montes) and Daedalia Planum, and concluding with two
subsections on the volcanic ?ow ?elds on the western and eastern
slopes of the Tharsis Rise.
4.1. Olympus Mons
The pre-upgrade Arecibo imagery showed the Olympus Mons
region to be dominated by bright off-shield ?ows, although the
shield itself also showed some bright features (Harmon et al.,
1999). This is borne out by the post-upgrade Arecibo imagery of
the region. Fig. 3 shows a radar image of Olympus Mons and
Fig. 4 shows an expanded view of the general region. Context for
region.
rus 2996 J.K. Harmon et al. / IcaFig. 4 is provided by the MOLA shaded relief and vertical roughness
maps in Fig. 5.
Some prominent structural features of the shield include the
perimeter scarp (3-A, 3-B) and the superimposed impact craters
Pangboche (3-C) and Karzok (3-D). The bright glint off the south-
east section of the scarp (3-A) is likely enhanced by the radar ??lay-
over?? effect, whereas the brightness of the northwest section of
Fig. 4. Radar image of the Olympus Mons region. This is the sum of imagery from Oct
Olympus Mons include: H and N, Olympus Mons scarp-draping ?ows; J, Olympus plain
aureole unit; Q, eolian deposits; R, glacial or slide deposits; S, putative aureole runoff dep
maximum re?ectivity measured over this region.
Fig. 5. Orbiter-based context maps of the Olympus Mons region: (a) MOLA shaded-relief
scale roughness. The map boundaries are the same as for the radar image in Fig. 4.20 (2012) 990?1030scarp (3-B) may be enhanced backscatter off slope rubble. The lack
of layover glints from the southwest section of scarp is probably
due to slope reduction by scarp-draping ?ows. Craters Pangboche
and Karzok show dark ejecta and moderately bright ?oors. Pangb-
oche shows up particularly well because its dark halo and rays con-
trast sharply with the surrounding bright terrain. The geologic map
of Scott et al. (1981a) has Karzok as younger than, and Pangboche
ober 10, November 14, and November 17, 2005. The labeled features surrounding
s member ?ows; K, dark-halo craters; L, Sulci Gordii; M, Cyane Sulci; P, Lycus Sulci
osits. The gray-scale maximum is set to re?ectivity of 0.284 (5.47 dB), or 85% of the
map; (b) MOLA vertical-roughness map, with lighter shading indicating higher km-
rus 2as older than, the surrounding terrain. However, the radar imagery
shows Pangboche, like Karzok, to be a fresh crater superimposed
on the Olympus ?ows. The darkness of the haloes around these
craters is consistent with meters-thick, ?ne-grained ejecta deposits
draped over the bright ?ows (see Section 8 on ??dark-halo craters??).
The Olympus Mons caldera complex (3-E) is only slightly darker
than the bright summit ?ows, which suggests that the caldera ?oor
is moderately rough and not heavily mantled. The brightest part of
the summit plateau (3-F) coincides with the shield?s highest con-
centration of tube-fed lava ?ows (Richardson et al., 2009). Other
bright patches surrounding the caldera and Pangboche, and
extending down the shield ?ank to the west and north, coincide
with concentrations of lava tubes, fans, and channels (Bruno
et al., 2005; Richardson et al., 2009). The dark northwest corner
of the shield (3-G) corresponds to an area largely devoid of lava
?ow features (Basilevsky et al., 2005a,b; Richardson et al., 2009).
The lower southern and southwest slopes of the shield plateau also
show some dark areas, but these are crossed by radial bright
spokes that correlate well with lava tubes that mostly terminate
in lava fans near or below the basal scarp (Richardson et al.,
2009). These streaks also agree with ?ow streamlines on early
geologic maps of the region (Scott et al., 1981b). Basilevsky et al.
(2005a,b) have identi?ed ??soft?? terrains on the western shield pla-
teau that they suggest are dust, volcanic ash, or glacial deposits.
This could explain the radar-dark regions that we see around the
western edge of the shield.
The bright apron south of the shield has the appearance of being
a single unit, but in fact is more complex. The bright off-shield lobe
(3-H, 4-H) abutting the southwest quadrant of the shield has been
mapped as post-scarp or ??scarp-draping?? ?ows (designated either
map unit Aos or Aom2) originating from the shield (Scott et al.,
1981b; Scott and Tanaka, 1981a, 1986; Plescia, 2004). It is interest-
ing that the off-shield Aos lavas are much more uniformly bright
than the southwest part of the shield plateau from which they
are presumed to have ?owed. Either the off-shield Aos lavas were
fed by, and spread out from, a series of restricted lava channels on
the shield, or portions of the southwest edge of the shield plateau
were mantled with radar-dark deposits during or after emplace-
ment of the Aos ?ows. Also, the high radar brightness of this unit
suggests that the Aos ?ow surfaces roughened in the process of
spreading out over the off-shield plains.
South of the off-shield Aos ?ows, and abutting the southeast
edge of the shield, are the bright ??Olympus Plains Flows?? or ??Olym-
pus Mons Plains Member?? (3-J, 4-J), designated unit Aop in geo-
logic maps (Scott et al., 1981b; Scott and Tanaka, 1981a, 1986).
The Aop and Aos units are easily distinguishable based on their dif-
ferent MOLA vertical roughness (Fig. 5b). A portion of the contact
between the Aos and Aop units (along 220?225E at 13N) shows
up as a subtle brightness contrast in the radar image (Figs. 3 and
4), with the Aos unit being brighter (hence rougher at small scales).
The Aop ?ows are younger than the Aos ?ows and do not originate
from the shield but rather from vents or low-relief volcanic con-
structs east and southeast of the shield (Scott and Tanaka, 1986;
Hodges and Moore, 1994). Punched into these Aop plains are three
impact craters (4-K) with dark ejecta (see Section 8); the western-
most of these shows some extended radar-dark rays and was
mapped by Scott and Tanaka (1981a) as a younger, superimposed
feature. This portion of the Aop plains shows some streaking sug-
gestive of variations in ?ow surface texture. The bright Aop unit
continues to wrap around the east side of the shield, ?owing
around Sulci Gordii (4-L) and Cyane Sulci (4-M) to its northern ter-
minus near Acheron Fossae. Here the brightness variations are
even more marked, with the ?ows around Sulci Gordii being dis-
J.K. Harmon et al. / Icatinctly brighter than those farther north. This indicates different
?ow textures suggestive of different source vents and/or episodic
emplacement. Also, the ?ows on the east side of Cyane Sulci appearto include a contribution from the Ceraunius Fossae Formation
?ows originating farther east, as we will discuss in Section 4.6. Fi-
nally, note the bright lobe (3-N, 4-N) just off the northeast side of
the shield. This is not mapped as plains member Aop but rather as
an off-shield lobe or scarp-draping ?ow of the shield member Aos
(Scott and Tanaka, 1986; Plescia, 2004). This radar feature, and the
adjacent bright patches on the lower northeast slopes of the shield
itself, show a high concentration of lava ?ows in spacecraft optical
imaging (Scott et al., 1981a; Richardson et al., 2009). Like the off-
shield Aos ?ows (3-H, 4-H) in the southwest, this northeastern lobe
appears brighter than the source ?ows on the shield itself, again
suggesting that the ?ows roughened as they spread onto the plains.
Other major Olympus Mons off-shield members include the
aureole deposits and debris aprons. The aureole deposits, though
topographically extremely rough (see Fig. 5), have a relatively
low radar brightness indicating low decimeter-scale roughness or
rock cover. Aureole deposits east of the Olympus Mons shield in-
clude Sulci Gordii (4-L) and Cyane Sulci (4-M), which appear as
darker islands surrounded by the bright Aop ?ows. To the north
and west of the shield are the vast aureole deposits of Lycus Sulci.
The Aoa4 member (Scott and Tanaka, 1986) of the Lycus Sulci aure-
ole can be seen (4-P) in the northwest corner of Fig. 3 (see also
Fig. 2). Though faint, it is still brighter than the adjacent region
(4-Q) to the east, which is mapped as eolian deposits (unit Ae)
by Scott and Tanaka (1986). A thick eolian mantle is certainly con-
sistent with the radar darkness of the 4-Q region. Adjacent to this is
another dark lobe (3-R, 4-R), abutting the west and northwest edge
of the shield, which was mapped as ??slide?? material (unit As) by
Scott and Tanaka (1986) and a ??debris apron?? by Hodges and
Moore (1994). More recently it has been interpreted as a glacial de-
posit analogous to the ??fan-shaped deposits?? of the three Tharsis
Montes shields (Milkovich and Head, 2003; Milkovich et al.,
2006). If this deposit consists of glacial debris, then its radar-dark
appearance indicates that its surface cannot be rocky and must
instead consist of some radar-absorbing material such as a ?ne
glacial till or eolian mantle. Keszthelyi and Grier (2002) have iden-
ti?ed ??very thick?? mantling deposits over portions of this radar-
dark lobe, particularly below the northwest corner of the shield.
Finally, outside the southwest edge of the Lycus Sulci aureole is a
bright boomerang-shaped feature (4-S) that has been mapped as
either South Amazonis volcanics (Tanaka et al., 2005) or runoff
debris from the aureole (Fuller and Head, 2002). The fact that this
unit appears distinct from the rest of Amazonis in MOLA roughness
maps (Fig. 5b) suggests that it is an aureole-related feature rather
than a volcanic feature, as we will discuss in more detail in Section
6.2.
4.2. Alba Patera
The enormous volcano Alba Patera was not well situated for ra-
dar imaging. The 2005 imagery covered the region with negligible
north?south ambiguity, but at high incidence angles that de-
pressed the brightness of the radar features (see Figs. 1 and 2).
The more northerly tracks in 2012 brought out some of Alba?s radar
features more clearly, but at the expense of ambiguity foldover in
some regions. Nevertheless, combining the imagery from the 2005,
2008, and 2012 oppositions has now given us a fairly good radar
picture (Fig. 6) of this volcano. We ?nd that, like Olympus Mons,
Alba Patera shows complex radar structure.
Alba?s summit caldera (6-A) is almost entirely radar-dark. This
is consistent with evidence for heavily-mantled or pyroclastic sur-
faces in the Alba summit region (Mouginis-Mark et al., 1982b,
1988; Plescia, 2004; Keszthelyi et al., 2008). Around the ?anks of
20 (2012) 990?1030 997the volcano are several regions of moderate radar brightness. A
bright streak (6-B) immediately east of the caldera correspond to
the upper reaches of some narrow sheet ?ows, while a much larger
rus 2feature (6-C) in the southeast quadrant of the lower shield corre-
sponds to a region of massive sheet ?ows (Cattermole, 1988,
1990). West and northwest of the caldera is an extended bright re-
gion (6-D) that contains some major E/W-aligned lava ridges (Cat-
termole, 1988, 1990). A fainter feature can be seen to the south on
the lower southwest part of the shield (6-E). Both western features
Fig. 6. Radar image of Alba Patera. This is the sum of imagery from October 14,
October 15, and November 17, 2005, and January 12, 2008. The labeled features are:
A, Alba Patera caldera; B, narrow Alba sheet ?ows from the caldera; C, Alba Patera
southeast sheet ?ows; D and E, Alba late-stage lava apron; F, low-thermal-inertia
region; W, Ceraunius Fossae Formation (CFF). The image has been smoothed over
3  3-pixel blocks and the gray-scale maximum is set to the maximum re?ectivity
of 0.202 (6.95 dB) measured over this region. Also shown (inset) is an image of the
Alba Patera summit region, on the same scale, from data obtained on February 14?
17, 2012.998 J.K. Harmon et al. / Ica(6-D,E) lie within a west-?ank lobe mapped by Cattermole as the
youngest summit-related ?ows and identi?ed by Plescia (2004)
as a late-stage lava apron. Separating 6-D and 6-E is a dark feature
(6-F) that coincides with a region of low thermal inertia (Zimbel-
man and Mouginis-Mark, 1987a). This may be a band of mantling
material, possibly pyroclastic, stretching across the center of Cat-
termole?s western lobe. Heavy mantling may also explain the lack
of bright features on the summit plateau north and northeast of the
caldera (6-A). Small-valley networks in this region suggest a friable
surface made up of either pyroclastic deposits (Zimbelman and
Mouginis-Mark, 1987b; Mouginis-Mark et al., 1988) or ice-rich
dust (Ivanov and Head, 2006).
4.3. Ascraeus Mons
Ascraeus Mons shows up prominently in a radar image of cen-
tral and northern Tharsis shown in Fig. 7. In our discussion of this
volcano we will also refer to MOLA context images in Fig. 8 and a
detail image (with extra labeling) of the Ascraeus Mons region in
Fig. 9a.
The Ascraeus Mons shield has a very bright, symmetrical radar
feature (7-A, 9a-A) centered on the caldera. The caldera itself (9a-
G) is dark, indicating a ?oor surface that is smooth and/or heavily
mantled. This is consistent with orbiter imagery, which shows the
caldera ?oor to be a very young, smooth surface devoid of craters,
vents or ?ows and presumably covered with a mantle of dust
(Crumpler and Auberle, 1978; Murray et al., 2010). The caldera
has a low thermal inertia indicating at least several centimeters
of dust (Zimbelman, 1984; Zimbelman and Edgett, 1992a,b), while
the radar darkness would be consistent with a meter or more of
dust. The surrounding bright oval shield feature is somewhat smal-ler than the shield itself (compare Figs. 7 and 8), extending roughly
3/5 of the way across the shield except in the southeast, where it
reaches the shield base. This bright feature is consistent with
studies based on orbiter imagery, which reveal a dense array of
fresh-appearing lava ?ow channels extending from the caldera
rim to well down the upper ?anks of the shield (Mouginis-Mark,
1981, 1982b; Zimbelman and Greeley, 1983; Zimbelman, 1985a;
Zimbelman and McAllister, 1985; Mouginis-Mark and Christensen,
2005; Hiesinger et al., 2007). Rheological studies of the Ascreaus
summit ?ows suggest that the lavas are likely to have a basaltic/
andesitic composition and an a?a ?ow texture (Zimbelman,
1985b; Zimbelman and McAllister, 1985; Hiesinger et al., 2007).
Such a rough texture would be consistent with a radar-bright fea-
ture, although the extreme brightness and high circular polariza-
tion ratios suggest that the ?ow texture is likely to be blockier
than is typical of a?a lavas (see Section 7). The brightness of the
radar feature indicates that the summit and upper ?anks of the
shield cannot be heavily mantled with dust. Zimbelman (1984,
1985a) argued that the low thermal inertia of the summit area
requires at least 2 cm of cover with ?ne-grained material. A few
centimeters of mantle would be easily penetrated by the radar
wave, but a meter or so of dust would probably produce enough ra-
dar attenuation to be inconsistent with the observed brightness of
the upper shield. It has been noted that the fresh optical appear-
ance of the summit ?ows argues against the presence of volcanic
ash deposits in this region (Mouginis-Mark, 1982b; Zimbelman
and Greeley, 1983; Zimbelman, 1984), and the radar imagery sup-
ports this.
The drop in radar brightness on the lower ?anks of the shield is
consistent with the fact, long known from Viking imagery, that the
Ascraeus lava ?ow features become more subdued at lower eleva-
tions (Zimbelman and Greeley, 1983; Zimbelman, 1985a). This
could be a result of the limited length of the young summit ?ows,
an increase in the depth of obscuring mantle deposits at the lower
elevations, or both. Recent studies of the eastern ?ank show the
younger ?ows to be concentrated toward the summit region (Hie-
singer et al., 2007). The radar brightness drops east of 257.5E (9a-
H), where it merges with the more moderate (but still enhanced)
brightness typical of the East Tharsis plains. Zimbelman (1985a)
argued from crater obscuration studies that surface dust may
exceed 15 m depths at the volcano?s base. However, there is no evi-
dence from the radar imagery for such thick mantling deposits on
the volcano?s lower eastern ?ank. The situation appears to be
different on the volcano?s western ?ank, where the bright shield
echoes disappear almost completely west of 254E longitude (9a-
J). This is part of a large radar-dark lobe that extends well beyond
the shield?s western base to nearly 249E longitude. This dark lobe
includes the so-called ??fan-shaped deposit?? or FSD (9a-K), a unit
designated as ??slide material?? (unit As) by Scott and Tanaka
(1986) and identi?ed as a likely glacial deposit by Kadish et al.
(2008). The darkness of this unit indicates that any glacial lag de-
posit is in the form of a ?ne-grained till rather than rocky rubble,
or that the deposits are mantled by radar-absorbing dust. Between
the FSD and the shield is a ??degraded ?ank?? region (9a-L) where
Kadish et al. (2008) identi?ed dunes that may have formed when
westerly winds dropped dust at the shield base. The radar-dark-
ness of this region is at least consistent with this. Also, if such wes-
terly winds deposited dust farther east, this could help to explain
the radar-darkness of the lower western slopes of the shield itself.
The possibility of radar-dark glacial deposits collecting on the low-
er western portion of the shield should also be considered.
The region south of the shield appears interesting and complex
in orbiter optical imagery as well as the radar imagery. The domi-
20 (2012) 990?1030nant feature here is a SSW-trending constructional feature (Fig. 8a)
referred to variously as a lava fan, apron, or rift apron (Hodges and
Moore, 1994; Scott and Wilson, 1999; Plescia, 2004; Bleacher et al.,
rus 2J.K. Harmon et al. / Ica2007b; Murray et al., 2010). Although the apron is usually
described as a lava-?ow feature (Hodges and Moore, 1994; Plescia,
2004), its central core appears in orbiter imagery as smooth terrain
cut by sinuous rilles (Fig. 10a) and in our radar imagery as a dark
feature (9a-M). Several authors proposed this to be a site of late-
stage meltwater release and pointed to some features that may
be mud ?ows or water-lubricated debris ?ows (Scott and Wilson,
1999; Mouginis-Mark and Christensen, 2005; Murray et al.,
2010). This might be consistent with the radar-darkness of the
region SSW of the shield (9a-M), if one assumes a typical mud ?ow,
say, to be smooth-textured and non-rocky. Another explanation for
the soft, radar-dark apron terrain is that it is pyroclastic. Scott and
Wilson (1999) did, in fact, suggest that the apron rilles might be
formed by water seepage in ash deposits.
Fig. 7. Radar image of north-central Tharsis and the Tharsis west slope. The left side (west
right side is the sum of imagery from October 14, October 15, and November 17, 2005. T
from the Tharsis ridge; C, Ascraeus south vent ?eld; D, Pavonis east lava apron; E, ?ow
Mons; G, Pavonis Mons bright shield collar; H, Pavonis Mons fan-shaped deposit; J, Pavon
dark lava ?ows presumably mantled by eastern tongue of Stealth; P, Ulysses Fossae; Q
Tharsis ?ows (??anvil??); S, Poynting Crater and ejecta (??Poynting wedge??); T, long lava ?ow
Fossae Formation (CFF); X, notch feature where CFF ?ows bend northwest; Y, Ceraunius
re?ectivity of 0.30 (5.23 dB), or 92% of the maximum re?ectivity measured over this
(4.69 dB), or 80% of the maximum re?ectivity measured over this region.20 (2012) 990?1030 999On the east ?ank of the south rift apron (255E long.) space-
craft imagery (Fig. 10a) shows a transition from the smooth apron
terrain to rougher terrain with ?ow features identi?ed as lava
channels (Bleacher et al., 2007b; Schierl et al., 2012; Signorella
et al., 2012). At this same meridian we see a sharp west-to-east
transition from dark to bright in our radar image. The heads of
the lava channels show up as arcuate radar-bright features
(9a-N) that drape due south of the Ascreaus Mons shield. The chan-
nels appear as if emerging from the southeast edge of the smooth
unit in the spacecraft imagery (Fig. 10a), which suggests that the
apron lavas were emplaced ?rst and then had their upper reaches
covered by the radar-absorbing smooth unit. Even if the smooth
unit is ?uvial, the radar brightness of the emergent channel fea-
tures supports their identi?cation as volcanic ?ows.
of 248E) is the sum of imagery from November 14 and November 17, 2005, and the
he labeled features are: A, Ascraeus Mons shield; B, northwest-trending lava ?ows
?eld from Ascraeus north lava fan; F, northeast-trending lava ?ows from Ascraeus
is south volcanic ?eld; K, Noctis Fossae; L, Biblis Patera; M, Ulysses Patera; N, radar-
, northeast-trending lava ?ows on topographic bench; R, combined north-trending
from the east ?ank of region B; U, dark-halo craters; V, Jovis Tholus; W, Ceraunius
Fossae; Z, outlier CFF deposits. The gray-scale maximum for the left side is set to a
region. The gray-scale maximum for the right side is set to a re?ectivity of 0.34
rus 21000 J.K. Harmon et al. / IcaThe south apron lava channels ?ow downslope to the east and
curve around to join the vast expanse of radar-bright ?ows (9a-
P) making up the eastern Tharsis plains (see Section 4.7). On the
west ?ank of the rift apron the terrain is mostly radar-dark and
indistinguishable from the rest of the dark lobe west of the shield.
This is consistent with the smoother appearance of the terrain on
this side of the apron. Apparently, the terrain-softening in?uences
were more active on this side of the rift, and this may account for
the radar-dark terrain south and west of the FSD. A curving lobe of
moderately bright apron ?ows (9a-Q) can be seen in the outer
southwest part of the apron, in a region where orbiter imagery
shows fresh-appearing lava ?ows emerging from the smooth unit
and curling around to the west. Bordering this on the south is a
very bright ?ow lobe (9a-R) that appears to be curving around
out of the south end of the Ascraeus apron and joining the promi-
nent lobe (7-B) of bright lavas ?owing northwest off the Tharsis
ridge between Ascraeus and Pavonis Montes.
Fig. 8. Orbiter-based context maps of north-central Tharsis: (a) MOLA shaded-relief map
roughness. The map boundaries are the same as for the radar image in Fig. 7.
Fig. 9. Detail images of the Tharsis Montes from Figs. 7 and 11: (a) Ascraeus Mons reg
Ascraeus Mons image (a) are: A, Ascraeus Mons upper shield; G, Ascraeus Mons caldera;
bright Ascraeus upper shield ?ows; K, Ascraeus Mons fan-shaped deposit; L, degraded-?
east ?ank of south rift apron; P, east Tharsis plains ?ows; Q, moderately bright south apr
north apron smooth unit. The labeled features for the Pavonis Mons image (b) are: B, nor
?ows; G, Pavonis Mons bright shield collar; H, Pavonis Mons fan-shaped deposit; J, Pav
caldera; M, Pavonis Mons eastern shield base; N, degraded-?ank region; P, Pavonis north
rift apron; S, dark feature from small island of smooth rilled terrain. The labeled features f
shaped deposit; C, Arsia north rift apron ?ows; D, east wing ?ows from Arsia north apr
caldera; W, dark Arsia Mons south shield ?ank; X, degraded-?ank region; Y, dark Arsia Mo
terrain. The rectangles in the Ascraeus (a) and Pavonis (b) images show the boundaries20 (2012) 990?1030South of the apron (1.8?4.8N lat.) is a N/S-aligned linear cluster
of low shields and vents (Hodges and Moore, 1994; Plescia, 2004;
Bleacher et al., 2007a,b), the north end of which can be seen at the
bottom of Fig. 10a. This region (7-C, 9a-C), dubbed the ??small vent
?eld?? by Bleacher et al. (2007a,b), is radar-bright. The brightness is
particularly strong over the southern half of the vent ?eld, indicat-
ing that the associated lava ?ows are roughest there. The radar
feature appears to blend in with bright vents and ?ows (7-D) due
east of Pavonis Mons (see Section 4.4), although the clustering seen
in orbiter imagery (Bleacher et al., 2007a) indicates that the Asc-
raeus ??small vent ?eld?? is a distinct unit.
Extending NNE of Ascraeus Mons (9a-S) is another lava fan or
apron (Hodges and Moore, 1994; Plescia, 2004; Murray et al.,
2010). Lavas ?owing down the western side of the apron axis ac-
count for the large radar-bright basin (7-E) embaying Ceraunius
Fossae. Lavas ?owing down the eastern side and off to the north-
east (7-F, 9a-F, 14-F) make up the brightest radar feature
; (b) MOLA vertical-roughness map, with lighter shading indicating higher km-scale
ion; (b) Pavonis Mons region; (c) Arsia Mons region. The labeled features for the
H, east edge (257.5E) of bright Ascraeus upper shield ?ows; J, west edge (254E) of
ank region; M, smooth, rilled core of Ascraeus south rift apron; N, lava channels on
on ?ows; R, bright south apron ?ows; S, Ascraeus north rift apron; T, dark terrain on
thwest-trending lava ?ows from the Tharsis ridge; D, Pavonis east lava apron plains
onis south volcanic ?eld; K, Pavonis Mons dark summit plateau; L, Pavonis Mons
rift apron; Q, northeast-trending lava ?ows on topographic bench; R, Pavonis south
or the Arsia Mons image (c) are: A, Arsia Mons bright shield collar; B, Arsia Mons fan-
on; E, bright Arsia south apron ?ows; J, Pavonis south volcanic ?eld; V, Arsia Mons
ns eastern shield ?anks; Z, Arsia south rift apron core, associated with smooth rilled
of the THEMIS south apron images in Fig. 10a and b, respectively.
rus 2J.K. Harmon et al. / Icar030 ? 8:1 dB
 
east of Ascraeus Mons (see Section 4.7). Some
smooth units similar to those on the south apron can be seen at
the base of the north apron (Murray et al., 2010), and these can
account for some restricted radar-dark areas (9a-T) around the
north base of the shield. However, the northern smooth unit is
smaller than the south-apron smooth unit and shows a much smal-
ler radar-dark feature.
4.4. Pavonis Mons
Pavonis Mons has a much different radar appearance from that
of Ascraeus Mons (Fig. 7 and Fig. 9b). In contrast to Ascraeus?s uni-
formly bright upper shield, Pavonis shows a radar-dark summit
plateau (9b-K) surrounded by a bright ring or collar (7-G, 9b-G).
A bright glint can also be seen from the radar-facing inner rim wall
of the caldera (9b-L). The radar-dark summit plateau appears
bright in MOC (Mars Orbiter Camera) optical images and dark in
THEMIS nighttime infrared images. Furthermore, the radar-bright
collar coincides with a prominent dark collar in MOC images and
largely overlaps a bright collar in THEMIS nighttime images. The
optical dark collar was already known from Viking imagery and
has been attributed to exhumation of a dark substrate by eolian
erosion of brighter surface dust (Lee et al., 1982; Toyota et al.,
2011). All of this is consistent with a summit plateau covered with
a radar-absorbing, low-thermal-inertia dust mantle and
surrounded by a collar of exposed radar-bright lavas. The bright
Fig. 10. THEMIS daytime infrared images of the south rift aprons of (a) Ascraeus Mons, an
radar images in Fig. 9a and b. Note the smooth rilled terrain on the west sides, the lava-
regions.20 (2012) 990?1030 1001collar extends all the way to the shield?s eastern base but only
about half to 2/3 of the way across the shield on the western side
(compare Figs. 7 and 8). A faint contact line (9b-M) can be seen at
the eastern base where the shield ?ows are embayed by plains
?ows (7-D, 9b-D) of the Pavonis east lava apron (see below).
Pyroclastic volcanism is a possible alternative to eolian dust for
explaining the radar-darkness of the summit plateau. The possibil-
ity of late-stage pyroclastic activity on Pavonis Mons was raised
with the discovery of putative cinder cones and explosive vents
in the summit region (Wood, 1979; Edgett, 1990; Zimbelman
and Edgett, 1992b; Crumpler et al., 1996; Head and Wilson,
1998). Airfall ash deposits from these or other pyroclastic vents
could act as a radar-absorbing mantle obscuring any underlying
effusive lavas on the summit plateau. This could explain the
contrast with the radar appearance of Ascraeus Mons, whose
radar-bright summit retains optically fresh evidence of effusive
volcanism and shows no evidence of pyroclastic activity (Mougi-
nis-Mark, 1981, 1982b; Zimbelman and Greeley, 1983; Zimbelman
and Edgett, 1992b).
Like Ascraeus Mons, Pavonis Mons shows a large radar-dark
lobe (7-H, 9b-H) extending well beyond its western ?ank. The
so-called ??Pavonis Mons fan-shaped deposit?? or FSD (Shean et al.,
2005) constitutes most of this dark lobe (compare Fig. 7 and
Fig. 8b). This deposit is analogous to, but much larger than, the
Ascraeus Mons FSD and has been assigned a similar glacial origin
by Shean et al. (2005). Although Shean et al. map the fan as
d (b) Pavonis Mons. The boundaries of these images are shown as rectangles on the
channeled terrain on the east sides, and the small-shield ?elds in the south-central
rus 21002 J.K. Harmon et al. / Icacomprising several distinct subunits, the radar feature appears uni-
formly dark. This darkness implies that the entire fan is covered
with ?ne-grained material and is devoid of surface or near-surface
rocks. The northwest boundary of the dark lobe corresponds very
closely with the outermost ridge (a ??drop moraine?? in Shean
et al.) of the FSD. Farther south, and due west of the Pavonis
summit, one sees the dark lobe extending a little way beyond the
outer fan ridge. The lower northwest slopes of the shield (9b-N),
between the bright collar and the inner boundary of the FSD, are
either radar-dark or show only faint brightness. This is similar to
the situation on the lower west ?ank of Ascraeus Mons, and may
be an indication of mantling by eolian or glacial deposits.
Like Ascraeus Mons, Pavonis Mons has aprons extending north-
east and southwest of the shield (Hodges and Moore, 1994; Plescia,
2004). The north apron (9b-P) appears in orbiter images as a rela-
tively small lobe of smooth terrain dissected by rilles. This region is
mostly radar-dark, except for a brighter patch on its east side. This
apron butts up against the bright ?ows (7-B, 9b-B) trending north-
west from the Pavonis east lava apron (see below), producing the
semicircular brightness contact seen near 249E, 4N. Much of
the south apron (9b-R) is also rather radar-dark and shows some
of the same smooth rilled terrain (Fig. 10b) seen on the Ascraeus
south apron and Pavonis north apron. This provides additional evi-
dence that the softer-appearing, rilled terrain making up the cores
of the Tharsis Montes rift aprons is also radar-smooth (i.e., non-
rocky). This could be consistent with the suggestion by Murray
et al. (2010) that mud ?ows might exist on other Tharsis Montes
rift aprons besides that of Ascraeus Mons. However, we also note
Fig. 11. Radar image of southern Tharsis, including Daedalia Planum. This is a montage of
imagery from November 14 and November 17, 2005, and the southern panel is from Nov
and February 14?17, 2012. The 2010?2012 imagery has also been used to patch N/S-ambi
The labeled features are: A, Arsia Mons bright shield collar; B, Arsia Mons fan-shaped de
bright Arsia south apron ?ows; F, Daedalia At6 ?ows; G, Daedalia At5 ?ows; H, bright
extension of bright Daedalia ?ows; W, bright south Daedalia ?ow ?nger; X, Zumba Cr
maximum is set to a maximum re?ectivity value of 0.393 (4.06 dB).20 (2012) 990?1030that Keszthelyi et al. (2008) identi?ed eolian dunes in HiRISE (High
Resolution Imaging Science Experiment on Mars Reconnaissance
Orbiter) images of the Pavonis south rift apron at 246.5E, 0.8S,
which raises the possibility that some of the radar darkness of
the Tharsis Montes rifts may be attributable to radar-absorbing
mantle deposits.
On the east side of the Pavonis south apron, radar-bright lava
channels trend downhill to the south and east (Fig. 10b) in a man-
ner similar to the Ascraeus south apron. Here and to the south is a
N/S-aligned linear cluster of small shields and vents similar to the
small-vent ?eld south of Ascraeus Mons (Plescia, 2004; Greeley
et al., 2005, 2006; Bleacher et al., 2007a,b). In their more recent pa-
pers, Bleacher et al. (2009, 2010) refer to this unit as the ??Pavonis
Mons South Volcanic Field.?? The entire vent ?eld, which extends
from 2S to 8S, shows up as a prominent radar-bright feature
(7-J, 9b-J) at the bottom of Figs. 7 and 9b; this feature (11-J) is also
identi?ed in the southern Tharsis image in Fig. 11. The northern
vents in this ?eld are visible at bottom center in the THEMIS image
in Fig. 10b; also visible in this image is an isolated island of smooth
rilled terrain, which shows up as a small dark spot (9b-S) in the ra-
dar imagery. East of the South Volcanic Field is a contiguous area of
moderately high radar brightness that extends to the barrier of
Noctis Fossae (7-K), at the west edge of the Syria Rise. This bright
region includes: a second cluster of small shields just west of the
Fossae at 250.5E, 5S (Hauber et al., 2009); a shield-free but
?ow-covered plains region separating this cluster from the South
Volcanic Field; and an E/W-trending arc of vents centered at
248E, 5S (Hauber et al., 2009). The radar imagery shows that this
three panels spliced at 3.65S and 20.8S latitudes. The northern panel is the sum of
ember 14, 2005. The central panel is the sum of imagery from January 26?29, 2010
guity foldover in regions southeast of Daedalia Planum and northwest of Arsia Mons.
posit; C, Arsia north rift apron ?ows; D, east wing ?ows from Arsia north apron; E,
Daedalia ?ow lobes; J, Pavonis south volcanic ?eld; K, Noctis Fossae; P, western
ater; Y, Pickering Crater; Z, southern terminal Daedalia ?ow lobe. The gray-scale
ap;
rus 2entire complex of small vents and lava plains south of Pavonis
Mons has moderate to high small-scale surface roughness, with
the roughest surfaces being in the South Volcanic Field proper.
Northeast of the South Volcanic Field is an even brighter radar
feature (7-D, 9b-D) that covers a large area east of Pavonis Mons.
Harmon et al. (1999) referred to this feature in the pre-upgrade
imagery as the ??Pavonis east lava apron,?? borrowing an informal
designation by Hodges and Moore (1994). Strictly speaking, the
Hodges?Moore apron refers to the near-shield lavas that ?ow
north around the eastern base of Pavonis Mons, wrap around the
older Pavonis northeast apron, and combine with ?ows from the
south end of the Ascraeus south apron to produce the prominent
bright ?ow lobe (7-B) extending northwest down the Tharsis west
Fig. 12. Orbiter-based context maps of southern Tharsis: (a) MOLA shaded-relief m
roughness. The map boundaries are the same as for the radar image in Fig. 11.
J.K. Harmon et al. / Icaslope (Section 4.6). However, the large radar-bright feature (7-D)
also encompasses a large cluster of small shields and vents be-
tween Pavonis Mons and Noctis Fossae (Hauber et al., 2009) as well
as the Ascraeus Mons small-vent ?eld just to its north (Section 4.3).
On the east side, the very bright Pavonis east apron feature blends
in with the moderately bright lavas ?owing down the Tharsis east
slope and spreading across the East Tharsis plains (Section 4.7).
Spacecraft imagery suggests that the small vents in the Pavonis
east apron (including the Ascraeus small-vent ?eld) are younger
structures superimposed on the long east-trending ?ows from
the Tharsis rift zone. Although the lava ?ows from these vents
are probably localized within the vent ?eld itself and do not con-
tribute to the East Tharsis plains, their high radar brightness indi-
cates they are among the roughest volcanic surfaces on Mars.
4.5. Arsia Mons and Daedalia Planum
Arsia Mons had the distinction of showing the brightest depo-
larized feature in the pre-upgrade imagery (Harmon et al., 1999).
Although the volcano also appeared very bright in the 2005
images, the quality of the imagery in that region was compromised
by proximity to the Doppler equator (see Fig. 2). Fortunately, in
January 2010 and February 2012 we were able to obtain additional
data from a better sub-Earth aspect, and with these data we have
been able to patch the corrupted imagery over the Arsia Mons
region. In Fig. 11 we show our ?nal composite image covering
southern Tharsis and Daedalia Planum. Here the 2010/2012 imag-
ery over Arsia Mons and northern Daedalia Planum is combinedwith 2005 imagery of southern Daedalia and the region north of
Arsia Mons. The 2010/2012 imagery has also been used to patch
N/S-ambiguity foldover in the southeast and northwest corners
of Fig. 11. In Fig. 12 we show MOLA shaded relief and vertical
roughness maps of the same region shown in Fig. 11. In Fig. 9c
we show a detail image of Arsia Mons and environs.
Arsia Mons shows a very bright collar (11-A, 9c-A) covering
much of the upper shield and wrapping around the darker caldera
(9c-V). The bright north slope shows an average r030 of 5.6 dB and
some extremely bright patches with r030 ? 4:5 dB (Table 2) This
supports our pre-upgrade ?nding (Harmon et al., 1999) that this
is the brightest depolarized radar feature on Mars. The high radar
brightness indicates extremely rough surfaces for the lava ?ows
(b) MOLA vertical-roughness map, with lighter shading indicating higher km-scale
20 (2012) 990?1030 1003in this region. The bright collar covers most of the summit plateau
and the steep portion of the shield ?anks down to the 10-km alti-
tude contour. The location of the radar feature is consistent with
spacecraft imagery showing long, narrow lava ?ows emanating
from the summit region (Mouginis-Mark, 1982b; Hodges and
Moore, 1994). The bright collar lies within the late-Hesperian to
early-Amazonian lava ?ow unit AHt3 in the geologic map of Scott
and Tanaka (1986) and also within the early-Amazonian ?ow unit
At4 in the geologic map of Scott and Zimbelman (1995). However,
we also ?nd radar-dark portions of the shield edi?ce lying within
the AHt3 and At4 units. These include a dark segment (9c-W) south
of the caldera, the graben zone on the west and northwest rim of
the caldera, a ??degraded ?ank?? region (9c-X) on the lower western
shield (see below), and a dark collar (9c-Y) around the lower east-
ern and southeastern shield ?anks. While it is possible that lava
?ows are either smoother-textured or absent in these regions,
the fact that some of this radar-dark terrain is also dark in THEMIS
nighttime infrared (IR) images suggests the possible in?uence of
pyroclastic or eolian mantling deposits with low thermal inertia.
In fact, Mouginis-Mark (2002) has identi?ed possible ash deposits
at several locations on the Arsia Mons shield, including the south-
ern rift zone and the western rim graben. Nevertheless, the very
existence of the radar-bright shield collar strongly suggests that
most of the upper shield is free of thick mantling deposits, whether
pyroclastic or eolian.
The Arsia Mons caldera (9c-V) is darker than the shield collar
but still bright (8.3 dB) in an absolute sense. This contrasts with
Ascraeus Mons, with its distinctly dark caldera. Mouginis-Mark
rus 2(1982b) noted the contrast between the Ascraeus Mons caldera,
with its featureless ?oor, and the Arsia Mons caldera, with its vis-
ible post-collapse lava ?ows. Also, dome-like features have long
been known to exist in the Arsia caldera ?oor (Carr et al., 1977;
Zimbelman and Edgett, 1992a,b; Crumpler et al., 1996), and these
were later con?rmed as lava shields (Mouginis-Mark and Christen-
sen, 2005). More recently, HiRISE imagery of the caldera (Keszth-
elyi et al., 2008) has identi?ed low shields and lava ?ows and
placed an upper limit of one meter for the depth of any mantling
deposits.
Like Ascraeus and Pavonis Montes, Arsia Mons has an associated
northwest-trending fan-shaped deposit (FSD) that a recent study
has interpreted as glacial in origin (Shean et al., 2007). This unit,
which stands out in the MOLA roughness map (Fig. 12b), is the
largest of the three FSDs associated with the Tharsis Montes. Shean
et al. (2007) map the Arsia FSD extending northwest as far as
229.5E. Our imagery shows this region (11-B, 9c-B) to be radar-
dark, which is similar to the other two FSDs and consistent with
a rock-poor, radar-absorbing mantle. Unlike the Pavonis FSD, the
Arsia FSD does not have its distal boundary clearly de?ned by
contrasting emergent bright lava ?ows. Instead, the radar-dark ter-
rain continues farther northwest, beyond the FSD ??terminal mor-
aine?? and into the ??Stealth?? radar feature (see Section 4.6). Also,
the western edge of the bright shield collar stops short of the
mapped inner (eastern) boundary of the FSD. This is similar to
the situation on the western ?anks of both Ascraeus and Pavonis
Montes and, in fact, the radar-dark lower western slopes (9c-X)
of the Arsia shield bear some resemblance to the ??degraded ?ank
material?? (Kadish et al., 2008) on the west side of Ascraeus Mons
(Section 4.3). Hence, the radar-darkness of this region may be
attributable to dust deposited at the shield base by westerlies or
to the ash deposits proposed by Mouginis-Mark (2002).
Also like Ascraeus and Pavonis Montes, Arsia Mons has lava fans
or aprons on its NNE and SSW ?anks (Fig. 12). These aprons are
thought to derive from eruption centers on the volcano?s ?anks,
rather than from the summit, and appear to represent the most re-
cent major eruptions in the region (Crumpler and Auberle, 1978;
Hodges and Moore, 1994; Scott and Zimbelman, 1995; Plescia,
2004). Both northern and southern aprons show up as radar-bright
features. The north apron radar feature (11-C, 9c-C) fans out in
western and eastern wings corresponding to lavas ?owing down
either side of the Tharsis ridge crest. The western wing has a ?n-
gered appearance suggesting individual ?ow units, whereas the
eastern wing looks more amorphous and merges with the Pavonis
Mons South Volcanic Field (11-J, 9c-J). Just south of the Pavonis
South Volcanic Field is another bright radar feature (11-D, 9c-D)
that spacecraft imagery suggests is where lava ?ows from the east
wing of the Arsia north apron have wrapped around to the south
and become partially obscured by the younger Pavonis South Vol-
canic Field to the north. This feature stands out as a topographi-
cally smooth unit in Fig. 12b.
Arsia?s south apron shows a semi-circular radar-bright feature
(11-E, 9c-E) that is particularly prominent on its western side. This
feature, which was also prominent in the pre-upgrade imagery
(Harmon et al., 1999), corresponds to well-de?ned lava ?ows in
spacecraft images. Inside this feature is a radar-darker region
(9c-Z) that lies on the core axis of the rift apron itself. The fact that
this same region shows muted texture in orbiter images suggests
that there may have been some terrain softening by smoother sur-
?cial material deposited near the rift axis after the apron lavas
were emplaced. Some similarity can be seen with muted terrains
in the cores of the Ascraeus and Pavonis south rift aprons, although
the Arsia feature is not as radar-dark as those on the other two
1004 J.K. Harmon et al. / Icavolcanoes.
Extending nearly 2000 km south of Arsia Mons are radar-bright
lava ?ows that cover much of Daedalia Planum and dominate thesouthern half of Fig. 11. These ?ows are all presumed to come from
Arsia Mons or its southern rift zone, although tracing the exact
sources has been dif?cult owing to obscuration by more recent
south apron deposits. All studies suggest that the Daedalia ?ows
were emplaced over several episodes spanning the mid-Hesperian
to mid-Amazonian epochs, and spatial variations in radar bright-
ness over the region support an episodic emplacement scenario.
The moderately bright ?ows (11-F) extending south of the south-
ern rift apron to about 17S latitude correspond approximately to
the upper-Amazonian At6 ?ows in the geologic map of Scott and
Tanaka (1986), while the brighter ?ows (11-G) extending farther
south to about 24S correspond approximately to Scott and
Tanaka?s older (mid-Amazonian) At5 ?ow unit. Hence, the radar
imagery suggests that the At5 ?ows have rougher surface texture
than the younger At6 ?ows. Bright Daedalia ?ows extending far
to the west (11-P) appear to be extensions of the At5 ?ows in the
radar imagery, despite their designation as older At4 and AHt3
?ows by Scott and Tanaka (1986). Extending southwest of the main
body of bright At5 ?ows (11-G) are moderately bright ?ows whose
southern terminus (11-Z) embays the cratered highlands and
whose boundaries mostly lie within the older AHt3, Ht2, and Ht1
map units of Scott and Tanaka (1986). Interestingly, other mapped
AHt3 and Ht2 regions located due south and east of the main At5
bright region (11-G,H) appear completely dark in the radar imag-
ery (Fig. 11).
Recently, Giacomini et al. (2009, 2010) mapped Daedalia ?ows
according to a more complex scheme involving 13 separate age
units (designated D1?D13). Several prominent radar-bright ?ow
lobes extending a few degrees south of 20S latitude in the south-
ern At5 region (11-H) correspond to the youngest (D13) mapped
?ows in Giacomini et al. (2009, 2010). These ?ows are closely
spaced and have lobate ?ow fronts with irregular and rough sur-
faces (Giacomini et al., 2009). In this same region Crown et al.
(2009, 2010) noted optically bright and dark ?ows with channels,
levees, spillovers, and breakouts. Based on HiRISE imagery, Keszth-
elyi et al. (2008) describe these radar-bright ?ows as having an
intensely knobby or ridged texture with surfaces similar to terres-
trial a?a, rubbly pahoehoe, or blocky ?ows with brecciated tops.
They note, however, that the scale of the brecciation is larger than
that seen on terrestrial a?a ?ows. All of this is consistent with the
high radar brightness of the region. Ramsey and Crown (2010)
claim a ??moderate?? dust cover for these ?ows, while Keszthelyi
et al. (2008) point out that the ?ow tops are obscured by a ?ne-
grained mantle but that boulders do crop out. This is consistent
with the sort of thin (less than meter-depth) mantle that should
be penetrable by the radar wave.
The radar imagery of the south end of Daedalia (Fig. 11 and
Fig. 13c) shows interesting detail. Some of this detail can also be
discerned in THEMIS daytime (Fig. 13a) and nighttime (Fig. 13b)
infrared images. The most prominent single ?ow feature in this re-
gion is a long radar-bright ?nger (11-W, 13-W) that extends south-
west more than 700 km and that Giacomini et al. (2009) consider
to be the best-de?ned lava ?ow in southern Daedalia. Giacomini
et al. map this ?ow as unit D6, which places it intermediate on
their age scale. They suggest that this might be an ??in?ated?? ?ow,
which could account for its extreme length. More recently, Graff
and Zimbelman (2012) have found evidence for in?ated pahoehoe
?ows in radar-bright terrain at the northern end (226.2E, 23.9S)
of this same long ?ow.
There is a close correspondence between some small radar-dark
patches clustered around 227E, 29S (Fig. 13c) and dark (low ther-
mal inertia) patches in Fig. 13b. Several of these dark patches are
inselbergs of old terra (probably including remnant crater rims)
20 (2012) 990?1030that stick up out of the surrounding lava plains (Fig. 13a). Others
can be traced to isolated ?at plains regions where converging lava
?ows have failed to completely merge.
rus 2J.K. Harmon et al. / IcaIn the middle of these small dark patches is another dark spot
(11-X, 13c-X) that corresponds to the 3.3-km-diameter impact cra-
ter Zumba (13a,b-X). This is one of a handful of very young martian
impact craters that show distinct rays in IR images (Tornabene
et al., 2006). Most of these are, like Zumba, located in young lava
plains. Although Zumba?s long rays are clearly seen in the THEMIS
images (Fig. 13a and b), they do not show up in the radar images.
The small dark feature that we do see is apparently where ?ne
ejecta deposits in or near the crater have blanketed the rough lava
?ow surfaces (see Section 8). Extended radar-bright rays, associ-
ated with rocky ejecta or secondary cratering, are not normally
seen on craters this small on other planets or the Moon and would
be dif?cult to discern against the bright lava plains in any case. An-
other ??dark-halo crater?? (see Section 8) shows a prominent dark
spot at 230E, 23S (Fig. 11). This 7-km-diameter crater, unlike
Zumba, does not show rays in THEMIS IR images.
Another notable feature is the crescent-shaped bright lobe
(11-Y, 13c-Y) at 227E, 33S. This shows where bright lava ?ows
of the main Daedalia feature have breached the northwest rim of
Pickering Crater (13a,b-Y) and ?lled the western portion of crater
?oor. The observed radar lobe agrees precisely with the mapped
extent of invading ?ood lavas in Scott and Tanaka (1986) and Cap-
rarelli and Leitch (2009). Caprarelli and Leitch have the eastern
?oor of Pickering covered by older ??platy?? ?ows (mapped as Ht1
by Scott and Tanaka). Caprarelli and Leitch also suggest that the
younger radar-bright ?ows ?ooding the western ?oor are associ-
ciated with lava ?ows extending over similarly long distances to
the west and east of the Tharsis Montes divide. The morphologyFig. 13. THEMIS (a) daytime and (b) nighttime infrared images of the south-Daedalia region, along with the corresponding radar image (c) from Fig. 11. The
labeled features are: W, long Daedalia ?ow ?nger; X, Zumba Crater; Y, Pickering
Crater; Z, southern terminal Daedalia ?ow lobe.of the west-slope ?ows is particularly complex, in part because
of obscuration by late-stage mantling deposits.
Nowhere are the effects of mantling and resurfacing more
apparent than over an extended region west and northwest of Ar-
sia Mons (Fig. 11). This region is mostly radar-dark, despite the fact
that Tharsis ?ows have been mapped all the way from the Arsia
fan-shaped deposit (FSD) to Mangala Valles (Scott and Tanaka,
1986). That the Arsia FSD (11-B) should be radar dark is not
surprising, given the darkness of the Pavonis and Ascraeus FSDs.
It is plausible that a ?ne glacial deposit would hide underlying
rough-surfaced lava ?ows from the radar but still allow large-scale
?ow relief features to remain visible through the deposit in orbiter
imagery. One would expect the well-de?ned At5 ?ows that are
emergent from (and surround) the FSD to be radar-bright and thus
to delineate the distal boundary of the deposit (as with the Pavonis
FSD). This is not the case, as the ?ows remain mostly radar-dark
west and northwest of the FSD. The most likely explanation is that
the radar waves are being absorbed by the mantling deposits of the
so-called ??Stealth?? feature, an equatorial radar-dark region that
may be a volcanic ash deposit (Muhleman et al., 1991; Butler,
1994; Edgett et al., 1997; Edgett, 1997; Ivanov et al., 1998). Stealth
overlaps the outer portion of the Arsia FSD, the Arsia ?ows farther
west and northwest, and much of the Medusae Fossae Formation
north of these Arsia ?ows. A tongue of Stealth also extends east be-
tween Arsia Mons and the minor shields Biblis Patera (7-L) and
Ulysses Patera (7-M) (Muhleman et al., 1991; Edgett et al., 1997).
This would explain the radar-darkness of this region (7-N), despiteated with Amazonian volcanism, whereas Scott and Tanaka map
these as part of the same late-Hesperian Ht2 unit covering much
of south Daedalia.
The bright lobe (11-Z, 13c-Z) at the southern terminus (13a,b-Z)
of Daedalia Planum shows where the Daedalia ?ows have spilled
through one ?nal restriction before embaying the cratered high-
lands of Terra Sirenum (see also Fig. 12). (The small dark spot at
center right on this lobe corresponds to a dark-halo crater.) Crown
and Berman (2012) and Crown et al. (2012) have mapped this
terminal lobe as mid-Amazonian sheet ?ows. This lobe and the
western Pickering feature both show the same moderate radar
brightness as most of the main south Daedalia feature. Since these
regions are all mapped as Hesperian-age unit Ht2 in Scott and
Tanaka (1986), there seems to be disagreement between the earlier
and the more recent geologic maps on the age of the south Daeda-
lia ?ows.
Finally, in an unrelated note, we point to the region of Syria Pla-
num between 255?265E and 10?20S, where spacecraft imagery
has identi?ed clusters of small shields, vents, and lava ?ows
(Hodges and Moore, 1994; Baptista et al., 2008; Hauber et al.,
2009; Richardson et al., 2010). Our imagery (Fig. 11) shows this re-
gion to be entirely radar-dark, unlike the low-shield ?elds near the
Tharsis Montes. This might be consistent with a suggestion by Bap-
tista et al. (2008) that Syria Planum experienced a different and
very speci?c style of volcanism, which is plausible given its greater
(Hesperian) age and isolation from the main axis of Tharsis
Montes.
4.6. West-slope ?ows and minor shields; Ceraunius Fossae
In the previous section we showed where radar-bright lavas
from Arsia Mons have ?owed south nearly 2000 km across Daeda-
lia Planum. In the next two sections we discuss radar features asso-
20 (2012) 990?1030 1005coverage by At5 ?ows, as well as the fact that the bright ?ows sur-
rounding the two paterae appear as if disconnected from their
source.
rus 2Biblis Patera and Ulysses Patera stand out in the radar imagery
because of the stark contrast between their radar-dark shields
(7-L,M) and the surrounding bright ?ows. These paterae are older
than the major Tharsis shields (Hodges and Moore, 1994; Plescia,
1994), so their darkness may simply re?ect a long-term accumula-
tion of eolian dust or a blanketing by a pre-Stealth pyroclastic air-
fall. Ulysses Patera may even have produced its own pyroclastics
(Plescia, 1994), although its neighbor Biblis Patera probably did
not. Both shields appear to have been built up from low-viscosity
lavas (Plescia, 1994), so a smoother lava surface texture could also
have contributed to their radar-darkness. The caldera ?oor of Ulys-
ses Patera shows post-collapse ?ows (Plescia, 1994), and these
probably account for the faint central bright features seen in the
radar imagery. Biblis Patera shows no such caldera ?ows (Plescia,
1994), which is consistent with the nearly uniform darkness of
the Biblis radar feature.
The lower ?anks of both paterae are embayed and surrounded
by younger Tharsis ?ows (unit At5 of Scott and Tanaka (1986)),
whose radar brightness contrasts with and sharply delineates the
patera edges. Orbiter images and geologic maps show that these
younger ?ows originate from the Tharsis Montes (Scott and Tana-
ka, 1981b; Scott et al., 1981b). However, the only radar-bright
connection between the main Tharsis ridge and the bright cir-
cum-patera ?ows is a narrow, faint neck extending northwest from
the Pavonis south apron and volcanic ?eld. Well-de?ned lava ?ows
extending all the way to the paterae from the Arsia north apron are
clearly visible in orbiter imagery but are largely radar-dark in the
middle region (7-N) between Arsia Mons and the paterae. Appar-
ently these and some of the Pavonis south apron ?ows are being
obscured by radar-absorbing mantle deposits associated with the
eastern tongue of Stealth. Supporting this is the fact that the south-
ern edge of the bright circum-patera radar feature shows a fuzzi-
ness suggesting a gradual thinning of the mantling deposits at
the outer edge of Stealth. Stealth or Stealth-like deposits may also
account for the fading out of the bright At5 ?ows west of Biblis
Patera. Evidence for this comes from the identi?cation of an eolian
dune ?eld at 234E, 5N, just northwest of Biblis Patera (Edgett,
1997). To the north and west of this point is the Medusae Fossae
Formation, which is also intrinsically radar-dark (see Section 6.2).
The bright At5 ?ows continue north of the two paterae until
they run up against Ulysses Fossae (7-P), a fractured plateau of old-
er crustal material (unit Hf of Scott and Tanaka (1986)). Ulysses
Fossae is mostly radar-dark, making it one of several examples of
??fractured terrain?? (others being Noctis Fossae, Ceraunius Fossae,
and Fortuna Fossae) that are topographically rough (Fig. 8b) but
smooth or non-rocky at small scales. Ulysses Fossae is embayed
by Tharsis ?ows on the south and east sides and by Olympus Mons
basal plains and aureole units on the west side.
To the northeast of Ulysses Patera, a broad neck of At5 ?ows
(7-Q) extends northeast between Ulysses Fossae and the Pavonis
FSD. These lavas, which apparently derive from both the Arsia
north apron and Pavonis south apron, ?ow downhill off the topo-
graphic ??bench?? (Fig. 8a) on which Biblis and Ulysses Paterae sit.
This is the brightest radar feature northwest of Pavonis Mons,
which indicates that these distal ?ows either have developed very
rough surfaces by this point or are less heavily mantled (by Stealth
deposits or other pyroclastics?) than are the ?ow surfaces to the
south and west.
The ??bench?? ?ows (7-Q) terminate at a radar-dark line that cor-
responds to a graben connecting the apex of the Pavonis FSD (7-H)
with Ulysses Fossae. Beyond this point, orbiter imagery and the
MOLA maps (Fig. 8) suggest that the lava ?ows derive from more
northerly sites on the Tharsis ridge, such as Pavonis Mons itself
1006 J.K. Harmon et al. / Icaor the Pavonis east lava apron. In fact, the earlier geologic maps
of Scott and Tanaka (1981b) and Scott et al. (1981b) identi?ed a
separate tongue of young Pavonis Mons ?ows (designated unitstm6 and Apm1 in the two maps, respectively) emerging from the
apex of the even younger Pavonis FSD. Although the later map of
Scott and Tanaka (1986) does not distinguish these from the At5
?ows, recent orbiter imagery and the MOLA maps (Fig. 8) show
these emergent Pavonis ?ows clearly and identify them as the
most widespread ?ow unit in northwest Tharsis. These ?ows stand
out in the MOLA roughness map (Fig. 8b) as a dark (topographically
smooth) anvil-shaped feature that spreads out north from a narrow
base at the north end of the Pavonis FSD. A corresponding anvil-
shaped radar-bright feature (7-R) approximately follows the path
of the MOLA smooth unit. This is one example (others being in
Amazonis and Elysium) of a lava plains unit that appears extremely
smooth topographically in MOLA vertical roughness maps but
shows high small-scale roughness in radar imagery, a phenomenon
that was ?rst recognized by Fuller and Head (2002).
On the east side of the ??anvil?? feature (7-R), and beyond the
northeast edge of the Pavonis FSD (7-H), is a radar-dark wedge
(7-S) that is roughly centered on Poynting Crater (whose rim is
faintly discernable in the radar image). Although Poynting Crater
has not been the subject of any published studies, inspection of
THEMIS daytime imagery clearly shows the crater postdating some
of the lava ?ows in the region. This is especially clear on the west
side, where ?ows appear as if emerging from under the crater?s
western ejecta lobes. However, the THEMIS images also show
younger lava ?ows from the southeast (our radar-bright 7-B ?ows)
lapping up against the crater?s eastern rim and ejecta. This suggests
that the Poynting impact occurred in the middle of a period of epi-
sodic lava ?ow emplacement over this region. The radar-dark
wedge (7-S) would then correspond to a region where the crater
and its lobate ejecta were superimposed on older (pre-impact)
?ows from the Pavonis region and also presented a topographic
barrier to later (post-impact) ?ows. The radar darkness of the
Poynting ejecta lobes indicates that the near-surface ejecta blanket
is made up of ?ne-grained (rockless) material.
Curving around the east side of the ??Poynting wedge?? is a nar-
row, radar-bright feature (7-T) that originates from the north edge
of the Pavonis east lava apron or from the Ascraeus south apron.
This radar feature is associated with a 700-km-long lava ?ow
(Zimbelman, 1984; Hanley and Zimbelman, 1995) that was appar-
ently emplaced in a deep channel at high effusion rates (Garry
et al., 2007) and that radar sounding by the SHARAD (Shallow Ra-
dar) instrument indicates is a high-permittivity material consistent
with a dense basalt (Carter et al., 2009). The radar brightness indi-
cates that this long ?ow, which has a fresh appearance in orbiter
images, is rough-surfaced at small scales and not heavily mantled.
Older ?ows between this bright ?ow and the Ascraeus FSD (9a-K)
appear radar-darker, which is consistent with their muted optical
appearance in orbiter images (suggesting mantling or some terrain
softening effect associated with the Ascraeus south apron).
Several prominent dark-halo craters (7-U) have been identi?ed
in northwest Tharsis (see Section 8). Two of these are located on
the western and northern edges of the ??anvil?? ?ows: a 23-km-
diameter crater at 240.9E, 9.6N; and a 24-km-diameter crater
at 249.1E, 18.0N. These craters show radar-dark ejecta haloes
extending out 100 km or more. These haloes take signi?cant bites
out of the southwest and northeast edges of the radar-bright fea-
ture, which accounts for most of the difference in the shapes of
the ??anvil?? in the radar (Fig. 7) and MOLA vertical-roughness
(Fig. 8b) maps. The northeastern crater is close to a lava ?ow
(250.9E, 16.8N) that has been inferred to be high-permittivity
(dense) basalt based on radar sounding with SHARAD (Carter
et al., 2009). Noting the low radar brightness of this region in our
preliminary report on the 2005 Arecibo imagery (Harmon and No-
20 (2012) 990?1030lan, 2007), Carter et al. suggested the presence of a dust cover that
absorbs the S-band radar signal. It now appears that the absorbing
mantle proposed by Carter et al. is real, but is in the form of distal
rus 2ejecta from the dark-halo crater at 249.1E, 18.0N rather than eo-
lian dust. Finally, we note two other dark-halo craters in the region.
One of these is a relatively small feature from a 10-km-diameter
crater at 243.8E, 15.6N. The second is a large (125-km) dark spot
that can be traced to a 20-km-diameter crater at 260.1E, 19.3N,
on the east side of the large lobe of bright ?ows (7-E) extending
north from Ascraeus Mons.
Near the northern terminus of the bright ??anvil?? ?ows (7-R) is
the minor shield volcano Jovis Tholus (7-V). The Jovis shield is ra-
dar dark, like Biblis and Ulysses Paterae. This is consistent with the
shield?s muted texture and dearth of small craters, which suggests
a covering mantle such as an ash deposit (Plescia, 1994). Immedi-
ately ESE of the Jovis shield is a small, low shield volcano that
resembles some of the low shields in the aprons of the major Thar-
sis volcanoes (Hodges and Moore, 1994; Plescia, 1994, 2004). This
small shield, which is apparently younger than Jovis Tholus, is
bright in the radar image. Several other radar-dark patches in the
Jovis Tholus region can be traced to impact craters or to islands
of fractured terra.
The Ceraunius Fossae Formation (or CFF) is a prominent radar-
bright feature north of the Tharsis Montes and east of Olympus
Mons. The CFF appears as a gullwing-shaped feature (7-W) in the
radar image. Scott and Tanaka (1986) identify the formation as
?ssure-fed lava ?ows of late-Hesperian to early-Amazonian age.
Subsequent studies assigned a more recent (mid-Amazonian) age
that has some of the eruptions being contemporaneous with
late-stage volcanism from the Tharsis Montes (Tanaka, 1990; Hod-
ges and Moore, 1994). The radar-brightness indicates that the CFF
lavas are rough and cannot have accumulated a thick dust mantle.
In addition to ?ssures, numerous small shield volcanoes have been
identi?ed in the central portion and east wing of the CFF (Hodges
and Moore, 1994; Wong et al., 2001). These could be sources of
some of the bright CFF ?ows. Scott and Tanaka (1986) map the east
wing of the CFF as extending beyond the northern and eastern lim-
its of the radar-bright feature. However, the revised mapping by
Tanaka (1990) redesignates this region as late-Hesperian and
shrinks the NE boundaries of the CFF into much better agreement
with the boundaries of the radar feature (see also Borraccini et al.,
2005, 2006). The map of Scott and Tanaka (1986) also shows the
CFF extending below 20N latitude into the region east of the Jovis
Tholus shield (7-V). Our image shows this region to be dark and
thus not covered by the same rough-textured ?ows as farther
north, although part of this could be due to obscuring ejecta from
a crater at 246E, 18.3N. Geologic maps have the west wing of the
CFF terminating at, and presumably overlapped by, younger Olym-
pus plains (Aop) lavas ?owing northwest (Scott et al., 1981a;
Mouginis-Mark, 1982a; Mouginis-Mark et al., 1982a; Scott and Ta-
naka, 1986; Hodges and Moore, 1994). The mapped contact (Scott
and Tanaka, 1986) between these two ?ow units coincides with the
V-notch feature (7-X) in the radar image. Note that there is no ra-
dar brightness contrast at the contact. Also, MOLA maps (Fig. 8)
and THEMIS images show CFF ?ows rounding the notch and ?ow-
ing off to the northwest. It seems likely, then, that the mapped Aop
?ows extending north of 20N toward Acheron Fossae actually in-
clude a substantial admixture of CFF lavas. The region around
240E, 20N is, in fact, a complex convergence zone for lava ?ows
from three sources: Olympus plains, the CFF, and the Tharsis Mon-
tes (Mouginis-Mark, 1982a; Mouginis-Mark et al., 1982a).
The eponymous Ceraunius Fossae (or CF) comprises three lobes
of fractured terra of late-Noachian to early-Hesperian age that are
characterized by dense clusterings of N/S-aligned graben (Scott
and Tanaka, 1986; Tanaka, 1990; Borraccini et al., 2005, 2006).
The CF rises above, and is embayed, by the younger CFF ?ows. Un-
J.K. Harmon et al. / Icalike the CFF, the CF (7-Y) is almost entirely radar-dark, indicating a
surface that, while topographically rough on kilometer scales
(Fig. 8b), is smooth at small (decimeter) scales. In this respect itresembles other features such as Ulysses Fossae and Fortuna Fos-
sae (Section 4.7), which are also radar-dark islands of fractured ter-
ra surrounded by bright lava ?ows. The only bright features in the
CF are several N/S-trending bright streaks. These appear to be
places where CFF ?ows (or, in one or two cases, Tharsis Montes
?ows) have invaded some of the low-lying graben ?oors of the
CF. Also of interest are two bright spots near 250E, 30N (7-Z),
which Scott and Tanaka (1986) map as an isolated Noachian ??sub-
dued crater unit?? (unit Npl2) within the northern lobe of the CF.
The revised mapping by Tanaka (1990) give these as outlier CFF
deposits, which is consistent with their radar brightness.
4.7. East Tharsis ?ows and minor shields
Like its west slope, Tharsis?s eastern ?ank shows radar-bright
lavas ?owing downslope over distances exceeding 1000 km. These
?ows dominate the radar image shown in Fig. 14. In Fig. 15 we
show the MOLA shaded relief and vertical roughness maps over
the same region.
A comparison of radar and orbiter imagery indicates that most
of the radar-bright northeast plains are ?ows originating from Asc-
raeus and Pavonis Montes and their aprons. The drape-like lava
channels due south of Ascraeus Mons merge into striated ?ow
features trending toward the northeast (Fig. 14). Some of these
bright ?ows wrap around the north side of radar-dark Fortuna Fos-
sae (14-A) before heading due east. Other ?ows wrap around the
south side of Fortuna Fossae and then turn north to join the north-
ern ?ows. Fortuna Fossae itself, a large heart-shaped radar-dark
feature (14-A), is an island of fractured terra (unit Hf of Scott and
Tanaka (1986)) similar to the Ulysses Fossae feature northwest of
Tharsis (see previous section). The bright ?ows south of Fortuna
Fossae show a sharp southern edge where they shoulder up against
the Syria Rise (Fig. 15a). Farther west, this bright southern edge
shows spikes or feathering (14-B) where the Tharsis lavas have in-
vaded the lower reaches of some of the graben of Noctis Fossae
(14-K) (see Fig. 15).
East of Fortuna Fossae the bright southern ?ows skirt around a
large radar-dark region (14-C) bordered on the east by Echus Chas-
ma (14-D). This dark area was mapped as a northern extension of
the Syria Planum Formation (unit Hsu) by Scott and Tanaka (1986)
and includes, as its eastern half, a region recently dubbed the
??Echus Plateau?? (Zealey, 2007). Spacecraft imagery shows Amazo-
nian-age ?ows converging northeast of Fortuna Fossae, continuing
due east, and then spilling into the northern reaches of Echus Chas-
ma. Nearly all of the Echus Chasma ?oor is radar-bright, which is
consistent with near-complete covering by rough lavas. This agrees
with the geologic map of Scott and Tanaka (1986), which has the
At5 ?ows extending south and embaying most of this leg-shaped
chasma. A more recent study (Chapman et al., 2010) suggests a
more complex volcanic sequence for Echus Chasma, in which older
??platy-ridged?? ?ows (their unit Apf), possibly originating in the
chasma itself, are overlain or embayed by younger At5 ?ows.
Platy-ridged ?ows have also been identi?ed in the chasma by Man-
gold et al. (2010). These ?ows are similar to radar-bright platy
?ows seen in Cerberus and Marte Vallis (Section 5.3) and southern
Amazonis (Section 6.2). Chapman et al. (2010) mapped Apf out-
crops in the ankle, foot, and lower calf of Echus, and portions of
these are among the radar-brightest regions of the chasma. These
authors suggest that the ??foot?? region is covered by a thin lacus-
trine material (unit Apec) similar to the sedimentary deposit (unit
Avf) mapped in this same region by Scott and Tanaka (1986). How-
ever, the fact that the forefoot region contains the brightest radar
feature in Echus Chasma indicates that any overlying sediments
20 (2012) 990?1030 1007must be very thin.
Surrounding Echus Chasma is a radar-dark halo that is even dar-
ker than the rest of Echus Plateau to the west and Lunae Planum to
rus 21008 J.K. Harmon et al. / Icathe east. Silvestro and Pondrelli (2007) found the edges of the
plateaus surrounding Echus Chasma to be covered with layered
deposits that are dissected by water-cut sapping valleys on the pla-
teau west of the chasma ??foot.?? Mangold et al. (2008) agreed on the
origin of the sapping valleys and suggested that they represent ero-
sion of a weak surface layer of poorly-indurated or ?ne-grained
material. The dark halo surrounding the chasma would be consis-
tent with radar absorption in such a ?ne-grained surface layer.
Fig. 14. Radar image of the East Tharsis plains. This is the sum of imagery from October
Fortuna Fossae; B, lava ?ows invading Noctis Fossae graben; C, Echus Plateau; D, Echus
Ascraeus Mons; G, Kasei Valles ?ows; H, Tharsis Tholus; J, transition region between olde
Tholus caldera; N, Uranius Patera caldera; P, dark-halo crater; Q, dark feature encompass
from the north rims of Valles Marineris channels can be seen at bottom right. Also sho
obtained on February 15?24, 2012. The gray-scale maximum is set to a re?ectivity of 0
Fig. 15. Orbiter-based context maps of the East Tharsis plains: (a) MOLA shaded-relief m
scale roughness. The map boundaries are the same as for the radar image in Fig. 14.20 (2012) 990?1030Zealey (2007) identi?ed glacial outwash features on the eastern
edge of the Echus Plateau, so it is possible that the western part
of the dark halo represents radar absorption in ponded sediments
of glacial till. Zealy also speculated that the entire Echus Plateau
was a remnant glacier or ice sheet around which the Amazonian
lavas were forced to ?ow. This might help to explain why the lava
?ows turn so sharply north after rounding the south side of Fort-
una Fossae. A glacial mantle deposit could also explain why some
14, October 15, November 17 and November 23, 2005. The labeled features are: A,
Chasma; E, northern limit of bright plains ?ows; F, northeast-trending ?ows from
r and younger plains ?ows; K, Noctis Fossae; L, Uranius Tholus caldera; M, Ceraunius
ing a dune ?eld on its east side and dark-halo crater on its west side. Layover glints
wn (insert) on the same scale is an image centered on Uranius Patera, from data
.25 (6.02 dB), or 70% of the maximum re?ectivity measured over this region.
ap; (b) MOLA vertical-roughness map, with lighter shading indicating higher km-
lava ?ows on the west side of the dark plateau (14-C) have a some-
what muted optical appearance and are radar-dark, despite show-
ing obvious ?ow fronts in orbiter images. It should be noted,
however, that Mangold et al. (2008) have reported ?nding no
evidence of glacial features on the Echus Plateau.
North of Echus Chasma the radar-bright terrain traces where
the At5 ?ows extend as a peninsula (284E, 16N) into the radar-
dark head-region of Kasei Valles. In this region, and farther north
and west, the radar features conform more closely to the recent
geologic map of Chapman et al. (2010) than to the older map of
Scott and Tanaka (1986). The Chapman et al. map shows the At5
?ows extending up to 20N latitude, where they share a boundary
with the older (Hesperian) Ht4 ?ows. This boundary is clearly seen
in orbiter imagery (Fig. 16a) and agrees precisely with the northern
edge (14-E) of the radar-bright plains. The Chapman et al. map also
shows a small peninsula of At5 extending into the northern branch
of Kasei Vallis (283E, 21N) that coincides with a moderately
radar-bright lobe (14-G). By contrast, the Scott and Tanaka map
shows this region to be covered by lower-Amazonian At4 ?ows that
extend north of 20N. However, while the Scott and Tanaka map
does not explain the radar-bright boundary at 20N, its distinction
between At4 and At5 units probably re?ects a real north-to-south
decrease in ?ow ages. From Fig. 14 it can be seen that the ?ows
wrapping around the north end of Fortuna Fossae are brighter than
the ?ows farther north, with a distinct boundary visible about
halfway between the Fossae (14-A) and Tharsis Tholus (14-H). A
portion of this transition region (14-J) is shown in the spacecraft
image in Fig. 16b. Here the younger, brighter southern ?ows can
be seen overlapping the older, darker ?ows to the north. It should
be noted that the radar brightness transition does not correspond
well with Scott and Tanaka?s At4/At5 boundary, which runs just
north of Tharsis Tholus. Keszthelyi et al. (2000) noted that some
of the older ?ows east of Tharsis Tholus showed a ridged-
and-grooved morphology seen in younger ?ood lavas in Elysium
and Amazonis. They noted that these ?ows had a degraded appear-
ance with a signi?cant eolian cover, which could account for their
relative darkness compared to the younger southern ?ows. Farther
west, between Tharsis Tholus and Ascraeus Mons, the northern
?ows are more radar-bright and show more structure, which
suggests that they (like the ?ows surrounding Fortuna Fossae)
are rougher or less heavily mantled than the ?ows east of Tharsis
Tholus. Some of these ?ows trend northeast out of the Ascraeus
Mons south rift apron, while the brightest (14-F) ?ow is oriented
ENE out of the eastern side of the Ascraeus north rift apron.
The region covered in Fig. 14 includes several minor shields
similar to those on the Tharsis west slope (previous section). One
of these, Tharsis Tholus (14-H), is located in the north-central part
of the east Tharsis volcanic plains. The entire heart-shaped shield is
radar-dark except for a bright glint off the major fault that inter-
sects the west rim of the caldera. Viking and MOC images show a
smooth shield devoid of individual ?ow features (Hodges and
Moore, 1994; Plescia, 2003a), while some THEMIS and Viking
images show linear features suggestive of heavily mantled lava
?ows (Plescia, 2003a). A heavy mantle is consistent with the low
thermal inertia of the shield (Plescia, 2003a). A thick mantle was
also proposed by Platz et al. (2009, 2011), who attribute the layer
J.K. Harmon et al. / Icarus 220 (2012) 990?1030 1009Fig. 16. THEMIS daytime infrared images of (a) northern edge of bright East Tharsis pla
region. The labeled features E, J, and Q correspond to the same locations on the radar imins ?ows, and (b) central portion of East Tharsis plains showing overlapping-?ow
age in Fig. 14.
(2009, 2011) have the shield surrounded and partly buried by
younger and less heavily mantled lavas, which is consistent with
rus 2the radar imagery. The radar image (Fig. 14) shows a lobe of bright
lavas coming from the WNW and wrapping around most of the
shield base, with darker (presumably older) lavas surrounding
the remainder. Platz et al. (2009, 2011) identi?ed a ponding of
these lavas on the shield?s west ?ank, which can account for a
west-to-east elevation drop.
North of Tharsis is the minor-shield cluster comprising Uranius
Patera, Ceraunius Tholus, and Uranius Tholus. The smallest of these
shields, Uranius Tholus (14-L), is radar-dark. This is consistent with
the lack of well-de?ned lava ?ows and indications that much of the
shield is mantled, possibly with pyroclastic deposits (Plescia,
2000). Ceraunius Tholus (14-M) is also mostly lacking in recogniz-
able lava ?ows and may also be pyroclastically mantled (Plescia,
2000). The Ceraunius shield is radar-dark except for some modest
brightness on its southeast ?ank. Since this portion of the Cerau-
nius shield shows a high concentration of ?uvial troughs, it is pos-
sible that the faint radar feature is associated with surface
roughening through erosion or deposition. The west ?ank of the
shield is completely radar-dark. This side of the shield has a
smooth appearance and is probably the youngest surface on Cerau-
nius (Crumpler et al., 1996; Plescia, 2000). Crumpler et al. suggest
the west ?ank is pyroclastically mantled, which would be consis-
tent with its radar darkness. Plescia, on the other hand, identi?es
effusive lava ?ows in this region which, if real, must be suf?ciently
mantled to be radar-dark or of a low-viscosity type giving low sur-
face roughness. The largest of the three shields, Uranius Patera (14-
N), shows the strongest radar feature. This is a bright patch that
starts at the south caldera rim and connects with bright off-shield
?ows south of the shield. The rest of the shield is radar-dark, ex-
cept for a faint ring around the caldera rim. These Uranius Patera
features show up even better in the inset ?gure at upper right in
Fig. 14, which shows a radar image from February 2012 (when
the volcano was imaged at a much smaller incidence angle). Ura-
nius Patera shows more effusive ?ow features and less pyroclastic
mantling than its two neighboring shields (Plescia, 2000), which
could explain some of the radar brightness. Why its southwest
?ank ?ows should be rougher than elsewhere on the shield is
not clear, however, nor are there any distinguishing morphologic
features on this side of the volcano (except for a narrow trough).
Several dark-halo craters (Section 8) have been identi?ed in the
East Tharsis plains. These include a 5-km-diameter crater at
257.4E, 1N; an 8-km crater at 267.8E, 15.8N; a 10-km crater
(14-P) at 262.6E, 9.6N; a 10-km crater on the lower northeast
?ank of the Ascraeus Mons shield at 259E, 12N; and a 14-km cra-
ter at 272.3E, 17.9N. The last of these lies on the west side of a
150-km-diameter, quasi-circular dark region (14-Q) that also
encompasses what appears to be a large dune ?eld complex (see
Fig. 16a). Radar absorption in this dune ?eld apparently dominates
the eastern part of the dark feature (14-Q) and can explain its sharp
eastern edge. This unusual, possibly pyroclastic or sedimentary,
deposit must be an old and indurated unit, since it is interleaved
with lava ?ows and pocked with craters. It stands out in the MOLA
roughness map (Fig. 15b) as topographically rougher than the sur-
rounding lava ?ows. Further study should be made of this interest-
ing but undocumented feature.
5. Elysiumto dust and/or ash. Such a thick mantling is consistent with the ra-
dar darkness of the shield. Both Plescia (2003a) and Platz et al.
1010 J.K. Harmon et al. / IcaThe Elysium region is the second great Mars volcanic province,
after Tharsis, and the second largest expanse of radar-bright terrain
in the depolarized imagery. That Elysium is the source of strongdepolarized radar backscatter was ?rst apparent from Arecibo
CW observations (Harmon and Ostro, 1985). The pre-upgrade Are-
cibo imagery (Harmon et al., 1992b, 1999) provided the ?rst com-
plete radar maps of the region and identi?ed the major sources of
strong radar backscatter. These images showed Elysium Mons to be
a major source of radar-bright lavas but also showed that the dom-
inant radar-bright feature in Elysium is the Cerberus Basin and the
associated Marte Valles out?ow channel. Around the same time it
was becoming clear from spacecraft imagery that Cerberus is the
site of recent ?uvio-volcanic activity and that the associated lava
?ows include some of the youngest surfaces on the planet. This
has made it one of the most intensively studied regions of Mars
in recent years.
A 2005 image encompassing most of the radar-bright Elysium
features is shown in Fig. 17. Corresponding MOLA shaded relief
and vertical roughness maps are provided for context in Fig. 18.
Since the sub-Earth track (11S lat.) was entirely in the radar-dark
southern cratered highlands, there is no discernable ambiguity fol-
dover from southern features in our Elysium radar image (Fig. 17).
The major radar-bright features in the image include ElysiumMons
and its associated ?ows in the northwest part of the image, the
great Cerberus Basin spanning most of the lower central region,
and the narrow Marte Vallis feature arising in Cerberus and trend-
ing to the northeast to a debouchment in northern Amazonis. Also
of interest is an extended region of moderately enhanced bright-
ness in western Elysium, beyond the western edge of Fig. 17.
5.1. Elysium volcanoes
The three volcanoes on the Elysium Rise are Elysium Mons,
Albor Tholus, and Hecates Tholus (Fig. 19). Elysium Mons (19-A),
the largest and marginally youngest of the three volcanoes, is the
source of most of the extended lava ?ows in the region (Mougi-
nis-Mark et al., 1984; Greeley and Guest, 1987; Hodges and Moore,
1994). It is also the only Elysium volcano to show an extensive ra-
dar-bright feature. Nearly the entire high-elevation (above
+2.0 km) portion of the shield is radar-bright. The brightest region
is offset somewhat to the east of the caldera, as are the topographic
contours (see Plescia, 2004). The caldera itself (19-A) shows up as a
subtle radar feature but does not differ markedly in brightness
from the rest of the shield. Although the brightest region on Ely-
sium Mons is mostly con?ned above the +2.0-km elevation con-
tour, some narrow bright ?ngers can be seen ?owing farther
downslope, mostly on the eastern side. This is of interest, as some
of these extended Elysium Mons ?ows have been deduced to have
a rheology similar to that of Tharsis shield ?ows (Hiesinger et al.,
2009). Also, a major bright lobe can be seen extending southeast
to Albor Tholus (19-B). All of these bright features show up as
prominent ?ows in orbiter images. Plescia (2007) pointed out an
asymmetry in the Elysium Mons summit geometry that suggests
that west-?ank ?ows came from an older, higher, possibly smaller
caldera, whereas lavas ?lling the present caldera would ?ow down
the east and southeast ?ank. If correct, this could explain the ob-
served eastward asymmetry in the radar-bright ?ows.
Extending southwest of the main radar-bright shield of Ely-
sium Mons, in a region of circumferential graben, is a fainter
bright collar or apron (19-C) that corresponds to highly subdued
?ows in orbiter images. Mouginis-Mark et al. (1984) claimed this
feature to be a lobe of young ?ood lavas, which would be con-
sistent with its smoothness in the MOLA roughness map
(Fig. 18b). Zimbelman and McBride (1989) identi?ed this as a
low-thermal-inertia region whose subdued appearance and low
crater counts suggest ?ow obscuration by pyroclastic deposits.
20 (2012) 990?1030The radar imagery does not favor a heavily mantled surface,
although a thin mantle could account for the lower brightness
compared to the main shield ?ows. Separating this collar feature
Fig
rou
Fig
0.2
rus 2J.K. Harmon et al. / Icafrom the bright Elysium Mons ?ows north of Albor Tholus is a
dark patch (19-D) on the south base of Elysium Mons where
the extended ejecta blanket of a dark-halo crater (146E, 22N)
covers the underlying ?ows (see Section 8). This is identi?ed
as a fresh (c3-class) superimposed crater in the geologic map
of Scott and Allingham (1976).
The region (19-E) to the north (lat. >28N) of Elysium Mons and
west of Hecates Tholus is radar-dark, despite the presence of long,
well-preserved lava ?ows (Mouginis-Mark and Yoshioka, 1998).
These ?ows either have smooth surfaces or are heavily mantled.
Extending well to the ESE of Elysium Mons is a moderately
radar-bright band (19-F) that corresponds to prominent lava ?ows
in orbiter imagery. These merge with Grjot? Valles ?ows in north
ern Cerberus (Section 5.3) to form the long ??ESE band?? featur
(Fig. 17) identi?ed in the pre-upgrade imagery (Harmon et a
1999). Fainter bright ?ows (19-G) can be seen running paralle
to, but south of, the ESE band. These terminate at the barrie
presented by the Tartarus Montes, which are radar-dark. Farthe
southwest are other bright ?ows northeast and northwest of Cer
berus Palus. All of these bright ?ows in the quadrant southeast o
Elysium Mons fall into units Ael1 and AHEe in the geologic map
of Greeley and Guest (1987) and Tanaka et al. (2005), respectivel
. 18. Orbiter-based context maps of the Elysium region: (a) MOLA shaded-relief map; (b) MOLA vertical-roughness map, with lighter shading indicating higher km-scale
ghness. The map boundaries are the same as for the radar image in Fig. 17.
. 17. Radar image of the Elysium region. This is the sum of imagery from October 3 and October 6, 2005. The gray-scale maximum is set to the maximum re?ectivity of
76 (5.59 dB) measured over this region.-
e
l.,
l
r
r
-
f20 (2012) 990?1030 1011s
y,
mated thickness of 100 m. Hauber et al. (2005) also proposed pyro-
clastic mantling of the northwest ?ank from explosive eruptions at
a separate caldera at the northwest base of the Hecates shield. All
of this west-side mantling would explain the low THEMIS night-
time temperatures and radar-darkness of this side of the volcano.
Plescia (2007) has suggested that the entire volcano is mantled
and that the mantling on the northwest ?ank is suf?ciently heavy
to obscure any volcanic features. However, he interprets the man-
tling to be eolian rather than pyroclastic. If the entire volcano is
mantled, as Plescia suggests, then the mantling on the east side
must be suf?ciently thin (at most a meter or so) to allow some ra-
dar penetration to underlying rough ?ows. Mouginis-Mark and
Christensen (2005) argued that the prominent valley networks
on Hecates are consistent with ?uvial erosion of surface ash depos-
its, while Plescia (2007) argues that some lava ?ow features are
observable and that the valley networks are the result of volcanic
or ?uvial erosion of a non-pyroclastic surface. The fact that we
see any radar features at all suggests that moderately rough lava
?ows exist on the Hecates shield and that some of these are either
exposed or only lightly mantled on the eastern ?ank.
5.2. Western Elysium
Elysium Mons and the western Cerberus volcanics (Section 5.3)
mark the western limit of strong radar backscatter from Elysium.
However, some weak to moderate depolarized features can be seen
in the imagery farther west (Fig. 20). Although we refer to this re-
gion as Western Elysium, much of this is now considered part of
rus 220 (2012) 990?1030Fig. 19. Radar image (detail of Fig. 17) of the Elysium Mons region showing the
main Elysium shield volcanoes. The labeled features are: A, Elysium Mons caldera;
B, Albor Tholus caldera; C, Elysium Mons collar feature; D, dark-halo crater; E,
radar-dark Elysium Mons northern ?ows; F, east?southeast-trending Elysium Mons
?ows (??ESE band??); G, additional Elysium Mons ?ows; H, Hecates Tholus caldera.
The gray-scale maximum is set to the maximum re?ectivity of 0.268 (5.72 dB)
measured over this region.
1012 J.K. Harmon et al. / Icaand as such are presumed to have erupted from Elysium Mons or
its associated vents.
Albor Tholus (19-B) is smaller and lower than Elysium Mons
and more closely resembles the Tharsis minor shields in its radar
appearance. The entire summit plateau of the shield is radar-dark
(except for a bright ring from the interior caldera rim). The summit
plateau has a smooth appearance in orbiter imagery that suggests a
mantling deposit, possibly pyroclastic (Hodges and Moore, 1994;
Crumpler et al., 1996). Supporting this are THEMIS nighttime
images showing a dark (low-thermal-inertia) collar covering the
upper shield. It appears likely, then, that the radar-dark summit re-
gion has a thick, radar-absorbing mantle of eolian or pyroclastic
origin. Plescia (2007) suggests that a mantling deposit covers the
entire ?ank of the volcano. However, radar-bright terrain on the
southern, western, and eastern parts of the lower shield indicate
areas where any mantling must be suf?ciently thin to allow bright
?ows to show through. These same regions also show some lava
?ow features in orbiter optical images (Crumpler et al., 1996)
and have higher nighttime temperatures in THEMIS images. Being
located on the shield slopes, these radar-bright ?ows must origi-
nate from Albor Tholus itself. These ?ows appear to merge with
Elysium Mons ?ows embaying the ?at plains around the shield
base.
As with Albor Tholus, the radar appearance of Hecates Tholus
(19-H) seems to be largely determined by mantling effects. A mod-
erately bright radar feature is seen on Hecates, but this is mostly
con?ned to the east and southeast portions of the shield, with
some spillover onto the adjacent plains. Nearly the entire western
half of the shield is radar-dark, except for a bit of brightness on the
SSW edge near the shield base. Based on low crater counts and
subdued features, Mouginis-Mark et al. (1982b, 1984) proposed
the existence of air-fall ash deposits on the western side of the
shield, with the deposits just west of the caldera having an esti-Utopia (Tanaka et al., 2003a,b, 2005).
The Western Elysium bright features roughly comprise two
WNW-trending lobes that, west of 130E longitude, are separated
by a radar-dark region between 20 and 30N latitude. This dark
middle region (20-A) covers Hephaestus Rupes and is mapped as
sediments of the Vastitas Borealis interior unit (ABvi) in the map
of Tanaka et al. (2005). The radar-bright features north of the dark
Fig. 20. Radar image of western Elysium. This is from observations made on
October 3, 2005. The image has been smoothed over 3  3-pixel blocks. The labeled
features are: A, Vastitas Borealis-unit sediments; B, D, topographically smooth
Elysium Rise ?ows; C, debris ?ows or debris-laden lahars; E, Viking Lander 2 site; F,
Mie Crater; G, eroded or sapped terrain; H, proposed MER-2003 ??Elysium?? landing
site; J, bright Elysium Rise ?ows; K, western tip (??beak??) of Cerberus Fossae ?ows; L,
western lava path; M, western lava ponding basin. The gray-scale maximum is set
to 0.145 (8.40 dB), or 50% of the maximum re?ectivity measured over this region.
lobe all appear to derive from fossae radiating from the west ?ank 5.3. Cerberus and Marte Valles
Unlike Tharsis, the bulk of Elysium?s radar-bright features are
not associated with the large shield volcanoes and associated vents
but rather with plains volcanism in the Cerberus Basin to the
southeast. The ?rst full recognition of Cerberus as a major volcanic
region came with a paper by Plescia (1990), and this received some
support from the pre-upgrade Arecibo imagery (Harmon et al.,
1992b, 1999). Among the factors that make this region so interest-
ing are an apparent association between volcanic and ?uvial pro-
cesses and evidence that Cerberus volcanism includes some of
the youngest major resurfacing events on Mars.
The major Cerberus radar features fall roughly into ?ve regions:
(1) western Cerberus, including Athabasca Valles and its catch-ba-
sin, Cerberus Palus; (2) northern Cerberus, including Grjot? and
Rahway Valles and Tartarus Montes; (3) south Cerberus, where
bright lavas embay the southern cratered highlands; (4) central
Cerberus, a complex region of plains ?ows and small shields and
the source region for Marte Valles; and (5) Marte Valles itself,
which trends northeast out of eastern Cerberus and debouches in
northern Amazonis Planitia. We will discuss each of these regions
in turn.
Athabasca Valles is one of several radar-bright channels in Ely-
sium that appear to have been cut by ?ssure-fed water and/or lava
J.K. Harmon et al. / Icarus 220 (2012) 990?1030 1013of the Elysium rise or Elysium Mons. The radar-bright band (20-B)
immediately north of the dark lobe (and stretching roughly from
130E, 22N to 110E, 34N) is mapped as lava ?ows of the Elysium
Rise unit (AHEe) in Tanaka et al. (2005). Russell and Head (2003)
map it as lava ?ows of their ??smooth lobate?? unit, where the
??smoothness?? refers to low topographic roughness in MOLA rough-
ness maps. The bright banding (20-C) seen farther north is mapped
as debris ?ows of the Granicus and Tinjar Valles (unit AEtb) by
Tanaka et al. (2005) and as debris-laden lahars by Russell and Head
(2003). North of these are still more bright features, including a
patch at 130E, 37N (20-D) that Tanaka et al. map as Elysium rise
unit lavas (AHEe) and that Russell and Head map as ??smooth lobate
unit?? lavas. This brightness fades north of 40N latitude. Included
in this darker region is the Viking Lander 2 site at 133.5E, 48N
(20-E). It is interesting to note that Mie Crater (140E, 48.5N)
(20-F), which was once considered a possible source of the rock
?eld at the nearby VL-2 site, does not show a distinctive radar-
bright feature. Hence, there is no evidence from the radar imagery
that the rocks at VL-2 are Mie ejecta or that the rock cover at VL-2
is signi?cantly enhanced over that elsewhere in this region of Uto-
pia Planitia.
The bright features south (lat. <20N) of the dark lobe (20-A)
have a different provenance from those to the north. Much of the
bright band (20-G) that extends from 132E to 100E falls within
the erosional unit HBu2 of Tanaka et al. (2005), with some overlap
onto their older HBu1 unit. However, there appears to be an even
closer association with the boundary-plains unit Hb1b of Tanaka
et al. (2003a). This region shows stepped scarps and depressions
that may have been formed by collapse processes involving lique-
faction, so the radar brightness may indicate surface roughness
related to these. The older Hb1a unit to the south is radar-dark
and hence smoother at small (decimeter) scales. The proposed
MER-2003 ??Elysium?? landing site (124E, 12N) (20-H) of Tanaka
et al. (2003a) falls on the northern edge of this Hb1a unit, near
the boundary with Hb1b. In their study of this potential landing
site, Tanaka et al. suggested that this region might contain depos-
ited detritus from erosion of Noachian highland rocks. The radar
darkness of this region con?rms their evaluation of this as a safe
site for a lander.
Geologic mapping (Tanaka et al., 2005) suggests that the radar-
bright terrain farther east, between 130?140E and 4?20N is more
likely to be associated with volcanic roughness. Most of this bright-
ness falls within the Tanaka et al. Elysium Rise unit (AHEe) and
includes areas with mapped ?ow fronts. The southern bright patch
centered near 136E, 7N (20-J) connects with the much brighter
western tip (20-K) of the Late-Amazonian Cerberus Fossae (AEc3)
?ows via a narrow bright channel (20-L). This gives the impression
that the bright patch (20-J) west of 140E may actually be Cerberus
?ows that debouched from this channel. However, the geologic
map of Tanaka et al. (2005) indicates that the Cerberus ?ows from
the channel terminate in a small radar-bright pond near 139E, 5N
(20-M) and that the more extended bright patch is associated with
older, unrelated Elysium Rise ?ows. One interesting question re-
mains as to why the region (20-N) from 140 to 145E and south
of Elysium Mons is so radar-dark. One partial explanation is that
AHEe ?ows are overlain by ejecta from Eddie Crater (142E,
12.5N) and another larger crater (141E, 9N) to its south. This
does not, however, explain the radar darkness farther north and
southwest of Elysium Mons. Apparently, the Elysium Rise ?ows
in this region are either smooth-surfaced or more heavily mantled
than in other areas.?ows and then ?lled with ?ood lavas (Burr et al., 2002a,b; Berman
and Hartmann, 2002; Plescia, 2003b; Lanagan, 2004; Jaeger et al.,
2007, 2010; Hurwitz and Head, 2012). A blowup of the radar imag-
ery of this region is shown in Fig. 21. The source of both water and
lava is a ?ssure (21-A) at the western end of the Cerberus Fossae
fracture system. Both orbiter and radar imagery show the Athaba-
sca lavas ?owing southwest in two main channels: a narrow north-
ern channel (21-B), for which some papers exclusively reserve the
name Athabasca Valles; and a wider southern channel (21-C),
whose upstream region shows a radar-dark wedge (21-D) where
the ?ow is split by Persbo Crater and whose eastern bank is con-
?ned by a radar-dark ridge (21-W).
At their southwest terminus, the Athabasca ?ows debouch into
the Cerberus Palus Basin (Jaeger et al., 2007, 2010), which is also
called the Athabasca Basin by Vaucher et al. (2009a) and the Wes-
Fig. 21. Radar image (detail of Fig. 17) of western Cerberus. The labeled features
are: A, location of Cerberus Fossae source ?ssure for Athabasca Valles ?ows; B,
Athabasca Valles north channel; C, Athabasca Valles south channel; D, Persbo
Crater; E, Cerberus Palus; F, proposed MER-2003 ??Athabasca?? landing site; G,
spillway connecting Cerberus Palus with southern ??beak?? basin; H, parasitic ?ows
from Athabasca Valles breach; J, parasitic ??sub-basin 4?? feature; K, southern ??beak??
basin; L, western lava path; M, western lava ponding basin; N, Zephyria Planum
section of Medusae Fossae Formation; P, Aeolis Planum section of Medusae Fossae
Formation; Q, Lethe Basin; R, Elysium Rise ?ows; W, topographic divide between
Athabasca Valles and central Cerberus. The gray-scale maximum is set to the
maximum re?ectivity of 0.225 (6.48 dB) measured over this region.
rus 2tern Elysium Basin by Balme et al. (2011). The brightest radar
re?ections in western Cerberus are found in eastern and northern
Cerberus Palus (21-E). It is here that some of the most spectacular
examples of ??platy-ridged?? ?ows are found (Keszthelyi et al., 2000,
2008; Lanagan, 2004). This supports indications from other
radar-bright features (e.g., Echus Chasma, Marte Vallis, Amazonis)
that platy-ridged lava ?ows tend to have high small-scale surface
roughness.
It has been proposed that the platy-ridged surfaces in western
Cerberus are not lava ?ows but rather the remnants of fractured
ice-rich materials (Murray et al., 2005; Kossacki et al., 2006; Balme
et al., 2011). Such non-volcanic theories were challenged by Jaeger
et al. (2007), based partly on the high decimeter-scale roughness
indicated by the Earth-based radar observations. We do not
dismiss the possibility that ice-based processes could have pro-
duced rough-textured surfaces giving strong depolarized radar
backscatter. On the other hand, the fact that the western Cerberus
radar features are similar to bright features from known or
suspected volcanic ?ow surfaces (platy-ridged or otherwise) else-
where on Mars suggests that the volcanic hypothesis is the simpler
and more plausible explanation for the surfaces in Athabasca Val-
les and Cerberus Palus. Since ?ood waters are thought by some to
have played an important role in the resurfacing of the Athabasca
region (Plescia, 1990; Burr et al., 2002a,b; Balme et al., 2011), it is
possible that some of the radar brightness in the region is not asso-
ciated with primary volcanic surfaces but rather with aqueous
erosion and deposition. While this might account for some of the
more modest radar brightness enhancements in the region, it prob-
ably cannot explain extremely bright features such as those seen in
Cerberus Palus.
It is apparent that radar brightness decreases as one goes up-
stream through Athabasca Valles toward the source ?ssure, sug-
gesting that the ?ow surfaces become smoother near the source.
This effect is particularly noticeable within about 50 km of the
source fossa (21-A), where the radar feature becomes very faint.
This may be related to the fact that the upstream portions of the
?ow appear to have ??de?ated?? as the lavas drained downstream
and ponded, leaving very thin ?ows near the source vent (Jaeger
et al., 2007, 2010). It is possible that the upstream ?ows were
not subjected to the same levels of surface brecciation as the pond-
ed platy-ridged ?ows in the Cerberus Palus debouchment zone.
Another possibility is that the region near the source vents was
subject to some late aqueous erosion (Burr et al., 2002a,b) or to a
late resurfacing by inherently smooth (radar-dark) materials such
as muds, pyroclastics, or smooth-textured lavas (Burr et al.,
2002a). It should be noted that MOLA topographic roughness
(Fig. 18b) shows the reverse behavior, with higher roughness near
the Athabasca source vent and smoother topography downstream
and in Cerberus Palus. This may re?ect some topographic smooth-
ing associated with downstream ?ow thickening.
An indication of the keen interest in Athabasca Valles is the fact
that it was once considered one of six prime candidates for a MER
landing site (Christensen et al., 2005). The landing ellipse was cen-
tered at 155E, 9N (21-F), which placed the western half of the el-
lipse in older Elysium Rise ?ows (AHEe of Tanaka et al., 2005) and
the eastern half in an upstream portion of Athabasca Valles. Our
imagery indicates that there would have been a potential landing
hazard from surface roughness anywhere in this ellipse, especially
on the east (Athabasca) side, although the hazard would not have
been as extreme as in downstream portions of Athabasca Valles
or in Cerberus Palus.
The Athabasca and Cerberus Palus ?ows appear to have spread
into outlying basins via a series of spillways. Jaeger et al. (2010)
1014 J.K. Harmon et al. / Icahave some of the Cerberus Palus lavas exiting the basin through
a spillway (21-G) in the southwest and spreading out into a second
basin. A beak-shaped radar-bright feature (21-K) agrees closelywith the mapped extent of this smaller basin and is consistent with
the debouching of rough-surfaced Cerberus Palus lavas from the
spillway (21-G). The maps of Vaucher et al. (2009a) and Jaeger
et al. (2010) show a narrow lava channel extending from the south-
ern tip of this feature to a small pond basin at 139E, 5N; this
channel and basin are referred to, respectively, as the ??western
lava path?? and ??western lava basin?? in Vaucher et al. (2009a). Both
the lava path (21-L, 20-L) and ponding basin (21-M, 20-M) show up
as bright features in the radar imagery (see also Section 5.2). The
maps of Jaeger et al. (2010) and Balme et al. (2011) also show a
spillway that breached the eastern side of the Athabasca Valles
channel and allowed lavas and/or waters to ?ow toward the south-
east. This parasitic ?ow shows up as a radar-bright feature (21-H)
that comprises ??sub-basin 1?? and ??sub-basin 3?? of Balme et al.
(2011). Although Balme et al. interpreted the features in these
sub-basins to be purely ?uvial, the high radar brightness indicates
?ooding by lavas diverted from the southern branch of Athabasca
Valles. Our imagery does not show the radar feature (21-H)
extending anywhere near as far to the south as in the Jeager
et al. and Balme et al. maps, except for a small bright patch (21-
J) in Balme?s ??sub-basin 4??. This suggests that distal parts of the
parasitic ?ow developed a smoother surface or that the ?ow was
episodic. Between the parasitic basin (21-H) and eastern Cerberus
Palus (21-E) is a region of moderate radar brightness (21-Q) whose
southern end includes Lethe Basin, which is connected via Lethe
Valles spillway to Cerberus Palus (21-E) and via another spillway
to the parasitic basins (21-H) (Jaeger et al., 2010; Keszthelyi
et al., 2010; Balme et al., 2011). This moderate brightness also ex-
tends north of Lethe Basin and across the plains separating the
northern and southern branches of Athabasca Valles. Since this
falls within the Hbu2 unit of Tanaka et al. (2005), it is possible that
some of the radar roughness is erosional/depositional rather than
volcanic. The radar feature continues well north of Athabasca
Valles into a region (21-R) that Tanaka et al. mapped as Elysium
Rise unit AHEe, so the brightness here is probably backscatter off
moderately rough lavas from the Elysium Mons region.
The region immediately south and west of Cerberus Palus is ex-
tremely radar-dark. Particularly dark is the region (21-N) due
south of the eastern end of Cerberus Palus. This area was mapped
geologically as a western portion of the Medusae Fossae Formation
or MFF (Greeley and Guest, 1987; Tanaka et al., 2005), and now has
the local designation Zephyria Planum (Kerber and Head, 2010;
Kerber et al., 2011). The radar darkness is consistent with absorp-
tion in the friable pyroclastic deposits generally considered to con-
stitute the MFF (Scott and Tanaka, 1982; Tanaka et al., 2005;
Kerber and Head, 2010; Kerber et al., 2011). The re?ectivity of
20.3 dB is consistent with zero detectable echo, when ambiguity
foldover from the southern cratered highlands is accounted for.
The dark region (21-P) south of the ??western lava path?? has been
mapped as the Aeolus Planum section of the MFF (Tanaka et al.,
2005; Kerber and Head, 2010).
Another radar-bright region that appears to have been resur-
faced by ?ssure-fed ?ows of both water and lava is the area of
northern Cerberus above 7N latitude (Fig. 22). Much of the radar
brightness in this region correlates with the AEc2 (??Cerberus Fossae
2??) unit mapped by Tanaka et al. (2005). This geologic unit is
mapped as older than the main central-Cerberus ?ows connecting
with Marte Valles and includes the Grjot? Valles and Rahway Val-
les complexes. The dominant source ?ssure for much of the north-
ern Cerberus resurfacing is a northern branch of Cerberus Fossae
(22-A) that extends from 165E, 15N to 160.5E, 16.5N and has
been identi?ed as the source of the Grjot? Valles ?oods (Burr
et al., 2002b; Plescia, 2003b; Burr and Parker, 2006). The Grjot?
20 (2012) 990?1030Valles water/lava ?ows (22-B) have a mapped northern boundary
at 16?17N latitude. At this boundary (22-C) these ?ows contact,
and presumably overlap, the southeast limit of the radar-bright
rus 2Elysium Mons lavas (22-F), making only a subtle brightness con-
Fig. 22. Radar image (detail of Fig. 17) of northern Cerberus. The labeled features
are: A, Cerberus Fossae source ?ssure for Grjot? Valles ?ows; B, Grjot? Valles ?ows;
C, contact region between Elysium Mons ?ows and Grjot? Valles ?ows; D, Grjot?
Valles ?ood troughs; E, Grjot? Valles bypass channel through Tartarus Montes; F,
Elysium Mons ?ows; G, Rahway Valles drainage network; H, eastern end of
Cerberus Fossae fracture system; J, bright lavas presumably erupted from eastern
Cerberus Fossae ?ssures; K and L, large craters breached and ?ooded with bright
lavas; M, side channel connecting Grjot? and Rahway Valles; P, dark-halo crater; Q,
??Mesogaea Crater??; R, Tartarus Montes; S, lava channel from Grjot? Valles; W,
topographic divide between Athabasca Valles and central Cerberus; X, Zunil Crater;
Y, Cerberus Fossae. The gray-scale maximum is set to the maximum re?ectivity of
0.270 (5.69 dB) measured over this region.
J.K. Harmon et al. / Icatrast. Together, these Elysium Mons ?ows (22-F, 19-F) and the
Grjot? Valles ?ows (22-B) comprise the so-called radar-bright
??ESE band?? (Fig. 17) noted in the pre-upgrade imagery (Harmon
et al., 1999). The radar-bright Grjot? Valles feature (22-B)
corresponds to a region where orbiter images show polygonal
and platy-ridged features like those seen in Athabasca Valles (Kes-
zthelyi et al., 2010), all of which is consistent with lava ?ows. The
Grjot? Valles ?oodwaters are deduced to have ?owed east from the
fossa and then south in two main channels or troughs through the
Tartarus Montes and into the northern part of the central Cerberus
basin (Berman and Hartmann, 2002; Burr and Parker, 2006). Ples-
cia (2003b) argued for a volcanic ?ll in these troughs based on their
smooth texture and the presence of ?ow fronts. Our imagery shows
these troughs (22-D) as radar-bright extensions of the Grjot? Valles
feature, which supports Plescia?s argument for a volcanic ?ll. In
fact, the entire Grjot? Valles ?ood tract as mapped by Burr and Par-
ker (2006) has a radar-bright appearance consistent with lava ?ll.
One can even see a narrow bright feature (22-E) where Burr and
Parker map a small bypass channel through the northern reaches
of Tartarus Montes.
The other major ?uvio-volcanic complex in northern Cerberus is
located southeast of the Grjot? Valles complex and at the northeast
corner of the main Cerberus basin. This complex includes Rahway
Valles (22-G), a dendritic valley network of a type not seen else-
where in Cerberus. Waters erupting from the eastern end of
Cerberus Fossae (174?170E, 6?7N) (22-H) may have fed the Rah-
way Valles drainage network, but it is also likely that some of the
valleys were fed by local sources in the headwaters region (22-G)
itself (Plescia, 2003b). There are radar-bright features covering
much of this region, and these conform closely to the eastern
boundaries of the same AEc2 volcanic unit (Tanaka et al., 2005)
making up the Grjot? Valles complex. This suggests that lavas cov-
ered Rahway Valles subsequent to the aqueous ?ooding, as has also
been proposed by Plescia (2003b). Our imagery suggests that someor all of these lavas erupted from the eastern end of Cerberus Fos-
sae (22-H). One of the brighter radar features (22-J) in this region is
located just north of the fossae at 170?173E, 7N, and it is likely
that some of these lavas continued ?owing northeast into the
pre-existing Rahway Valles drainage network (22-G). Some of
these ?ows also breached and ?ooded two large craters at
172.5E, 9.3N (22-K) and 170.7E, 10.2N (22-L), which accounts
for the two circular bright features in this region. It is also possible
that some of the Rahway Valles lavas were fed from the Grjot?
Valles complex, since we see a bright side-channel (22-M) extend-
ing from the eastern Grjot? Valles trough to the northwest end of
the Rahway Valles plains. This side-channel was also mapped by
Tanaka et al. (2005) and could have contributed Grjot? Valles
waters to Rahway Valles.
The moderately bright region of the north Cerberus basin just
south of the Grjot? Valles troughs, assigned unit AEc2 by Tanaka
et al. (2005), probably consists of lavas debouched from the
troughs or erupted from local Cerberus Fossae vents. In the middle
of this region is a circular dark spot (22-X) centered on the impact
crater Zunil. This 10-km-diameter crater is one of several fresh
martian craters (like Zumba) showing extended rays in THEMIS
infrared images (McEwen et al., 2005; Tornabene et al., 2006; Preb-
lich et al., 2007). Although the radar image does not show long
rays, the black spot identi?es Zunil as a dark-halo crater with a me-
ters-thick ejecta blanket that obscures the underlying Cerberus
lavas over distances of 30 km beyond the crater rim (see Section
8). The darkness of the feature indicates that the ejecta are primar-
ily ?ne-grained rather than rocky; any rocky ejecta associated with
rays and secondaries must be too sparse to show up as a bright
contrast feature against the bright Cerberus plains. Another simi-
lar-size dark spot (22-P) can be seen superimposed on the AEc2
?ows in the southern part of the Rahway Valles region. This corre-
sponds to another 10-km-diameter dark-halo crater at 174.3E,
7N. The so-called ??Mesogaea Crater?? (Veverka et al., 1976), at
169.2E, 8.5N, shows up as a small dark spot (22-Q). This spot is
no larger than the crater itself, which indicates that Mesogaea is
an older impact embayed by younger ?ows. Also, the visual south-
west dark streak for which this crater is known (Veverka et al.,
1976) shows no radar signature, which indicates that any eolian
deposit associated with the dark streak must be thin.
Wedged between Grjot? Valles in the north and central Cer-
berus Basin is a radar-dark region (22-R) corresponding to the
knobby terrain of Tartarus Montes, which is mapped as ??Nepen-
thes Mensae?? unit HNn by Tanaka et al. (2005). The mapped unit
extends farther east and north and is transected or embayed by
the Grjot? Valles N/S troughs (22-D) and the Rahway Valles volcan-
ics. The entire mapped HNn unit is radar-dark, which indicates a
surface of low small-scale roughness despite the high topographic
roughness (cf. Fig. 18). This characterization is consistent with
Tartarus Montes?s status as remnant highland terrain.
Outlying segments of HNn terrain are also located well to the
northeast, where they constitute the Tartarus Colles group. The
Tartarus Colles knobs are embayed by ?ngers of radar-bright lava
?ow over a 5  5 region centered on 172E, 24N (see Fig. 17).
Although these embaying lavas were mapped as distal Elysium
Rise ?ows (unit AHEe) by Tanaka et al., 2005 and Hamilton et al.
(2011) recently suggested a more southerly source, possibly Grjot?
Valles. This interesting region has garnered special attention with
the discovery of platy-ridged ?ows and rootless cones suggestive
of phreatovolcanic activity in ice (Lanagan et al., 2001; Bishop,
2008; Keszthelyi et al., 2010). These ?ndings, and the curiously iso-
lated location of the radar feature, strongly suggest a local source
for the radar-bright lavas.
20 (2012) 990?1030 1015Northern Cerberus is a classical dark feature in optical tele-
scopic maps and shows many crater wind streaks in orbiter
images. The Cerberus dark material has been shown to be mobile
The central-Cerberus region (east of Athabasca Valles) has a
complex array of lava ?ows, volcanic constructs and other features
that produce complex structure in the radar imagery. In the north-
west corner of Fig. 23 is a dark band (23-W) corresponding to a
topographic divide (24-W) that Lanagan (2004) identi?ed as sepa-
rating the Athabasca Valles ?ows on the west side from the cen-
tral-Cerberus ?ows to the east. On the east side of this divide is a
very bright lobe (23-G, 24-G) centered near 158E, 6N. This is
mapped by Lanagan (2004) as a shield (although no central vent
has been identi?ed) and by Vaucher et al. (2009b) as an unleveed
lava ?ow. (Note the narrow, radar-bright lava arm ?owing off the
northwest corner of the lobe and ponding against the dark topo-
graphic divide.) Lanagan (2004) used crater counts to place this
feature as one of the youngest Cerberus volcanic units east of Ath-
abasca Valles. He also identi?ed platy-ridged ?ows on the lobe,
which is consistent with its high radar brightness. To the SSE is a
low shield (23-H, 24-H) identi?ed by both Lanagan (2004) and
Vaucher et al. (2009b). This shield is relatively radar-dark and does
not stand out against the adjacent volcanic plains, although a
bright bowtie-shaped feature (23-J) just to the south may indicate
rougher ?ows originating near its southeast base. Northeast of the
bright lobe (23-G) is another low shield (23-K, 24-K) identi?ed by
both Lanagan (2004) and Vaucher et al. (2009b). This shield is
much brighter than 23-H, but its brightness blends in with the
bright plains farther east. A third shield (23-L, 24-L) identi?ed in
Vaucher et al. (2009b) is located in the moderately bright plains
to the northeast and is slightly darker than those plains. The fact
that these three low shields (24-H,K,L), like their South Cerberus
R, small-shield cluster; S, bright central channel of Marte Valles source ?ows; T and
V, outcrops of Medusae Fossae Formation; U, Hibes Montes inselberg; W,
topographic divide between Athabasca Valles and central Cerberus; X, Zunil Crater.
The gray-scale maximum is set to the maximum re?ectivity of 0.252 (5.99 dB)
measured over this region.
rus 2(Veverka et al., 1976), and Head et al. (1991) proposed the Cerberus
albedo feature to be a sand sheet. It is possible that such a mantling
deposit could contribute to radar darkening in this region. The ra-
dar-dark wedge of Tartarus Montes (22-R) is contained within the
dark Cerberus albedo feature, so some of the radar-darkness of this
unit could be enhanced by mantling (although knobby highland
terrain is likely to be radar-dark in any case). Immediately west
of Tartarus Montes and north of Athabasca Valles are some moder-
ately radar-bright Elysium Rise ?ows containing an even brighter
narrow lava channel from Grjot? Valles (22-S). These lie within
the optically dark Cerberus albedo feature, so any mantling here
must be thin. The southern part of the dark albedo feature contains
the Athabasca Valles headwaters, and it is possible that some of the
radar-darkness of the Athabasca source region could be associated
with heavy mantling in the lee of this section of the Cerberus
Fossae rift.
One ?nal feature worth noting in northern Cerberus (Fig. 22) is
a dark line (22-Y) that coincides with a section of Cerberus Fossae
?ssure extending east of Tartarus Montes. Since this feature is
wider than the ?ssure itself, it is probably associated with a late-
stage erosion or resurfacing by materials (waters, lavas, muds)
erupting from the ?ssure. Orbiter images show lower optical albe-
dos near the ?ssure, which might indicate a recent layer of darker
material emplaced on the older (radar-bright) lavas. Such late-
stage, localized resurfacing by ?ssure eruptions might also be
related to the radar-darkness that we noted earlier near the
Athabasca Valles source ?ssure.
The southern tip of the Cerberus basin shows up as a radar-
bright feature (23-A) where young Cerberus lavas have embayed
rugged highland terrain to the south (Fig. 23). A corresponding fea-
ture can also be identi?ed in the MOLA roughness map (Fig. 18b).
Lanagan (2004) mapped the ??South Cerberus?? feature as a distinct
?ow unit and, from crater counts, showed this to be the youngest
surface in Cerberus east of Athabasca Valles. This young age is
apparent from the MOLA shaded relief map (Fig. 24), which shows
the South Cerberus ?ows (24-A, 23-A) overlapping older ?ows to
the north (24-B, 23-B), which overlap still older ?ows (24-C, 23-
C) in the Marte Valles source region in central Cerberus Basin.
Inspection of orbiter imagery by Lanagan (2004) and ourselves
has identi?ed platy-ridged lava ?ows in South Cerberus, which is
consistent with the high radar brightness of the region. The orbiter
imagery also shows some complex albedo and textural patterns
along the northeast and east edges of the feature, which show up
in the radar imagery (Fig. 23) as contrasting brightness banding
indicative of surface texture variations at the ?ow margins.
The source of the South Cerberus ?ows is uncertain. Schaber
(1980) identi?ed a possible source caldera at 164.2E, 0.2S, but
this has not been con?rmed by post-Viking orbiter imagery. The
nearest known volcanic constructs are a close pair of low shields
at 161E, 0N and 161E, 1S (Vaucher et al., 2009a,b). The more
northerly shield (23-D, 24-D) lies in the darker ?ow unit (23-B,
24-B) to the north and thus is unlikely to be the source of the South
Cerberus ?ows. The smaller southern shield (23-E, 24-E) lies within
the western boundary of the South Cerberus radar feature and thus
might be a source for some of the South Cerberus ?ows, including
the ?ows that Vaucher et al. (2009b) have shown to embay the
south base of the more northerly shield. If the South Cerberus lavas
issued from this shield, or from a related vent on this same western
side, then the predominant ?ow direction would be west-to-east
rather than the north-to-south direction favored by Plescia
(1990). Alternatively, the South Cerberus lavas could have erupted
from a buried, unrecognizable ?ssure vent located in the central or
north-central part of the unit. Plescia (1990, 2003b) has, in fact, ar-
1016 J.K. Harmon et al. / Icagued that most of the Cerberus ?ows are ?ssure-fed ?ood lavas
upon which the low shields are superimposed as localized erup-
tions, and this may well apply to the South Cerberus unit.Fig. 23. Radar image (detail of Fig. 17) of central and south Cerberus. The labeled
features are: A, South Cerberus feature; B, smooth darker ?ows; C, central Cerberus
?ows in Marte Valles source region; D, low shield on smooth ?ow unit; E, low shield
on South Cerberus ?ow unit; G, bright lava lobe; H, K, and L, low shields; J, bright
?ows; M, knobby-cratered terrain; N, northern ?ow lobe overlapping central
Cerberus C-?ows; P, northern ?ow lobe overlapping N-?ows; Q, unleveed lava ?ow;
20 (2012) 990?1030counterparts (24-D,E), have different radar brightnesses but blend
in with their own immediate surroundings suggests that the low-
rus 2J.K. Harmon et al. / Icashield ?ow rheology is spatially variable and is governed by the
lava properties peculiar to the local plains.
Farther east are bright central-Cerberus ?ows (23-C, 24-C) that
are continuous with, and clearly contribute to, the Marte Valles
?ows. These are the oldest and stratigraphically lowest ?ows in
the region, as is apparent from the MOLA shaded relief map
(Fig. 24). In the west these bright ??C?? ?ows embay remnants of
knobby and cratered terrain, which show up as radar-dark features
(23-M, 24-M); the largest of these dark features corresponds to
Tombaugh Crater (162E, 3.5N) and its ejecta. To the south, the
MOLA shaded relief map (Fig. 24) shows the ??C?? ?ows being over-
lapped by the younger, darker ?ows ??B?? ?ows (23-B, 24-B) that lie
north of the even younger South Cerberus (??A??) ?ows (23-A, 24-A).
The B ?ows appear topographically smooth in MOLA roughness
(Fig. 18b) and shaded-relief (Fig. 24) maps, and their relatively
low radar brightness indicates that their small-scale surface tex-
ture is also much smoother than that of the A and C ?ows. Since
the B ?ows are intermediate in age between the A and C ?ows,
their smooth surface texture must be related to a particular lava
type or ?ow rheology rather than age effects.
To the north, the bright C ?ows are overlapped by younger, dar-
ker ?ows (23-N, 24-N) that are overlapped by a still younger ?ow
lobe (23-P, 24-P) south of Zunil Crater (23-X, 24-X); the ?ow stra-
tigraphy and age relationships here are apparent from the MOLA
shaded relief map (Fig. 24). On the west ?ank of the ??P?? lobe is a
radar-bright streak (23-Q) which is mapped as a distinct unleveed
lava ?ow (24-Q) by Vaucher et al. (2009b) based on spacecraft
imagery and maps. Off the east ?ank of the ??N?? lobe is a cluster
of small radar-bright shields (23-R, 24-R) which were mapped by
Fig. 24. MOLA shaded relief map of central and south Cerberus covering the same region
Global Data Sets, Planetary Data System Node, Arizona State University.20 (2012) 990?1030 1017Vaucher et al. (2009b) from spacecraft imagery. The fact that these
show up as distinct radar-bright spots suggests that these shield
?ows are localized and probably do not contribute to the Marte
Valles ?ows.
The most impressive radar-bright feature in Elysium is the
enormous Marte Valles channel system that connects eastern
Cerberus with northern Amazonis Planitia. Marte Valles is thought
to be a water-cut channel that was later ?lled with lavas (Plescia,
1990, 2003b; Burr et al., 2002b). Burr et al. (2002b) and Vaucher
et al. (2009a) place the water source in a buried eastern section
of Cerberus Rupes, just south of the Rahway Valles headwaters.
However, the MOLA shaded relief map (Fig. 24) and radar images
suggest that the source of the Marte Valles lavas is not in this re-
gion of far eastern Cerberus but rather in north-central Cerberus.
They also suggest that the westernmost exposed Marte Valles
?ows are the ??C?? ?ows in Figs. 23 and 24. These lavas probably
originate from Cerberus Rupes vents between 168?160E and
7?9N, but the sources appear to be largely buried under the youn-
ger ?ow lobes and shields. Once erupted, the radar-bright lavas
(23-C, 24-C) ?owed ESE across central Cerberus. They narrowed
and brightened near the point (23-S) where they pass between
the small-shield ?eld (23-R, 24-R) and a radar-dark outcrop (23-
T, 24-T) of the Medusae Fossae Formation. From here they wrap
around the south side of one of the radar-dark Hibes Montes insel-
bergs (23-U, 24-U) and then turn northeast to ?ow across eastern
Cerberus and into the narrow Marte Valles channel system. This
roughly follows the ?ow route described by Plescia (1990).
In Fig. 25 we show an image of Marte Valles downstream of its
northeastern turn. Note that east of 176E the very bright main
, and with the same feature labels, as in Fig. 23. The image is from the online Mars
rus 21018 J.K. Harmon et al. / Icachannel (25-A) breaks up into complex braided ?ow structures.
These correlate well with optically dark, anastomosing lava chan-
nels in orbiter optical imagery (Plescia, 1990, 2003b; Berman and
Hartmann, 2002). Optically bright, radar-dark regions of Marte Val-
les correspond to streamlined islands between these ?ows. One
can see a long radar-dark ?lament (25-B) that runs continuously
along the eastern side of the main channel and into the braided re-
gion. This feature shows up in Viking optical imagery as a bright
lineament separating the dark ?ows on either side and is presum-
ably a region free of young Marte Valles lavas. On the right ?ank of
the Marte Valles ?ows, between 177?181E and 2?9N, there is a
less-radar-bright region (25-C) (similar in brightness to Rahway
Valles) that is mapped as an older Cerberus ?ow unit AEc2 by Ta-
naka et al. (2005). The radar image shows where portions of these
?ows have embayed (25-D) cratered and knobby terrain (unit
HNu) to the east. South of this, and bordering the right side of
the main (AEc3) Marte Valles ?ows, is rugged radar-dark terrain
(25-E) corresponding to Medusae Fossae sediments and debris
(unit AEc1). Farther east lies the topographically smoother, radar-
dark terrain (25-F) of the Lucus Planum section of the Medusae
Fossae Formation (unit AAm).
The continuous and extremely radar-bright portion of Marte
Valles terminates at about 184E, 18.5N (25-G). Beyond this point,
the radar features become less bright and somewhat broken up as
the ?ows break out from the main channel and divert around insel-
ter optical imagery to differentiate ?ows and other surface features
in this region, the S-band radar imagery can play an important
mapping role.
Pre-upgrade Arecibo imagery (Harmon et al., 1992b, 1999)
Fig. 25. Radar image of Marte Valles. This is the sum of imagery from October 3,
October 6, October 10, and November 14, 2005. The labeled features are: A, central
Marte Valles channel; B, radar-dark region between younger Marte Valles ?ows; C,
older Cerberus ?ow unit; D, older Cerberus ?ow breakouts/embayments; E,
Medusae Fossae Formation sediments; F, Lucus Planum section of the Medusae
Fossae Formation; G, northern terminus of brightest Marte Valles ?ows; H, crater
splitting Marte Valles ?ows; J, prominent lava ?ow terminus in MOC image
SP240703. The gray-scale maximum is set to the maximum re?ectivity of 0.238
(6.23 dB) measured over this region.showed the Marte Valles connection between Elysium/Cerberus
and northern Amazonis, which had already been mapped by Ples-
cia (1990) from Viking images. However, the radar imagery also
indicated that the Amazonis radar-bright feature extended farther
southeast. Harmon et al. (1999) discussed this and suggested that
the northern and southern Amazonis features might be distinct
features from separate sources, despite the fact that both regions
were included in the same Aa3 unit in the earlier geologic map of
Scott and Tanaka (1986). Subsequent geologic mapping (Fuller
and Head, 2002; Tanaka et al., 2005) recognizes the northern and
southern Amazonis ?ows as separate units, and this is supported
by our post-upgrade imagery.
Our best single image covering the entire Amazonis region is
shown in Fig. 26, which was derived from observations on three
dates in October?November 2005. For additional context we pro-
vide MOLA shaded relief and vertical roughness maps in Fig. 27
covering the same region as in Fig. 26. In Fig. 28 we show a sepa-
rate image of north Amazonis based on newer imagery from the
2008, 2010, and 2012 oppositions; although this image suffers
from some degradation from Doppler equator effects, it shows
some of the northern Amazonis features more clearly owing tobergs and a large crater (25-H). Apparently here the ?ow regime
changes to one giving reduced surface stresses and a less blocky
surface texture. Nevertheless, both radar and orbiter images show
the ?ows continuing northeast to a debouchment in northern
Amazonis (see Section 6.1). Interestingly, one of the iconic orbiter
images of Marte Valles (MOC image SP240703) shows the terminus
of a prominent lava ?ow at 185.5E, 17.6N (Keszthelyi et al., 2000,
2008; Berman and Hartmann, 2002; Burr et al., 2002b). This fea-
ture (25-J) is visible in the radar imagery but is fainter than the
?ows in the main channel to the west. Apparently the small-scale
surface texture is smoother than that of typical Marte Valles ?ows,
despite the rugged appearance in orbiter imagery. This may indi-
cate less surface brecciation owing to weaker ?ow-surge stresses
in this minor terminating side-channel. Alternatively, the lower
radar brightness could result from surface mantling deposits
(Keszthelyi et al., 2008).
Keszthelyi et al. (2000) and Lanagan (2004) found that the Mar-
te Valles lavas have a ??platy-ridged?? morphology. This provides
additional support to our observation that platy-ridged ?ows tend
to have very radar-bright surfaces with extreme small-scale
roughness.
6. Amazonis
Amazonis Planitia is a ?at basin located between Elysium Plani-
tia and Olympus Mons. Although devoid of volcanoes, Amazonis is
now known to contain large expanses of young lava plains. Early
Arecibo CW observations indicated Amazonis to be the third major
center (after Tharsis and Elysium) of strong depolarized backscat-
ter on the planet (Harmon and Ostro, 1985). Subsequent radar
imagery (Muhleman et al., 1991; Harmon et al., 1999) con?rmed
this and mapped out the sources of the depolarized echoes. The
post-upgrade imagery presented here shows that Amazonis exhib-
its some of the most complex radar brightness structure on Mars.
Since dust mantling and other factors can make it dif?cult for orbi-
20 (2012) 990?1030more favorable incidence angles than in 2005. Finally, since south-
ern Amazonis shows so much ?ne structure in the radar imagery,
we have provided a blow-up image of this region in Fig. 29.
rus 2J.K. Harmon et al. / Ica6.1. Northern Amazonis
The new radar imagery (Fig. 26) supports earlier evidence from
orbiter optical images (Plescia, 1990) and pre-upgrade Arecibo
radar imagery (Harmon et al., 1992b, 1999) that much of northern
Amazonis is covered with Cerberus lavas that debouched from
Marte Valles. However, the resurfacing history of northern
Amazonis was probably a complex one (Fuller and Head, 2002;
Campbell et al., 2008) whose different stages may not be readily
distinguishable on the basis of radar brightness features. An epi-
sodic history is implied in the geologic map of Tanaka et al.
(2005), which has a younger volcanic unit (Cerberus unit AEc3) fan-
ning away from the Marte Valles mouth and an older, larger unit
(AAa2n) spreading farther to the Olympus Mons aureole. Fuller
and Head (2002) argue for at least two discrete episodes of lava
debouchment from Marte Vallis. The ?rst of these spread lavas as
far east as 204E and north to about 35N, while the lavas from
the more recent episode were con?ned to a more restricted region
Fig. 26. Radar image of Amazonis region. This is the sum of imagery from October 6, Oct
Olympus Mons aureole and eastern edge of the North Amazonis feature; B?D, outer Nort
boundary contact between the North Amazonis and South Amazonis regions; G, Tooting
with unusual thermal/spectral properties; K, Eumenides Dorsum section of the Medu
Amazonis Mensae section of the Medusae Fossae Formation; N, Gordii Dorsum section
Formation; Q, moderately bright region of mixed geologic units; R, Pettit Crater; S, Nicho
maximum is set to the maximum re?ectivity of 0.241 (6.18 dB) measured over this re20 (2012) 990?1030 1019near the valley terminus. Between these volcanic episodes there is
presumed to have been a ?uvial episode in which waters were re-
leased catastrophically through Marte Valles and ?owed across the
basin to the barrier presented by the Olympus Mons aureole (26-A,
28-A; Fig. 27). The region of highest radar brightness in our images
(Figs. 26 and 28) corresponds roughly with the proposed extent of
Fuller and Head?s ?rst volcanic episode. The radar brightness over
the region covered by their second volcanic episode is high, but not
any higher than that over much of the extended region covered by
the ?rst episode. We see two fainter bright features (26-B,C, 28-
B,C) extending farther north to about 38N between 197 and
210E and another (26-D, 28-D) extending farther east to the outer
Olympus Mons aureole (26-A, 28-A). These fainter outer features
correspond well with the mapped limits of Fuller and Head?s inter-
vening water ?ood, so it is possible that their enhanced brightness
is associated with ?uvial debris or erosion rather than lava ?ow
roughness. The southern edge of the north-Amazonis feature is
sharp and well-de?ned west of 200E; here it shows small scallop
ober 10, and November 14, 2005. The labeled features are: A, contact between outer
h Amazonis ?uvial features; E, lobate southern edge of the North Amazonis ?ows; F,
Crater; H, radar-dark region in isolated rectilinear depressions; J, radar-dark region
sae Fossae Formation; L, bright channel features south of Amazonis Mensae; M,
of the Medusae Fossae Formation; P, Lucus Planum section of the Medusae Fossae
lson Crater; T, bright channel within the Medusae Fossae Formation. The gray-scale
gion.
rus 21020 J.K. Harmon et al. / Icafeatures (26-E, 28-E) where the debouched lavas have embayed
degraded craters and knobs of Noachian terrain to the south. Be-
tween 199E, 24N and 206.5E, 21.5N lies the complex boundary
region (26-F) between the north-Amazonis and south-Amazonis
regions. We discuss this boundary in more detail in Section 6.2.
A few studies have been made of characteristics of the north-
Amazonis lava ?ows. Keszthelyi et al. (2000) were the ?rst to point
out the presence of platy-ridged ?ows in the region. They showed
speci?c examples of such ?ows in the region centered around
195E, 25N, which lies in a radar-bright area but outside Fuller
and Head?s second volcanic resurfacing zone. Lanagan (2004) also
mapped the location of platy-ridged ?ows in northern Amazonis.
He found such ?ows in the Marte Valles exit fan, in the 195E,
25N region studied by Keszthelyi et al. (2000), in a radar-darker
region a bit north of this, and in an isolated radar-bright feature
at 198E, 31N. Vaucher et al. (2009b) mapped a leveed lava ?ow
that crosses Keszthelyi?s platy-ridged region and then curves north
Fig. 27. Orbiter-based context maps of the Amazonis region: (a) MOLA shaded-relief map
roughness. The map boundaries are the same as for the radar image in Fig. 26.
Fig. 28. Radar image of North Amazonis. This is the sum of imagery from January 6,
2008, January 26?29, 2010, and February 14?16, 2012. The labeled features are: A,
contact between outer Olympus Mons aureole and eastern edge of the North
Amazonis feature; B?D, outer North Amazonis ?uvial features; G, Tooting Crater; H,
radar-dark region in isolated rectilinear depressions; J, radar-dark region with
unusual thermal/spectral properties. There is some image degradation from
Doppler-equator effects. The gray-scale maximum is set to the maximum re?ec-
tivity of 0.250 (6.02 dB) measured over this region.at 197E, 24.5N. All of this ?ow lies within our high-brightness
terrain, but it does not particularly stand out against the generally
high radar brightness of adjacent areas. Tanaka et al. (2005) map
this ?ow as the eastern end of their AEc3 unit, but our radar feature
extends farther east at the same brightness.
The radar image shows a dark bite out of the southeast corner of
the north-Amazonis feature that we attribute to obscuration by the
extended ejecta blanket of Tooting Crater (26-G, 28-G). Tooting is a
very young, 29-km-diameter crater located at 207.8E, 23.2N
(Fig. 27), just west of the Olympus Mon aureole (Mouginis-Mark
and Garbeil, 2007). Our image shows that the radar-absorbing ejec-
ta extend 72?110 km from the crater rim, which implies that the
ejecta must be more than a meter thick at this distance (see Section
8). This is plausible, given the ejecta thicknesses estimated by
Mouginis-Mark and Garbeil (2007). We see only very faint bright
; (b) MOLA vertical-roughness map, with lighter shading indicating higher km-scale20 (2012) 990?1030features associated with Tooting, which indicates a dearth of sur-
face or near-surface rocks in either crater or ejecta. In addition to
Tooting Crater, we found three smaller radar-dark spots that corre-
spond to dark-halo craters superimposed on the north-Amazonis
feature. These are located at: 203.9E, 24.2N; 207E, 28.6N; and
206E, 27.3N (see Section 8).
To the northeast of the north-Amazonis bright unit are two ra-
dar-dark features of note. The ?rst of these (26-H, 28-H) extends
from 206E to 215E and 31N to 37N. This corresponds to an unu-
sual pair of rectilinear depressions (Fig. 27) that Fuller and Head
(2002) suggest have been isolated from more recent ?ows and sed-
iments by a surrounding barrier of ??proto-Olympus?? ?ows. The ra-
dar-darkness of this unit is consistent with this interpretation. Just
to the north of this is an even darker heart-shaped feature (26-J,
28-J), centered at 210E, 40N, that contains within it a smaller
region with an unusual combination of thermal and spectral prop-
erties that Rogers and Christensen (2002) suggest is the signature
of a consolidated but highly porous material. This might be consis-
tent with the low radar brightness, since a high-porosity surface is
unlikely to contain radar-scattering structures such as rocks.
6.2. Southern Amazonis
Southeast of the northern Amazonis radar feature is another ra-
dar-bright complex (Figs. 26 and 29) that appears in every way to
be a separate and distinct unit that, unlike the northern complex,
rus 2J.K. Harmon et al. / Icahas not been resurfaced by Cerberus ?ows from Marte Valles. This
southern unit is characterized by a complex array of mostly N/S-
trending bright features separated by radar-dark borders. Most of
the radar-bright features that we see are contained within the geo-
logic map unit AAa2s (Amazonis Planitia 2 south unit) of Tanaka
et al. (2005). Tanaka et al. have this unit provisionally extending
up to about 26?27N latitude (well to the north of Tooting Crater),
where it contacts their AAa2n unit (of Cerberus origin). However,
the radar imagery does not support this, as we see a boundary with
the northern unit that is closer to that deduced by Fuller and Head
(2002). We see the boundary extending from 198E, 23N to 207E,
21.5N. The boundary appears as a meandering radar-dark line
(29-F), except for a tiny contact near 199.5W. This boundary line
lies close to the latitudes where Fuller and Head see a topographic
break at the point where the higher, downsloping southern terrain
?attens out in the northern basin. The southern upslope could then
provide a topographic barrier to the northern ?ows (both waters
and lavas) from Cerberus and thus de?ne a natural boundary be-
tween the northern and southern provinces. The radar-dark
boundary line hints at some resurfacing process. Possibly this line
represents a ?ne-grained sedimentary deposit at the high-water
mark of the aqueous ?oods that ?lled the northern basin. This is
supported by the fact that the line roughly traces the provisional
shoreline proposed by Fuller and Head (2002) to de?ne the south-
east limit of the ?oods. The boundary line coincides with a very
Fig. 29. Radar image of South Amazonis (detail of Fig. 26). The labeled features are: A, brig
leveed ?ow; C, modi?ed volcanic plains; D, localized bright ?ow patch in volcanic plain
between North and South Amazonis regions; G, Tooting Crater; H, bright ?ow in topog
Eumenides Dorsum section of the Medusae Fossae Formation; L, phreatovolcanic and
Medusae Fossae Formation; P, dark-halo craters; Q, moderately bright region of mixed ge
Fossae Formation. The gray-scale maximum is set to the maximum re?ectivity of 0.22120 (2012) 990?1030 1021dark (topographically smooth) line in the MOLA roughness map
(Fig. 27b), which could be consistent with planar sediments. This
also provides supporting evidence that the radar-dark boundary
line is unrelated to the N/S-trending dark lines separating the ra-
dar-bright features to the south, since these show up in the MOLA
roughness map as brighter (topographically rougher) features, pos-
sibly levees.
The radar complexity of southern Amazonis re?ects the spatial
and stratigraphic complexity of the region, and the radar feature
boundaries conform closely with morphologic features in orbiter
imagery. Perhaps the most striking radar feature is the large bright
wedge (29-A) centered at 206E, 18N. The feature corresponds to a
?at-?oored topographic depression that is mapped by Tanaka et al.
(2003b) as being older than the surrounding higher units. Orbiter
imagery shows this ?oor to be covered with platy-ridged lava ?ows
(Tanaka et al., 2005). Fuller and Head (2002) map this wedge as a
possible extension of the north-Amazonis terrain. However, Tana-
ka et al. (2005) include it in the south-Amazonis unit (AAa2s), and
we favor their assignment based on the radar imagery. The radar
brightness of the feature is nearly uniform except for a darker
shading in the northeast corner (see also Fig. 28). This darkening
is apparently caused by radar absorption in a semi-transparent
mantling layer associated with distal ejecta from Tooting Crater
(29-G) that extends more than 200 km beyond the crater rim. This
ht wedge-shaped feature in topographic low; B, radar-dark ridge with central bright
s region; E, southern extension of leveed ridge B-?ow; F, boundary contact region
raphic low; J, higher bright ?ows ?ngering into the Medusae Fossae Formation; K,
platy-ridged ?ows; M, older large impact crater; N, Gordii Dorsum section of the
ologic units; S, putative aureole runoff debris; T, bright channel within the Medusae
(6.56 dB) measured over this region.
rus 2is consistent with recent mapping of Tooting by Mouginis-Mark
(2012), which shows secondary cratering extending out 220 km.
To the west of the bright wedge is a radar-dark ridge with a nar-
rower bright feature (29-B) running along its central axis. Orbiter
imagery indicates this ridge is a laterally leveed lava ?ow, with
the bright streak corresponding to the central channel. Still farther
west is a large radar-bright feature (29-C) at 200E that corre-
sponds to ??modi?ed volcanic plains?? (unit Hr) in Fuller and Head
(2002) and ??Amazonis Planitia 1 south unit?? (AHAa1s) in Tanaka
et al. (2005). Included within this unit is a very bright patch
(29-D) centered at 200E, 16N that also coincides with a smooth
region in the MOLA vertical roughness map (Fig. 27b); the radar
data suggest that this feature represents a more recent localized
resurfacing by rough-surfaced lavas.
On the other (east) side of the bright wedge is some moderately
bright terrain and a very bright boomerang-shaped feature (29-S)
that abuts the outer edge of the Olympus Mons aureole. Tanaka
et al. (2005) include all of this terrain in their AAa2s unit, but Fuller
and Head (2002) identify the boomerang feature (their unit Aor) as
runoff from the aureole and de?nitely not volcanic. Fuller and Head
suggest this may be a dewatering ?ow from the exposure of
ground ice during aureole emplacement, and the MOLA maps
(Fig. 27) support their contention that this is an aureole-related
unit rather than south-Amazonis volcanics. With a re?ectivity of
r030 ? 9:2 dB (Table 2), this is one of the brightest non-volcanic
feature in our radar imagery and, hence, one of the roughest
non-volcanic surfaces on Mars.
Southeast of the ??wedge?? (29-A) is a complex array of small
bright features separated by dark lineaments. The narrow bright
feature (29-E) immediately south of the wedge is actually a curved
southern extension of the leveed ridge ?ow (29-B) west of the
wedge. To its east is another small bright feature (29-H) that (like
the wedge) lies in a topographic depression and that also shows
platy-ridged ?ows. Still farther east is a larger bright tongue (29-
J) of higher terrain that ?ngers into the dark Medusae Fossae
deposits to the southeast. On the other side of the ridge feature
(29-E) is a radar-bright region (29-L) where Keszthelyi et al.
(2010) have identi?ed phreatovolcanic constructs and platy-ridged
lava ?ows. All of these bright features are separated by radar-dark,
levee-like terrain. Also in this region is an irregular dark spot (29-
M) at 207.3E, 10N that corresponds to a prominent impact crater
that appears to be older than, and embayed by, the surrounding
bright ?ows (Scott and Tanaka, 1981a). Three smaller, younger
dark-halo craters show up as black spots (29-P) in the radar-bright
terrain farther northwest (see Section 8).
The provenance of the south-Amazonis lavas remains uncertain
and is one of the interesting open problems in Mars surface map-
ping. Fuller and Head (2002) have these lavas originating in Tharsis
Montes and ?owing around the aureole, although verifying a Thar-
sis source may be impossible because of obscuration by Medusae
Fossae Formation (MFF) deposits. Tanaka et al. (2005), on the other
hand, suggest a more local source. They interpret the lavas of their
AAa2s unit as eruptions from irregular pits and rilles in eastern
Amazonis, although they also suggest that there may be sources
buried under MFF deposits. Based on the variety and highly struc-
tured appearance of the radar features, and the identi?cation of
platy-ridged ?ows, we think a local source is more probable than
a Tharsis source for the radar-bright lavas. Keszthelyi et al.
(2010) recently suggested that the surface of South Amazonis
may have been molded by a complex process involving the inter-
leaving and erosion of ?ood lavas and MFF deposits. It seems likely
that some process of this kind is required to account for the ?ne
structure seen in the radar and orbiter imagery of the region.
1022 J.K. Harmon et al. / IcaAs for their AHAa1s unit (29-C) to the west, Tanaka et al. (2005)
suggest a resurfacing by lava ?ows, debris ?ows, and ?uvial sedi-
ments from Mangala Valles or other, presumably buried, sources.The mapped AHAa1s unit wraps around the southwest side of the
AAa2s unit and continues south to the equator along the east ?ank
of the Eumenides Dorsum plateau (26-K, 29-K) of the Medusae Fos-
sae Formation. The corresponding radar-bright feature follows the
mapped unit fairly closely. The brightest radar features in this re-
gion are two segments (26-L) near the equator at 210E. These
are located in the ?oor of a channel that wraps around the south
end of the Amazonis Mensae plateau (26-M) of the MFF. The south-
erly of these two segments lies on a mapped extension of the Thar-
sis AHt3 ?ows in the geologic map of Scott and Tanaka (1986).
However, since this feature is unconnected to any bright features
farther east, a Tharsis origin seems unlikely. The proximity of these
bright segments to the mouth of Mangala Valles (where it empties
into the channel) suggests a possible connection with Mangala
ef?uent. Mangala Valles itself also shows some radar brightness
in branch valleys and adjacent ?oodplain and plateau. The bright-
ness is probably associated with ?uvial deposition, but could also
be associated with the volcanism that has been proposed for the
Mangala complex (Leverington, 2007).
Southern Amazonis shows large expanses of extreme radar
darkness that correspond to various plateau lobes of the Medusae
Fossae Formation. The MFF has long been thought to consist of
thick deposits of a friable material such as volcanic ash or ignim-
brite (Scott and Tanaka, 1982; Bradley et al., 2002; Kerber and
Head, 2010; Kerber et al., 2011), although an ice-rich material
has also been proposed (Watters et al., 2007). Portions of the
MFF overlap with the X-band radar ??Stealth?? feature discussed in
Section 4.6. Our pre-upgrade S-band observations (Harmon et al.,
1999) were poorly suited for imaging the MFF, because the radar
track directly crossed the unit. However, we were able to show,
based on template ?ts to the leading-edge echo along the Doppler
equator, that the depolarized signal disappeared where the tracks
crossed the MFF. This would be consistent with a homogeneous,
radar-absorbing material devoid of rocks or other discrete scatter-
ers, as might be expected with an ash deposit. Our post-upgrade
imagery is much better suited for imaging the MFF because of
the higher resolution and more southerly tracks. In Fig. 26 we
see three main radar-dark sections of the MFF. In the south central
portion of Fig. 26 is the Eumenides Dorsum plateau (26-K, 29-K).
To the east, on the other side of the AHAa1s bright band, is the Gor-
dii Dorsum plateau (26-N, 29-N). Also, in the lower left corner of
Fig. 26 is the Lucus Planum section (26-P) of the MFF that we men-
tioned in Section 5.3. These three MFF regions have an average
r030 ? 18:0 dB, which is close to the 19 dB that one would ex-
pect just from ambiguity foldover from the south. These, along
with the Zephyria Planum section of MFF south of Cerberus Palus
(Section 5.3), are among the darkest radar features on Mars (Table
2). This is consistent with the MFF being a smooth (at decimeter
scales), homogeneous, rock-free deposit giving near-complete
absorption of radar waves.
Some other radar features in southern Amazonis are worth not-
ing. First, some moderate radar brightness (26-Q, 29-Q) can be
seen west of 198E and east of Marte Valles. This has been mapped
by Tanaka et al. (2005) as a mix of units HAa (Arcadia Planitia unit),
HNu (Nepenthes Mensae unit), and AHAa1s (Amazonis Planitia 1
south unit). In this region, and just east of Marte Valles, is a partial
bright ring (26-R) corresponding to the north rim of Pettit Crater.
The surrounding area is rather radar-dark, which may be an indica-
tor of distal ejecta from the crater. Another ring can be seen from
the rim of Nicholson Crater (26-S), which lies within the radar-dark
MFF of western Eumenides Dorsum. Another crater with dark ejec-
ta can be seen at 192.9E, 10.6N (see Section 8). Finally, on the east
?ank of the Gordii Dorsum MFF is a bright feature (26-T, 29-T) with
20 (2012) 990?1030a narrow bright ?lament. Scott and Tanaka (1986) map this as a
small island of Tharsis volcanic unit AHt3 surrounded by MFF,
while Kerber and Head (2010) identify it with a channel in the MFF.
rus 220 (2012) 990?1030 10237. Circular polarization ratios and volcanic texture
Up to now we have considered only the depolarized (SC) com-
ponent of the echo because, for reasons described in the introduc-
tion, of its suitability for mapping spatial variations in small-scale
surface roughness. However, the ratio of SC to OC power, the
so-called ??circular polarization ratio?? lc, also provides important
information on the textural qualities of the roughness, especially
when comparisons are made with terrestrial radar polarimetry
and accompanying ground truth. Here we present a polarization-
ratio image of Tharsis?Elysium?Amazonis and discuss implications
for the volcanic surface texture over these regions.
The fact that Mars?s volcanic regions give high circular polariza-
tion ratios has long been known from CW radar observations at
Arecibo and Goldstone (Harmon et al., 1982; Harmon and Ostro,
1985; Moore and Thompson, 1991; Harmon et al., 1992a). The
CW Doppler spectra from these observations showed lc values
greater than 0.8 when Tharsis longitudes were observed near the
planet limbs. For comparison, the same observations indicated
lower lc values of 0.4?0.5 for the cratered highlands, which is clo-
ser to the typical ratios measured for the Moon and Mercury (Har-
mon and Ostro, 1985). Since the CW spectra integrated over long
Doppler strips that would necessarily include non-volcanic re-
gions, it was apparent that there must be extended regions of vol-
canic terrain with lc values close to or exceeding unity. Con?rming
this would, of course, require making actual spatial images of lc.
No attempt was made to make lc images from the pre-upgrade
Arecibo delay-Doppler data (Harmon et al., 1999), primarily be-
cause of the code-sidelobe problem (Section 2) related to the use
of a 2-tap code (which primarily affected the OC echo owing to
sidelobe leakage from the strong quasispecular echo component).
However, Harmon et al. (1999) did show that the SC re?ectivity
of Mars?s brightest radar features is about twice that of some of
the roughest a?a lava ?ows on Earth. They suggested this was more
likely to be an indicator of a high (near unity) circular polarization
ratio than of a high total (OC + SC) radar re?ectivity. This would be
consistent both with the high lc indicated by the old CW results
and with the fact that rough terrestrial a?a lava ?ows show ratios
of only 0.3?0.6 in the wavelength range 6?24 cm (Campbell
et al., 1993, 1999; Harmon et al., 1999; Campbell, 2002; Plaut
et al., 2004).
We have obtained a new polarization-ratio map of Mars that
con?rms the earlier indications that high lc predominates over
much of the volcanic terrain. In Fig. 30 we show a lc image ob-
tained by combining data from November 14 and 17, 2005. To beat
down the noise, we smoothed the raw image over 10  10-pixel
blocks, which coarsened the effective resolution to 0.5 or about
30 km. The main result to note in Fig. 30 is the green shade that
covers vast expanses of our radar-bright volcanic terrain. This
shade corresponds to lc = 0.80?1.15. At the highest incidence an-
gles (mainly toward the top and left edges) one sees a messy sprin-
kling of hues that is attributed to low signal-to-noise ratios where
both OC and SC signals become comparable with the OC clutter
noise. However, the green-hued tones in the main radar-bright fea-
tures are robust, with lc estimation errors being typically less than
5%. Some yellow- to red-hued spots within these features corre-
spond to locations with statistically signi?cant lc values exceeding
unity. These include an area of the Pavonis east lava apron with
lc = 1.33 ? 0.07, a spot on the northwest ?ank of Ascraeus Mons
with lc = 1.38 ? 0.09, and a spot in the Olympus basal plains with
lc = 1.40 ? 0.10.
Harmon et al. (1999) considered several scenarios for the high
depolarization from Mars lava ?ows. The simplest one is that Mars
J.K. Harmon et al. / Icalavas are blockier and give more multiple scattering than most ter-
restrial ?ows (including rough a?a ?ows). The classic example of ahighly depolarizing terrestrial ?ow is SP Flow, a basaltic andesite
lava ?ow in the San Francisco volcanic ?eld of north-central Ari-
zona (Schaber et al., 1980) with a surface dominated by faceted
blocks measuring tens of centimeters. Other examples of highly
depolarizing blocky ?ows are found around the Long Valley Cal-
dera in California (Plaut et al., 2004). AIRSAR images of SP Flow give
average lc values of 0.95 at 24-cm wavelength (Campbell et al.,
1999; Campbell, 2002), which is similar to that of the martian
?ows. However, a substantial fraction of SP ?ow has lc > 2. Such
high lc values are not seen in our Mars data, though it is quite
possible that they exist in small patches but are washed out by
averaging over our coarse (10  10-smoothed) resolution cell.
If blocky ?ows are to account for the highly depolarizing mar-
tian features, then one must explain why such ?ow surface texture
is so prevalent on Mars and includes not only platy-ridged ?ows
but most other radar-bright ?ows on the planet. One characteristic
shared by many martian lava ?ows is their extreme length. It is
possible that high effusion rates or other factors affecting ?ow
length may also produce high ?ow surface stresses resulting in
blocky textures. Keszthelyi et al. (2000, 2004) suggested that some
martian ?ows, such as the platy-ridged ?ows, may form an insulat-
ing surface crust that is fragmented (brecciated) in eruption surges
and breakouts through a process analogous to that observed for
some Icelandic ??rubbly pahoehoe?? ?ood basalts. Recently, Camp-
bell et al. (2009) applied this same Icelandic analogue to the Moon,
as a possible explanation for certain highly depolarizing (lc  1.5)
regions in Maria Serenitatis, Imbrium, and Crisium.
Another possibility suggested earlier by Harmon et al. (1999) is
that the ?ow near-surface layer supports multiple volume scatter-
ing, either in a frothy lava matrix or in a conglomerate of lava rub-
ble mixed with dust. Subsurface volume scattering has also been
proposed to explain elevated (lc  0.7) depolarization in certain
low-TiO2 regions of the lunar maria (Campbell et al., 2010), where
the scatterers are presumed to be impact-generated rocks. While
we cannot dismiss volume scattering as a contributor to the highly
depolarized Mars echoes, including this in martian volcanic ?ow
scenarios has always seemed somewhat arti?cial. Therefore, we
consider blocky surface scattering to be a more plausible explana-
tion for the Mars radar depolarization behavior, especially in light
of the recent studies of lava ?ow mechanics and surface morphol-
ogy on some potentially analogous terrestrial ?ows.
Finally, some comparisons with Venus radar observations are in
order. Campbell and Campbell (1992) reported lc = 0.05?0.4 from
Arecibo S-band observations of radar-bright lava ?ows in Eistla
Regio and Sedna Planitia, from which they concluded that these
Venusian ?ows were mostly smoother than a?a ?ows and non-
blocky in texture. Arecibo results from Campbell et al. (1999) found
higher lc values of 0.7 and 1.0?1.2 in the Beta Regio and Maxwell
Montes highlands, respectively, but only in subregions with anom-
alously high Fresnel re?ectivities in excess of 0.4. From this they
concluded that high Venusian lc values are attributable to en-
hanced multiple scattering off blocky, high-dielectric surfaces.
For Mars we see no evidence that the bright lava ?ows have anom-
alously high total (OC + SC) diffuse re?ectivities (compared to ter-
restrial ?ows), nor do we know of any reports of anomalously high
Fresnel re?ectivities from studies of the Mars quasispecular echo.
Since radar-bright martian lava ?ows yield comparably high lc
without the multiple-scattering boost from high dielectrics, we
conclude that the Mars ?ows are inherently rougher or blockier
than the anomalous Venus highland surfaces. Combining this with
the lower lc values of the Venus lowland ?ows suggests that, in
general, Mars lava ?ows have rougher small-scale surface texture
than lava ?ows on Venus.
and SC radar images from November 14 and November 17, 2005. Note that most of the
shades (lc = 0.80?1.15). The red?blue mottled regions at the higher latitudes and on the
is too high to give a meaningful ratio estimate.
rus 28. Dark-halo craters
As we have seen, Mars?s appearance in depolarized radar
images is dominated by bright volcanic features rather than impact
features. The radar signatures of impact craters are subdued and
mostly limited to faint rim enhancements and quasi-circular dark
patches where fresh impacts have been made into bright volcanic
plains. This is in marked contrast with radar images of the Moon,
Mercury, and Venus, which show prominent bright features asso-
Fig. 30. Radar image of the circular polarization ratio lc. This is computed from OC
bright radar features on Mars have circular polarization ratios corresponding to green
left and right edges of the ?gure correspond to regions where the image noise level
1024 J.K. Harmon et al. / Icaciated with crater ?oors, ejecta haloes, and even extended rays
(Zisk et al., 1974; Campbell and Burns, 1980; Thompson, 1987;
Harmon et al., 2007). Bright-halo craters on the Moon and Mercury
show bright ejecta rings that extend about one crater radius be-
yond the crater rim (Harmon et al., 2007; Ghent et al., 2010) and
that are attributed to enhanced diffuse backscatter from relatively
fresh, rocky ejecta deposits. To date we have identi?ed only a few
such bright-halo craters on Mars, the most striking examples of
which are located in the Chryse/Xanthe region. We will defer a dis-
cussion of these bright martian craters to our follow-up paper on
the Chryse/Xanthe imaging. Of more immediate interest for this
paper are the so-called ??dark-halo craters.?? These are craters with
radar-dark haloes extending several crater radii beyond the rim
and well beyond the outer edge of any bright ejecta halo. Dark-halo
craters have been identi?ed in radar images of the Moon (Ghent
et al., 2005), Venus (Schaber et al., 1992), and Mercury (Harmon,
2007; Harmon et al., 2007). The dark haloes have been interpreted
as thick deposits of ?ne-grained, block-poor ejecta (Ghent et al.,
2005, 2010). Since these deposits are subject to erasure by impact
gardening, dark-halo craters are considered to be relatively fresh
(Ghent et al., 2005). Although until now dark-halo craters had
not been identi?ed in Mars radar imagery, Ghent et al. (2010)
did ?nd martian craters that have low-temperature haloes in THE-
MIS nighttime IR images and that they considered to be analogous
to the radar dark-halo craters on the Moon and Venus. It seems
reasonable, then, to assume that the dark crater spots that we have
identi?ed in our radar images are likely martian counterparts to
the dark-halo craters and are related to the THEMIS dark-halo IR
craters.20 (2012) 990?1030The ?rst detailed comparative study of the dark-halo craters is
that of Ghent et al. (2010). They de?ned a ?ne-ejecta run-out
parameter r by
r ? ?rH  R?=R ?1?
where R is the crater radius and rH is the halo radius (de?ned as the
distance between the crater centroid and the most distant halo
lobe). Ghent et al. (2010) made a survey of dark-halo craters from
radar imagery of the Moon and Venus and THEMIS nighttime IR
images of Mars, from which they estimated power?law relations
for r(R). In Fig. 31 we reproduce their best-?t r(R) relations for
the dark-halo craters, along with their corresponding relation for
lunar bright-halo craters. Note that the line for the Venus dark ha-
Fig. 31. Ejecta run-out parameter r versus crater radius R for the dark-halo craters
listed in Table 3 (crosses), along with the corresponding ?tted power?law relation
r = 7.53 R0.076 (dashed line). Shown for comparison (solid lines) are the power?
law relations for dark-halo craters on the Moon, Venus, and Mars and bright-halo
craters on the Moon, from Ghent et al. (2010).
loes is higher and steeper (r / R1/2) than that for the Moon. Ghent
et al. attribute this to ?ne-ejecta entrainment in the dense Venusian
atmosphere, as opposed to the purely ballistic emplacement for the
lunar case. They also noted that the Mars line was somewhat higher
than the lunar line, despite the higher martian gravity. From this
they concluded that ?ne-ejecta emplacement on Mars includes
non-ballistic effects, although the relatively ?at r(R) power?law
argued against a Venus-like entrainment mechanism. Speci?cally,
they suggested that the martian haloes are spread by high-speed
winds produced in expansion of the impact-generated vapor plume,
whereas the Venusian haloes likely involve entrainment in turbu-
lent vortices generated by the ejecta curtain (which is consistent
with the steep R1/2 scaling).
We have surveyed our Mars depolarized radar images and com-
piled a list of 30 dark-halo craters suitable for comparing with the
scaling relations in Fig. 31. Our selection criteria required that
there be a quasi-circular dark patch or halo that coincides with a
known impact in orbiter imagery and, speci?cally, with a crater
whose ejecta are clearly superimposed on (i.e., postdate) the
surrounding bright plains or ?ows. For each of these craters we
computed the crater radius R, halo radius rH, and normalized
run-out parameter r. We list these parameters and the crater
locations in Table 3 and plot the r(R) points in Fig. 31.
The new comparison displayed in Fig. 31 appears to support and
strengthen the conclusions reached by Ghent et al. (2010) based on
their survey of martian dark IR haloes. Our radar-based Mars r
values average about 30% higher than the IR-based Mars line. This
increases the difference between the Mars and Moon relations and
thus strengthens the case for non-ballistic effects in martian dark-
halo emplacement. Our new Mars points remain below the Venus
line, which is consistent with the lower density of the martian
tures not only for the so-called platy-ridged ?ows common in
Elysium and Amazonis but also for the more conventional non-
J.K. Harmon et al. / Icarus 2atmosphere. Also, the shallow (r / R0.076) R-dependence of the
Table 3
Dark-halo crater locations and parameters.
No. Long. (E) Lat. () R (km) rH (km) r
1 146.02 22.02 8.7 68 6.9
2a 166.19 7.70 5.1 35 5.9
3 174.42 7.16 5.0 51 9.4
4 192.89 10.60 15.1 111 6.3
5 198.47 17.59 1.6 10 5.4
6 198.67 18.00 2.3 15 5.7
7 199.51 21.38 2.1 14 5.9
8 202.06 15.62 4.7 35 6.5
9 202.33 14.18 6.7 41 5.2
10 203.86 24.23 4.0 39 8.8
11 204.95 13.21 2.5 29 10.5
12 206.01 27.31 3.2 26 7.4
13 207.02 28.64 12.8 49 2.8
14b 207.77 23.18 13.7 111 7.1
15 222.55 35.37 2.5 15 5.1
16 223.13 11.12 1.9 19 9.0
17 224.78 10.65 5.9 42 6.1
18 226.36 11.40 2.5 20 6.8
19c 226.61 17.28 5.2 46 7.9
20d 226.93 28.67 1.6 17 9.1
21e 228.29 18.40 7.6 46 5.1
22 230.22 23.01 3.6 35 8.7
23 230.64 14.08 2.8 19 6.0
24 240.92 9.60 11.5 102 7.9
25 243.81 15.61 4.8 39 7.1
26 249.12 17.98 12.1 137 10.4
27 257.35 1.07 2.6 21 7.0
28 258.72 12.32 5.4 29 4.4
29 260.08 19.28 9.9 74 6.5
30 262.61 9.55 5.2 44 7.5Notes: R is crater radius, rH is the maximum dark ejecta runout distance from the
crater center, and r is the runout parameter de?ned by r = (rH  R)/R. Named
craters: aZunil, bTooting, cPangboche, dZumba, eKarzok.platy ?ows found on Tharsis. This suggests that blocky surfaces
are not peculiar to a particular ?ow type but rather are related to
emplacement conditions and processes common on Mars. The
most striking single characteristic of martian lava ?ows is their
extreme length, and we have suggested here that blocky ?ow sur-
faces may be a result of the high lava effusion rates and other ther-
mo-mechanical processes responsible for such long ?ows.
The radar-dark features are just as interesting, in some respects,
as the radar-bright ones. Low radar brightness is not solely an indi-
cator of surfaces that are non-volcanic or have low to moderate
rock coverage, but is also associated with radar-absorbing man-
tling deposits that may be obscuring what might otherwise be ra-
dar-bright features. The pre-upgrade imagery (Harmon et al., 1999)Mars radar points agrees with that of the IR-based line and thus
supports the suggestion by Ghent et al. that the martian dark-halo
ejecta have not been subjected to the same entrainment mecha-
nism as that working on Venus.
We have noted that some of our dark-halo Mars craters show a
faint to moderately bright central core, while others are completely
dark. The bright cores are small, measuring less than a crater diam-
eter in size to, at most, two crater diameters. This is consistent with
some rocky ejecta populating the ?oors and/or near-rim areas.
Those martian dark-halo craters lacking a bright core must have
craters ?oors and rim-collars that are de?cient in rocky rubble or
that are blanketed in the same ?ne-grained deposits making up
the dark haloes.
9. Summary and conclusion
In this paper we presented a radar study of the surface charac-
teristics of Mars?s major volcanic regions based on imagery ob-
tained with the upgraded Arecibo radar telescope. We have been
able to substantially improve on the pre-upgrade Arecibo results
thanks to a much ?ner image resolution and, for some regions, a
more favorable sub-Earth latitude aspect. Furthermore, the prolif-
eration in recent years of geologic studies based on post-Viking or-
biter imagery has provided a context within which the radar data
can be used to address some interesting current problems in Mars
surface science.
Mars?s volcanic regions stand out in radar images as bright
depolarized backscatter features. The post-upgrade imagery has
con?rmed and strengthened the earlier association of bright radar
features with relatively fresh lava ?ows. We have shown that all
three major volcanic provinces on Mars (Tharsis/Olympus, Ely-
sium, Amazonis) show these bright features. The brightest features
are found on the Tharsis Montes, although Elysium and Amazonis
also show features that are within 1.5?3 dB of the brightest Tharsis
features. We have also con?rmed earlier indications that the
brightest features have an unusually high absolute depolarized
(SC) re?ectivity when compared to terrestrial lava ?ows. In an ear-
lier paper (Harmon et al., 1999) we suggested that this may be
more an indication of a high degree of depolarization (as expressed
in the circular polarization ratio lc) than of a high total diffuse
re?ectivity. Our new polarization-ratio image, which shows vast
areas with lc  1, con?rms this and also explains the high appar-
ent ratios indicated by old dual-polarization CW spectra. Although
such highly depolarizing lava ?ows are not unknown on Earth, the
fact that they are the norm rather than the exception on Mars sug-
gests that conditions conducive to the production of blocky ?ow
surfaces are prevalent on the planet. We infer such blocky ?ow tex-
20 (2012) 990?1030 1025showed low radar brightness over the southern cratered highlands,
on islands of old terra sticking out of the northern volcanic plains,
and over the equatorial Medusae Fossae Formation. The post-up-
rus 2grade imagery has con?rmed these ?ndings and also revealed
many other radar-dark regions that we have identi?ed, de?nitely
or provisionally, with: (1) minor-shield mantling deposits; (2)
??Greater Stealth?? deposits; (3) the so-called ??fan-shaped deposits??,
putatively attributed to major-shield glaciation, (4) terrain-
softened surfaces on the cores of the major-shield rift aprons; (5)
regions known or suspected to be eolian features such as dune
?elds; and (6) the extended ejecta blankets of some fresh impact
craters (the so-called ??dark-halo craters??).
It is worthwhile reviewing here some of the more noteworthy
regional ?ndings and pointing out those warranting further study.
The Tharsis/Olympus region is the one to which we have devoted
the most space here, because of its size and complexity. It is also
the region that bene?ted most from the changed sub-Earth latitude
aspect, especially around Pavonis and Arsia Montes. The major
shields of Olympus Mons and Tharsis Montes vary signi?cantly
in their radar appearance. Both Ascraeus and Arsia Montes show
high radar brightness over most of their upper shields that is con-
sistent with extremely rough and lightly mantled lava ?ows. Pav-
onis Mons, which we properly imaged for the ?rst time, has a
much different radar appearance that includes a bright mid-shield
collar surrounding a dark, presumably heavily mantled, summit
plateau. The extremely patchy radar brightness seen on the Olym-
pus Mons shield in the pre-upgrade imagery was con?rmed, with
the new observations showing radar-bright features associated
with concentrations of lava channels and dark features identi?ed
with impact craters and mantled regions.
The radar imagery over Tharsis/Olympus has also provided
interesting results for geologic features associated with, but not
actually located on, the major shields. For example, the south rift
aprons of all three Tharsis Montes show regions where orbiter
imagery shows radar-bright lava ?ows emerging from softer, ra-
dar-darker terrain on the apron core. The radar darkness of some
of these core zones suggests a ?ne-grained, non-rocky surface,
which might be consistent with the ?uvial processes proposed
for the rilled apron terrain. More work is needed to determine
the nature of these intriguing apron features. Ascraeus and Pavonis
Montes are also associated with small-shield ?elds that show up as
some of the most prominent radar-bright features on Tharsis.
Olympus Mons shows a prominent radar-bright southern collar
that is associated with extensive plains volcanism. All three Tharsis
Montes shields show very dark features off their western ?anks
that are associated with the so-called ??fan-shaped deposits?? and
that may be ?ne-grained mantling deposits left by glaciers. Also
on the west slope of Tharsis are extended radar-bright lava ?ows
that apparently originate from Tharsis Montes and that are ob-
scured in places by superimposed Stealth and Pavonis fan deposits
and dark-halo craters. Vast radar-bright lava ?ows also extend
down the Tharsis east slope all the way to Kasei Valles and Echus
Chasma. One of the advantages of the improved resolution and
coverage is that it has allowed us to image the Tharsis minor
shields (tholi and paterae). These show up as almost entirely radar
dark features that, in many cases, contrast starkly with the sur-
rounding bright lava plains. The darkness of these shields is consis-
tent with surface mantling deposits of either eolian or pyroclastic
origin. The possibility of pyroclastic deposits on Tharsis, either on
the minor shields or in other regions such as Alba Patera, the Pav-
onis Mons summit plateau, and the Tharsis Montes rift aprons, is
an intriguing one that has been considered by a number of
researchers over the years and that still warrants further study.
Our next most intensively studied region is Elysium. Although
there was no need to improve on the favorable sub-Earth latitude
aspect of the pre-upgrade imagery, the improved resolution and
1026 J.K. Harmon et al. / Icaquality of the newer images has revealed many more feature de-
tails over this region. Furthermore, the ?uvio-volcanic Cerberus re-
gion of Elysium has become one of the most intensively studiedregions of the planet in the intervening years, which provides valu-
able geologic context for evaluation of the radar imagery. The new
imagery con?rms Elysium Mons as the only volcanic construct in
the region showing a prominent radar-bright feature. The new
imagery shows more detail in the ?ows radiating away from this
major shield, including the prominent ?ows extending well to
the east. The smaller shields of Hecates Tholus and Albor Tholus
do not have prominent radar signatures. Hecates Tholus shows a
modest bright feature on its east side but is radar-dark on the west
side, while Albor Tholus shows a bright feature on its lower south-
ern slopes but a completely dark upper shield. The radar-dark
shield areas on these two volcanoes are consistent with indications
of mantling deposits (eolian or pyroclastic) in orbiter imagery. All
of the major ?uvio-volcanic regions of Cerberus basin show up as
radar-bright features indicative of rough-surfaced lavas. These
regions include (1) Athabasca Valles and its Cerberus Palus catch
basin, (2) Grjot? Valles, (3) Rahway Valles, (4) South Cerberus,
and (5) Marte Valles. We also see a radar-bright feature from the
isolated Tartarus Colles region, which also shows evidence of
hydrovolcanism. The brightest features in Cerberus correspond to
regions showing the greatest concentration of platy-ridged lava
?ows, which suggests that the mechanics of ?ow emplacement
include high stresses that produce highly disrupted surfaces. Ra-
dar-dark features in and around Cerberus include highland terra
remnants and inselbergs, fresh impact craters (including Zunil),
and Medusae Fossae deposits. Western Elysium (west of Elysium
Mons) shows modestly enhanced radar brightness features that
correlate with units mapped variously as lava ?ows, debris ?ows,
erosional detritus, and lahars.
Amazonis Planitia is probably the region that bene?ted the
most from the improved resolution of the post-upgrade imagery.
The pre-upgrade imagery (Harmon et al., 1999) showed the high
radar brightness in Amazonis to be concentrated in two seemingly
separate but contiguous sections: a northern section associated
with Cerberus lavas debouched from Marte Valles and a southern
section that appeared likely to have an entirely different prove-
nance. Subsequent geologic maps and studies based on post-Viking
orbiter imagery have come to support separate origins for the two
sections and have also identi?ed lava ?ows in the southern unit
(some being of the platy-ridged type) that can account for the ra-
dar-brightness of this region. Our new higher-resolution radar
imagery also supports the distinctness of the two sections, based
mainly on two ?ndings: (1) the identi?cation of extremely detailed
brightness structure in the southern unit that contrasts with the
more amorphous structure seen in the north, and (2) the identi?-
cation of a radar-dark band in the contact region that may repre-
sent a topographic barrier or shoreline. The new radar imagery of
North Amazonis continues to support a Marte Valles origin for
these lavas, which seems to be uncontroversial. The brightest re-
gions in this unit correspond closely to the mapped extent of the
main volcanic resurfacing episode from Marte Valles, while fainter
features seen extending farther out may be ?uvial deposition fea-
tures. The radar complexity of South Amazonis is striking and
would seem to favor a local origin for most of these lavas, although
additional contributions from Mangala Valles or Tharsis are also
possible. Identi?cation of possible sources for the South Amazonis
volcanics is complicated by the overlapping Medusae Fossae For-
mation (MFF) deposits, which show up as prominent radar-dark
features. Our new imagery has con?rmed earlier indications from
the pre-upgrade data that the MFF deposits consist of a ?ne-
grained, rockless material, which would be consistent with the
ash-?ow hypothesis. At this point, the origins of the South Amazo-
nis lavas and MFF deposits remain among the more interesting
20 (2012) 990?1030outstanding problems in Mars geology.
Although we are approaching the limits of what can be done in
the way of Mars imaging from Earth with existing radar systems,
Such a radar would give much higher spatial resolution and deeper
subsurface penetration than our S-band imagery and would pro-
rus 2vide a valuable data set for comparisons with our shallower imag-
ery. Comparisons of this type would be particularly interesting
over Mars?s major volcanic regions, where radar absorption by
mantling deposits appears to be an important effect. Of course,
ultimately one would like to obtain ??ground truth?? for a radar-
bright Mars lava ?ow by actually sending a lander. So far, lava
?ows have been ruled out as landing sites on the grounds that they
pose surface-roughness hazards, are located at too high an altitude
for lander aerobraking, or are lacking in biological interest. Never-
theless, it is becoming increasingly clear that some radar-bright
lava ?ows are among the youngest and geologically most intrigu-
ing surface features on Mars and thus offer potentially spectacular
landing sites.
Acknowledgments
The National Astronomy and Ionosphere Center (Arecibo Obser-
vatory) was operated, under a cooperative agreement with the Na-
tional Science Foundation (NSF), by Cornell University (for the
2005?2010 observations) and by a consortium comprising SRI
International, Universities Space Research Association (USRA),
and Universidad Metropolitana (for the 2012 observations). The
S-band radar observations were also made possible with support
from the National Aeronautics and Space Administration (NASA).
Diana Husmann?s work at Arecibo was supported by a grant to
Cornell University from the Research Experience for Undergradu-
ates (REU) program of the NSF. We would like to thank the staff
of Arecibo Observatory for their support, and in particular the fol-
lowing people: Victor Negr?n, Alfredo Santoni, and Joe Greene for
transmitter maintenance and operations; Bill Sisk for digital hard-
ware support; and Phil Perillat for datataking software support. We
also thank John Chandler and Jon Giorgini for ephemeris support.
We are grateful to Chris Magri for kindly volunteering to do Mars
observations for us on October 14 and 15, 2005. Spacecraft images
used in this paper were from the online Mars Global Data Sets,
Planetary Data System (PDS) Node, Arizona State University
http://www.mars.asu.edu/data/. The Mars Orbiter Laser Altimetersome gains can be expected from continued Arecibo observations.
Some improvements in image quality may be achieved by observ-
ing during the closer oppositions and at full transmitter power.
Signi?cant improvement in the quality of polarization-ratio images
may also be possible by adopting some alternative delay-Doppler
scheme such as ??coherent frequency stepping?? (Harmon, 2002)
that gives less self-clutter noise than the long-code method used
here. Although Mars will be too far south to be observable from
Arecibo during the 2014?2018 oppositions, the planet?s return
with the October 2020 opposition will be extremely favorable be-
cause of the small (0.41 AU) planet distance and southerly (20S
lat.) sub-Earth tracks.
Over the long term, the most signi?cant advances in Mars radar
imaging can be expected to come from Mars-orbiting synthetic-
aperture radars (SARs). Pioneering work has been carried out in
recent years with the SHARAD and MARSIS Mars-orbiting instru-
ments. These are low-frequency sounding radars that use SAR
techniques for isolating deep subsurface re?ections along the ?ight
track rather than for two-dimensional imaging. No true imaging
radars have yet been deployed in Mars orbit, although proposals
have been made for an orbiting SAR that would operate at a higher
frequency and that would combine imaging with a shallow sound-
ing capability (Campbell et al., 2001, 2004; Paillou et al., 2006).
J.K. Harmon et al. / Ica(MOLA) topographic data used in the delay-Doppler mapping were
obtained from the PDS Geosciences Node, Washington University
in St. Louis http://pds-geosciences.wustl.edu/missions/mgs/meg-dr.html. We wish to thank Lynn Carter and another, anonymous,
reviewer for their constructive comments and suggestions.
References
Balme, M.R., Gallagher, C.J., Gupta, S., Murray, J.B., 2011. Fill and spill in Lethe Vallis:
A recent ?ood-routing system in Elysium Planitia, Mars. In: Balme, M.R.,
Bargery, A.S., Gallagher, C.J., Gupta, S. (Eds.), Martian Geomorphology, Special
Publ. 354. Geological Society, London, pp. 203?227.
Baptista, A.R. et al., 2008. A swarm of small shield volcanoes on Syria Planum, Mars.
J. Geophys. Res. 113, E09010. http://dx.doi.org/10.1029/2007JE002945.
Baron, J.E., Simpson, R.A., Tyler, G.L., Moore, H.J., Harmon, J.K., 1998. Estimation of
Mars radar backscatter from measured surface rock populations. J. Geophys.
Res. 103, 22695?22712.
Basilevsky, A.T. et al., 2005a. Morphology and geological structure of the western
part of the Olympus Mons volcano on Mars from the analysis of the Mars
Express HRSC imagery. Solar Syst. Res. 39, 85?101.
Basilevsky, A.T. et al., 2005b. Mars Express HRSC view of western Olympus Mons:
Evidence for ice-bearing deposit and high-altitude glaciation. Lunar Planet. Sci.
36. Abstract 1060.
Berman, D.C., Hartmann, W.K., 2002. Recent ?uvia, volcanic, and tectonic activity on
the Cerberus plains of Mars. Icarus 159, 1?17.
Bishop, M.A., 2008. Higher-order neighbor analysis of the Tartarus Colles cone
groups, Mars: The application of geographical indices to the understanding of
cone pattern evolution. Icarus 197, 73?83.
Bleacher, J.E., Greeley, R., Williams, D.A., Neukum, G., 2007a. Morphometric
characterization and comparison among the Tharsis Montes-related low
shield and ?ssure vent ?elds. Lunar Planet. Sci. 38. Abstract 1314.
Bleacher, J.E., Greeley, R., Williams, D.A., Cave, S.R., Neukum, G., 2007b. Trends in the
effusive style at the Tharsis Montes, Mars, and implications for the development
of the Tharsis province. J. Geophys. Res. 112, E09005. http://dx.doi.org/10.1029/
2006JE002873.
Bleacher, J.E. et al., 2009. Spatial and alignment analyses for a ?eld of small volcanic
vents south of Pavonis Mons and implications for the Tharsis province, Mars. J.
Volcanol. Geotherm. Res. 185, 96?102.
Bleacher, J.E. et al., 2010. Updates to the catalog of Tharsis province small volcanic
vents, Mars. Lunar Planet. Sci. 41. Abstract 1615.
Borraccini, F., Lanci, L., Wezel, F.C., Baioni, D., 2005. Crustal extension in the
Ceraunius Fossae, northern Tharsis region, Mars. J. Geophys. Res. 110, E06006.
http://dx.doi.org/10.1029/2004JE002373.
Borraccini, F., Lanci, L., Wezel, F.C., 2006. Does a detachment level exist beneath the
Ceraunius Fossae? Insights from graben mapping and lost-area balancing
analysis. Planet. Space Sci. 54, 701?709.
Bradley, B.A., Sakimoto, S.E.H., Frey, H., Zimbelman, J.R., 2002. Medusae Fossae
Formation: New perspectives from Mars Global Surveyor. J. Geophys. Res. 107
(E8). http://dx.doi.org/10.1029/2001JE001537.
Bruno, B.C., Fagents, S.A., Pilger, E., Rowland, S., Garbeil, H., 2005. Lava ?ows on
Olympus Mons, Mars: Estimates of ?ow speeds and volume ?uxes from MOC,
THEMIS and MOLA data. Lunar Planet. Sci. 36. Abstract 2012.
Burr, D.M., Parker, A.H., 2006. Grjot? Valles and implications for ?ood sediment
deposition on Mars. Geophys. Res. Lett. 33, L22201. http://dx.doi.org/10.1029/
2006GL028011.
Burr, D.M., McEwen, A.S., Sakimoto, S.E.H., 2002a. Recent aqueous ?oods from the
Cerberus Fossae, Mars. Geophys. Res. Lett. 29 (1). http://dx.doi.org/10.1029/
2001GL013345.
Burr, D.M., Grier, J.A., McEwen, A.S., Keszthelyi, L.P., 2002b. Repeated aqueous
?ooding from the Cerberus Fossae: Evidence for very recently extant, deep
groundwater on Mars. Icarus 159, 53?73.
Butler, B.J., 1994. 3.5-cm radar investigation of Mars and Mercury: Planetological
implications. Ph.D. Thesis, California Institute of Technology, Pasadena.
Campbell, B.A., 2001. Radar backscatter from Mars: Properties of rock-strewn
surfaces. Icarus 150, 38?47.
Campbell, B.A., 2002. Radar Remote Sensing of Planetary Surfaces. Cambridge
University Press, Cambridge.
Campbell, D.B., Burns, B.A., 1980. Earth-based radar imagery of Venus. J. Geophys.
Res. 85, 8271?8281.
Campbell, B.A., Campbell, D.B., 1992. Analysis of volcanic surface morphology on
Venus from comparison of Arecibo, Magellan, and terrestrial airborne radar
data. J. Geophys. Res. 97 (E10), 16293?16314.
Campbell, M.J., Ulrichs, J., 1969. Electrical properties of rocks and their signi?cance
for lunar radar observations. J. Geophys. Res. 74, 5867?5881.
Campbell, B.A., Arvidson, R.E., Shepard, M.K., 1993. Radar polarization properties of
volcanic and playa surfaces: Applications to terrestrial remote sensing and
Venus data interpretation. J. Geophys. Res. 98, 17099?17113.
Campbell, B.A., Campbell, D.B., DeVries, C.H., 1999. Surface processes in the Venus
highlands: Results from analysis of Magellan and Arecibo data. J. Geophys. Res.
104, 1897?1916.
Campbell, B.A. et al., 2001. Orbital imaging radar and the search for water on Mars.
Conf. on Geophys. Detect. Subsurface Water on Mars. Abstract 7040.
Campbell, B.A., Maxwell, T.A., Freeman, A., 2004. Mars orbital synthetic aperture
radar: Obtaining geologic information from radar polarimetry. J. Geophys. Res.
109, E07008. http://dx.doi.org/10.1029/2004JE002264.
20 (2012) 990?1030 1027Campbell, B.A. et al., 2008. SHARAD radar sounding of the Vastitas Borealis
Formation in Amazonis Planitia. J. Geophys. Res. 113, E12010. http://dx.doi.org/
10.1029/2008JE003177.
rus 2Campbell, B.A., Hawke, B.R., Carter, L.M., Ghent, R.R., Campbell, D.B., 2009. Rugged
lava ?ows on the Moon revealed by Earth-based radar. Geophys. Res. Lett. 36,
L22201. http://dx.doi.org/10.1029/2009GL041087.
Campbell, B.A. et al., 2010. Earth-based 12.6-cm wavelength radar mapping of the
Moon: New views of impact melt distribution and mare physical properties.
Icarus 208, 565?573.
Caprarelli, G., Leitch, E.C., 2009. Volcanic and structural history of the rocks exposed
at Pickering Crater (Daedalia Planum, Mars). Icarus 202, 453?461.
Carr, M.H., Greeley, R., Blasius, K.R., Guest, J.E., Murray, J.B., 1977. Some volcanic
features as viewed from the Viking Orbiters. J. Geophys. Res. 82, 3985?4015.
Carter, L.M. et al., 2009. Dielectric properties of lava ?ows west of Ascraeus Mons,
Mars. Geophys. Res. Lett. 36, L23204. http://dx.doi.org/10.1029/2009GL041234.
Cattermole, P, 1988. Mapping of volcanic units at Alba Patera, Mars. METV
Workshop on Nature and Composition of Surface Units on Mars. Lunar Planet.
Inst., pp. 37?39.
Cattermole, P., 1990. Volcanic ?ow development at Alba Patera, Mars. Icarus 83,
453?493.
Chapman, M.G. et al., 2010. Amazonian geologic history of the Echus Chasma and
Kasei Valles system on Mars: New data and interpretations. Earth Planet. Sci.
Lett. 294, 238?255.
Christensen, P.R. et al., 2004. The Thermal Emission System (THEMIS) for the Mars
2001 Odyssey Mission. Space Sci. Rev. 110, 85?130.
Christensen, P.R. et al., 2005. Mars Exploration Rover candidate landing sites as
viewed by THEMIS. Icarus 176, 12?43.
Crown, D.A., Berman, D.C., 2012. Geologic mapping of MTM-35137 quadrangle:
Daedalia Planum region of Mars. Lunar Planet. Sci. 43. Abstract 2055.
Crown, D.A., Berman, D.C., Rivas, R., Ramsey, M.S., 2009. Arsia Mons lava ?ows:
Insights into ?ow ?eld emplacement and stratigraphy from CTX and HiRISE
images. Lunar Planet. Sci. 40. Abstract 2252.
Crown, D.A., Ramsey, M.S., Berman, D.C., 2010. Mapping Arsia Mons lava ?ow ?elds:
Insights into ?ow ?eld emplacement processes and ?ow ?eld development.
Lunar Planet. Sci. 41. Abstract 2225.
Crown, D.A., Ramsey, M.S., Berman, D.C., 2012. Morphologic and chronologic studies
of lava ?ow ?elds in the southern Tharsis region of Mars. Lunar Planet. Sci. 43.
Abstract 2138.
Crumpler, L.S., Auberle, J.C., 1978. Structural evolution of Arsia Mons, Pavonis Mons,
and Ascraeus Mons: Tharsis region of Mars. Icarus 34, 496?511.
Crumpler, L.S., Head, J.W., Auberle, J.C., 1996. Calderas on Mars: Characteristics,
structure, and associated ?ank deformation. In: McGuire, W.J., Jones, A.P.,
Neuberg, J. (Eds.), Volcano Instability on the Earth and Other Planets, vol. 110.
Geol. Soc. Special Publ., pp. 307?348.
Edgett, K.S., 1990. Possible cinder cones near the summit of Pavonis Mons, Mars.
Lunar Planet. Sci. 21, 311?312.
Edgett, K.S., 1997. Aeolian dunes as evidence for explosive volcanism in the Tharsis
region of Mars. Icarus 130, 96?114.
Edgett, K.S., Butler, B.J., Zimbelman, J.R., Hamilton, V.E., 1997. Geologic context of
the Mars radar ??Stealth?? region in southwestern Tharsis. J. Geophys. Res. 102,
21545?21567.
Fuller, E.R., Head, J.W., 2002. Amazonis Planitia: The role of geologically recent
volcanism and sedimentation in the formation of the smoothest plains on Mars.
J. Geophys. Res. 107. http://dx.doi.org/10.1029/2002JE001842.
Garry, W.B., Zimbelman, J.R., Gregg, T.K.P., 2007. Morphology and emplacement of a
long channeled ?ow near Ascraeus Mons volcano, Mars. J. Geophys. Res. 112,
E08007. http://dx.doi.org/10.1029/2006JE002803.
Ghent, R.R., Leverington, D.W., Campbell, B.A., Hawke, B.R., Campbell, D.B., 2005.
Earth-based observations of radar-dark crater haloes on the Moon: Implications
for regolith properties. J. Geophys. Res. 110, E02005. http://dx.doi.org/10.1029/
2004JE002366.
Ghent, R.R., Gupta, V., Campbell, B.A., Ferguson, S.A., Brown, J.C.W., Fergason, R.L.,
Carter, L.M., 2010. Generation and emplacement of ?ne-grained ejecta in
planetary impacts. Icarus 209, 818?835.
Giacomini, L., Massironi, M., Martellato, E., Pasquar?, G., Frigeri, A., Cremonese, G.,
2009. In?ated ?ows on Daedalia Planum (Mars)? Clues from a comparative
analysis with the Payen volcanic complex (Argentina). Planet. Space Sci. 57,
556?570.
Giacomini, L., Carli, C., Massironi, M., Sgavetti, M., 2010. Morphological and spectral
analysis for the Daedalia Planum geological mapping. Lunar Planet. Sci. 41.
Abstract 1594.
Graff, M.A., Zimbelman, J.R., 2012. A search for in?ated lava ?ows on Mars. Lunar
Planet. Sci. 43. Abstract 1144.
Greeley, R., Guest, J.E., 1987. Geologic Map of the Eastern Equatorial Region of Mars.
USGS Map I-1802-B. US Geol. Surv., Reston, VA.
Greeley, R., Bleacher, J.E., Cave, S.R., Williams, D.A., Werner, S.C., Neukum, G., 2005.
Flank volcanism on Pavonis Mons, Mars revealed by the HRSC on Mars Express.
Bull. Am. Astron. Soc. 37, 656.
Greeley, R., Bleacher, J.E., Cave, S.R., Williams, D.A., Werner, S.C., Neukum, G., 2006.
Fields of small shield volcanoes in the Tharsis region: Mars Express HRSC
results. In: European Planet. Sci. Conf., Berlin, Germany, p. 355.
Hagfors, T., Kofman, W., 1991. Mapping of overspread targets in radar astronomy.
Radio Sci. 26, 403?416.
Hamilton, C.W., Fagents, S.A., Thordarson, T., 2011. Lava-ground ice interactions in
Elysium Planitia, Mars: Geomorphological and geospatial analysis of the
Tartarus Colles cone groups. J. Geophys. Res. 116, E03004. http://dx.doi.org/
1028 J.K. Harmon et al. / Ica10.1029/2010JE003657.
Hanley, D., Zimbelman, J.R., 1995. Topographic control of lava ?ow emplacement:
Earth, Mars, and Venus. Lunar Planet. Sci. 26, 545?546.Harmon, J.K., 1997. A radar study of the Chryse region, Mars. J. Geophys. Res. 102,
4081?4095.
Harmon, J.K., 2002. Planetary delay-Doppler radar and the long-code method. IEEE
Trans. Geosci. Remote Sens. 40, 1904?1916.
Harmon, J.K., 2007. Radar imaging of Mercury. Space Sci. Rev. 132, 307?349.
Harmon, J.K., Ostro, S.J., 1985. Mars: Dual-polarization radar observations with
extended coverage. Icarus 62, 110?128.
Harmon, J.K., Nolan, M.C., 2007. Arecibo radar imaging of Mars during the 2005
opposition. Int. Mars Conf. 7. Abstract 3136.
Harmon, J.K., Campbell, D.B., Ostro, S.J., 1982. Dual-polarization radar observations
of Mars: Tharsis and environs. Icarus 52, 171?187.
Harmon, J.K., Slade, M.A., Hudson, R.S., 1992a. Mars radar scattering: Arecibo/
Goldstone results at 12.6-and 3.5-cm wavelengths. Icarus 98, 240?253.
Harmon, J.K., Sulzer, M.P., Perillat, P.J., Chandler, J.F., 1992b. Mars radar mapping:
Strong backscatter from the Elysium basin and out?ow channel. Icarus 95, 153?
156.
Harmon, J.K., Slade, M.A., Butler, B.J., Head, J.W., Rice, M.S., Campbell, D.B., 2007.
Mercury: Radar images of the equatorial and midlatitude zones. Icarus 187,
374?405.
Harmon, J.K., Arvidson, R.E., Guinness, E.A., Campbell, B.A., Slade, M.A., 1999. Mars
mapping with delay-Doppler radar. J. Geophys. Res. 104, 14065?14089.
Hauber, E., Bleacher, J., Gwinner, K., Williams, D., Greeley, R., 2009. The topography
and morphology of low shields and associated landforms of plains volcanism in
the Tharsis region. J. Volcanol. Geotherm. Res. 185, 69?95.
Hauber, E. et al., 2005. Discovery of a ?ank caldera and very young glacial activity at
Hecates Tholus, Mars. Nature 434, 356?361.
Head, J.W., Wilson, L., 1998. Tharsis Montes as composite volcanoes? 3: Lines of
evidence for explosive volcanism in edi?ce construction. Lunar Planet. Sci. 29.
Abstract 1124.
Head, J.W., Singer, R.B., Geissler, P.E., 1991. Multispectral study of Cerberus dark
materials. Bull. Am. Astron. Soc. 23, 1176.
Hiesinger, H., Head, J.W., Neukum, G., 2007. Young lava ?ows on the eastern ?ank of
Ascraeus Mons: Rheological properties derived from High Resolution Stereo
Camera (HRSC) images and Mars Orbiter Laser Altimeter (MOLA) data. J.
Geophys. Res. 112, E05011. http://dx.doi.org/10.1029/2006JE002717.
Hiesinger, H., Pasckert, J.H., Reiss, D., 2009. Rheology of lava ?ows on Elysium Mons,
Mars. Lunar Planet. Sci. 40. Abstract 1983.
Hodges, C.A., Moore, H.J., 1994. Atlas of Volcanic Landforms on Mars. USGS Prof.
Paper 1534, US Geol. Surv., Washington, DC.
Hurwitz, D.M., Head, J.W., 2012. Testing the late-stage out?ow channel origin
hypothesis: Investigating both water erosion and lava erosion origins for
Athabasca Valles, Mars. Lunar Planet. Sci. 43. Abstract 1056.
Ivanov, M.A., Head, J.W., 2006. Alba Patera, Mars: Topography, structure, and
evolution of a unique late Hesperian?early Amazonian shield volcano. J.
Geophys. Res. 111, E09003. http://dx.doi.org/10.1029/2005JE002469.
Ivanov, A.B., Muhleman, D.O., Vasavada, A.R., 1998. Microwave thermal mapping of
the Stealth region on Mars. Icarus 133, 163?173.
Jaeger, W.L., Keszthelyi, L.P., McEwen, A.S., Dundas, C.M., Russell, P.S., 2007.
Athabasca Valles, Mars: A lava-draped channel system. Science 317, 1709?1711.
Jaeger, W.L. et al., 2010. Emplacement of the youngest ?ood lava on Mars: A short,
turbulent story. Icarus 205, 230?243.
Kadish, S.J., Head, J.W., Parsons, R.L., Marchant, D.R., 2008. The Ascraeus Mons fan-
shaped deposit: Volcano?ice interactions and the climatic implications of cold-
based tropical mountain glaciation. Icarus 197, 84?109.
Kerber, L., Head, J.W., 2010. The age of the Medusae Fossae Formation: Evidence for
Hesperian emplacement from crater morphology, stratigraphy, and ancient lava
contacts. Icarus 206, 669?684.
Kerber, L., Head, J.W., Madeleine, J.-B., Forget, F., Wilson, L., 2011. The dispersal of
pyroclasts from Apollinaris Patera, Mars: Implications for the origin of the
Medusae Fossae Formation. Icarus 216, 212?220.
Keszthelyi, L., Grier, J.A., 2002. Distribution of small scale volcanic, mantling, and
impact features on Olympus Mons and the Tharsis volcanoes. Lunar Planet. Sci.
33. Abstract 1458.
Keszthelyi, L., McEwen, A.S., Thordarson, T., 2000. Terrestrial analogs and thermal
models for martian ?ood lavas. J. Geophys. Res. 105, 15027?15050.
Keszthelyi, L. et al., 2004. Icelandic analogs to martian ?ood lavas. Geochem.
Geophys. Geosyst. 5, Q11014. http://dx.doi.org/10.1029/2004GC000758.
Keszthelyi, L. et al., 2008. High Resolution Imaging Science Experiment (HiRISE)
images of volcanic terrains from the ?rst 6 months of the Mars Reconnaissance
Orbiter Primary Science Phase. J. Geophys. Res. 113, E04005. http://dx.doi.org/
10.1029/2007JE002968.
Keszthelyi, L.P., Jaeger, W.L., Dundas, C.M., Mart??nez-Alonso, S., McEwen, A.S.,
Milazzo, M.P., 2010. Hydrovolcanic features on Mars. Preliminary observations
from the ?rst Mars year of HiRISE imaging. Icarus 205, 211?229.
Kossacki, K.J., Markiewicz, W.J., Smith, M.D., Page, D., Murray, J., 2006. Possible
remnants of a frozen mud lake in southern Elysium, Mars. Icarus 181, 363?374.
Lanagan, P.D., 2004. Geologic History of the Cerberus Plains, Mars. Ph.D. Thesis,
University of Arizona, Tucson.
Lanagan, P.D., McEwen, A.S., Keszthelyi, L.P., Thordarson, T., 2001. Rootless cones on
Mars indicating the presence of shallow equatorial ground ice in recent times.
Geophys. Res. Lett. 28, 2365?2367.
Lee, S.W., Thomas, P.C., Veverka, J., 1982. Wind streaks in Tharsis and Elysium:
Implications for sediment transport by slope winds. J. Geophys. Res. 87, 10025?
20 (2012) 990?103010041.
Leverington, D.W., 2007. Was the Mangala Valles system incised by volcanic ?ows?
J. Geophys. Res. 112, E11005. http://dx.doi.org/10.1029/2007JE002896.
rus 220 (2012) 990?1030 1029Mangold, N., Ansan, V., Masson, P., Quantin, C., Neukum, G., 2008. Geomorphic
study of ?uvial landforms on the northern Valles Marineris plateau, Mars. J.
Geophys. Res. 113, E08009. http://dx.doi.org/10.1029/2007JE002985.
Mangold, N. et al., 2010. Mineralogy of recent volcanic plains in the Tharsis region,
Mars, and implications for platy-ridged ?ow composition. Earth Planet. Sci. Lett.
294, 440?450.
McEwen, A.S. et al., 2005. The rayed crater Zunil and interpretations of small impact
craters on Mars. Icarus 176, 351?381.
Milkovich, S.M., Head, J.W., 2003. Olympus Mons fan shaped deposit morphology:
Evidence for debris glaciers. Int. Mars Conf. 6. Abstract 3136.
Milkovich, S.M., Head, J.W., Marchant, D.R., 2006. Debris-covered piedmont glaciers
along the northwest ?ank of the Olympus Mons scarp: Evidence for low-
latitude ice accumulation during the Late Amazonian of Mars. Icarus 181, 388?
407.
Moore, H.J., Thompson, T.W., 1991. A radar-echo model for Mars. Proc. Lunar Sci.
Conf. 21, 457?472.
Mouginis-Mark, P.J., 1981. Late-stage summit activity of martian volcanoes. Lunar
Planet. Sci. 12, 726?728.
Mouginis-Mark, P.J., 1982a. Lava ?ows as slope indicators in the Tharsis region of
Mars. Lunar Planet. Sci. 13, 554?555.
Mouginis-Mark, P.J., 1982b. Late-stage summit activity of martian volcanoes. Proc.
Lunar Sci. Conf. 12, 1431?1447.
Mouginis-Mark, P.J., 2002. Prodigious ash deposits near the summit of Arsia Mons
volcano, Mars. Geophy. Res. Lett. 29 (16). http://dx.doi.org/10.1029/
2002GL015296.
Mouginis-Mark, P.J., 2012. A new 1:200,000-scale geologic map of Tooting Crater,
Mars. Lunar Planet. Sci. 43. Abstract 1562.
Mouginis-Mark, P.J., Christensen, P.R., 2005. New observations of volcanic features
on Mars from the THEMIS instrument. J. Geophys. Res. 110, E08007. http://
dx.doi.org/10.1029/JE002421.
Mouginis-Mark, P.J., Garbeil, H., 2007. Crater geometry and ejecta thickness of the
martian impact crater Tooting. Meteorit. Planet. Sci. 42, 1615?1625.
Mouginis-Mark, P.J., Yoshioka, M.T., 1998. The long lava ?ows of Elysium Planitia,
Mars. J. Geophys. Res. 103, 19389?19400.
Mouginis-Mark, P.J., Zisk, S.H., Downs, G.S., 1982a. Ancient and modern slopes in the
Tharsis region of Mars. Nature 297, 546?550.
Mouginis-Mark, P.J., Head, J.W., Wilson, L., 1982b. Explosive volcanism on Hecates
Tholus, Mars ? Investigation of eruption conditions. J. Geophys. Res. 87, 9890?
9904.
Mouginis-Mark, P.J., Wilson, L., Head, J.W., Brown, S.H., Hall, J.L., Sullivan, K.D., 1984.
Elysium Planitia: Regional geology, volcanology and evidence for volcano-
ground ice interactions. Earth Moon Planets 30, 149?173.
Mouginis-Mark, P.J., Wilson, L., Zimbelman, J.R., 1988. Polygenic eruptions on Alba
Patera, Mars. Bull. Volcanol. 50, 361?379.
Muhleman, D.O., Butler, B.J., Grossman, A.W., Slade, M.A., 1991. Radar images of
Mars. Science 253, 1508?1513.
Muhleman, D.O., Grossman, A.W., Butler, B., 1995. Radar investigation of Mars,
Mercury, and Titan. Annu. Rev. Earth Planet. Sci. 23, 337?374.
Murray, J.B. et al., 2005. Evidence from the Mars Express High Resolution Stereo
Camera for a frozen sea close to Mars? equator. Nature 434, 352?356.
Murray, J.B. et al., 2010. Late-stage water eruptions from Ascraeus Mons volcano,
Mars: Implications for its structure and history. Earth Planet. Sci. Lett. 294, 479?
491.
Paillou, P., Lasne, Y., Heggy, E., Mal?zieux, J.-M., Ruf??, G., 2006. A study of P-band
synthetic aperture radar applicability and performance for Mars exploration:
Imaging subsurface geology and detecting shallow moisture. J. Geophys. Res.
111, E06S11. http://dx.doi.org/10.1029/2005JE002528.
Platz, T. et al., 2009. Growth and destruction cycles and eruption styles at Tharsis
Tholus, Mars. Lunar Planet. Sci. 40. Abstract 1522.
Platz, T. et al., 2011. Vertical and lateral collapse of Tharsis Tholus, Mars. Earth
Planet. Sci. 305, 445?455.
Plaut, J.J., Anderson, S.W., Crown, D.A., Stofan, E.R., van Zyl, J.J., 2004. The unique
radar properties of silicic lava domes. J. Geophys. Res. 109, E03001. http://
dx.doi.org/10.1029/2002JE002017.
Plescia, J.B., 1990. Recent ?ood lavas in the Elysium region of Mars. Icarus 88, 465?
490.
Plescia, J.B., 1994. Geology of the small Tharsis volcanoes: Jovis Tholus, Ulysses
Patera, Biblis Patera, Mars. Icarus 111, 246?269.
Plescia, J.B., 2000. Geology of the Uranius Group volcanic constructs: Uranius Patera,
Ceraunius Tholus, and Uranius Tholus. Icarus 143, 376?396.
Plescia, J.B., 2003a. Tharsis Tholus: An unusual volcano. Icarus 165, 223?241.
Plescia, J.B., 2003b. Cerberus Fossae, Elysium, Mars: A source for lava and water.
Icarus 164, 79?95.
Plescia, J.B., 2004. Morphometric properties of martian volcanoes. J. Geophys. Res.
109, E03003. http://dx.doi.org/10.1029/2002JE002031.
Plescia, J.B., 2007. Volcanology of the Elysium volcanoes. Lunar Planet. Sci. 38.
Abstract 2140.
Preblich, B.S., McEwen, A.S., Studer, D.M., 2007. Mapping rays and secondary craters
from the martian crater Zunil. J. Geophys. Res. 112, E05006. http://dx.doi.org/
10.1029/2006JE002817.
Ramsey, M.S., Crown, D.A., 2010. Thermophysical and spectral variability of Arsia
Mons lava ?ows. Lunar Planet. Sci. 41. Abstract 1111.
Richardson, P.W., Bleacher, J.E., Glaze, L.S., Baloga, S.M., 2009. The relationship
J.K. Harmon et al. / Icabetween lava fans and tubes on Olympus Mons in the Tharsis region, Mars.
Lunar Planet. Sci. 40. Abstract 1527.Richardson, J.A., Bleacher, J.E., Baptista, A.R., 2010. Identi?cation of volcanic ridge in
northern Syria Planum, Mars: Constraint on geologic history of Syria. Lunar
Planet. Sci. 41. Abstract 1427.
Rogers, D., Christensen, P.R., 2002. Identi?cation of a spectrally and
thermophysically unique region in northern Amazonis Planitia, Mars: Surface
analysis using TES and THEMIS data. AGU (Fall Suppl.). Abstract P72A-0490.
Russell, P.S., Head, J.W., 2003. Elysium?Utopia ?ows as mega-lahars: A model of
dike intrusion, cryosphere cracking, and water-sediment release. J. Geophys.
Res. 108 (E6). http://dx.doi.org/10.1029/2002JE001995.
Schaber, G.G., 1980. Radar, visual and thermal characteristics of Mars: Rough planar
surfaces. Icarus 42, 159?184.
Schaber, G.G., Elachi, C., Farr, T.G., 1980. Remote sensing data of SP Mountain and SP
Lava Flow in north-central Arizona. Remote Sens. Environ. 9, 149?170.
Schaber, G.G. et al., 1992. Geology and distribution of impact craters on Venus:
What are they telling us? J. Geophys. Res. 97, 13257?13301.
Schierl, Z. et al., 2012. Origin of sinuous channels on the SW apron of Ascraeus Mons
and the surrounding plains, Mars. Lunar Planet. Sci. 43. Abstract 1602.
Scott, D.H., Allingham, J.W., 1976. Geologic Map of the Elysium Quadrangle of Mars.
USGS Map I-935. US Geol. Surv., Reston, VA.
Scott, D.H., Tanaka, K.L., 1981a. Map Showing Lava Flows in the Southeast Part of the
Amazonis Quadrangle of Mars. USGS Map I-1802-A. US Geol. Surv., Reston, VA.
Scott, D.H., Tanaka, K.L., 1981b. Mars: Paleostratigraphic restoration of buried
surfaces in Tharsis Montes. Icarus 45, 304?319.
Scott, D.H., Tanaka, K.L., 1982. Ignimbrites of Amazonis Planitia region of Mars. J.
Geophys. Res. 87, 1179?1190.
Scott, D.H., Tanaka, K.L., 1986. Geologic Map of the Western Equatorial Region of
Mars. USGS Map I-1280. US Geol. Surv., Reston, VA.
Scott, E.D., Wilson, L., 1999. Evidence for a sill emplacement event on the upper
?anks of the Ascraeus Mons shield volcano, Mars. J. Geophys. Res. 104, 27079?
27090.
Scott, D.H., Zimbelman, J.R., 1995. Geologic Map of Arsia Mons Volcano, Mars. USGS
Map I-2480. US Geol. Surv., Reston, VA.
Scott, D.H., Schaber, G.G., Horstman, K.C., Dial, A.L., Tanaka, K.L., 1981a. Map
showing Lava Flows in the Northwest Part of the Tharsis Quadrangle of Mars.
USGS Map I-1266. US Geol. Surv., Reston, VA.
Scott, D.H., Schaber, G.G., Horstman, K.C., Dial, A.L., Tanaka, K.L., 1981b. Map
Showing Lava Flows in the Southwest Part of the Tharsis Quadrangle of Mars.
USGS Map I-1268. US Geol. Surv., Reston, VA.
Shean, D.E., Head, J.W., Marchant, D.R., 2005. Origin and evolution of a cold-based
tropical mountain glacier on Mars: The Pavonis Mons fan-shaped deposit. J.
Geophys. Res. 110, E05001. http://dx.doi.org/10.1029/2004JE002360.
Shean, D.E., Head, J.W., Fastook, J.L., Marchant, D.R., 2007. Recent glaciation at high
elevations on Arsia Mons, Mars: Implications for the formation and evolution of
large tropical mountain glaciers. J. Geophys. Res. 112, E03004. http://dx.doi.org/
10.1029/2006JE002761.
Signorella, J.D. de Wet, A.P., Bleacher, J.E., Collins, A. Schierl, Z.P., Schwans, B., 2012.
Volcanic or ?uvial channels on Ascraeus Mons: Focus on the source area of
sinuous channels on the southeast rift apron. Lunar Planet. Sci. 43. Abstract
2773.
Silvestro, S., Pondrelli, M., 2007. Geological map of the Echus Chasma region (Mars)
by using HRSC (MEX) data. Mem. Soc. Astron. Italia Suppl. 11, 228?231.
Simpson, R.A., Harmon, J.K., Zisk, S.H., Thompson, T.W., Muhleman, D.O., 1992.
Radar determination of Mars surface properties. In: Kieffer, H.H., Jakosky, B.,
Snyder, C.W., Matthews, M.S. (Eds.), Mars. Univ. of Arizona Press, London, pp.
652?685.
Smith, D., Neumann, G., Arvidson, R.E., Guinness, E.A., Slavney, S., 2003. Mars Global
Surveyor Laser Altimeter Mission Experimental Gridded Data Record. NASA
Planetary Data System, MGS-M-MOLA-5-MEGDR-L3-V1.0.
Sulzer, M.P., 1986. A radar technique for high range resolution incoherent scatter
autocorrelation function measurements utilizing the full average power of
klystron radars. Radio Sci. 21, 1033?1040.
Tanaka, K.L., 1990. Tectonic history of the Alba Patera?Ceraunius Fossae region of
Mars. Proc. Lunar Sci. Conf. 20, 515?523.
Tanaka, K.L. et al., 2011. Completing the new global geologic map of Mars. Lunar
Planet. Sci. 42. Abstract 2265.
Tanaka, K.L., Carr, M.H., Skinner, J.A., Gilmore, M.S., Hare, T.M., 2003a. Geology of the
MER 2003 ??Elysium?? candidate landing site in southeastern Utopia Planitia,
Mars. J. Geophys. Res. 108 (E12). http://dx.doi.org/10.1029/2003JE002054.
Tanaka, K.L., Skinner, J.A., Hare, T.M., Joyal, T., Wenker, A., 2003b. Resurfacing
history of the northern plains of Mars based on geologic mapping of Mars
Global Surveyor data. J. Geophys. Res. 108 (E4). http://dx.doi.org/10.1029/
2002JE001908.
Tanaka, K.L., Skinner, J.A., Hare, T.M., 2005. Geologic Map of the Northern Plains of
Mars. USGS Map 2888, US Geol. Surv., Denver, CO.
Thompson, T.W., 1987. High-resolution lunar radar map at 70-cm wavelength.
Earth Moon Planets 37, 59?70.
Thompson, T.W., Moore, H.J., 1989. A model for depolarized radar echoes from Mars.
Proc. Lunar Sci. Conf. 19, 409?422.
Tornabene, L.L. et al., 2006. Identi?cation of large (2?10 km) rayed craters on Mars
in THEMIS thermal infrared images: Implications for possible martian meteorite
source regions. J. Geophys. Res. 111, E10006. http://dx.doi.org/10.1029/
2005JE002600.
Toyota, T., Kurita, K., Spica, A., 2011. Distribution and time-variation of spike spears
at Pavonis Mons on Mars. Planet. Space Sci. 59, 672?682.
Vaucher, J., Baratoux, D., Mangold, N., Pinet, P., Kurita, K., Gr?goire, M., 2009a. The
volcanic history of central Elysium Planitia: Implications for martian
magmatism. Icarus 204, 418?442.
Vaucher, J., Baratoux, D., Toplis, M.J., Pinet, P., Mangold, N., Kurita, K., 2009b. The
morphologies of volcanic landforms at Central Elysium Planitia: Evidence for
recent and ?uid lavas on Mars. Icarus 200, 39?51.
Veverka, J., Sagan, C., Greeley, R., 1976. Variable features on Mars: VI ? An unusual
crater streak in Mesogaea. Icarus 27, 241?253.
Watters, T.R. et al., 2007. Radar sounding of the Medusae Fossae Formation Mars:
Equatorial ice or dry, low-density deposits. Science 318, 1125?1128.
Wong, M.P., Sakimoto, S.E.H., Garvin, J.B., 2001. MOLA topography of small
volcanoes in Tempe Terra and Ceraunius Fossae, Mars: Implications for
eruptive style. Lunar Planet. Sci. 32. Abstract 1563.
Wood, C.A., 1979. Cinder cones on Earth, Moon and Mars. Lunar Planet. Sci. 10,
1370?1372.
Zealey, W.J., 2007. Glacial, Periglacial And Glacio-Volcanic Structures on the Echus
Plateau. Int. Mars Conf. 7. Abstract 3002.
Zimbelman, J.R., 1984. Geologic Interpretation of Remote Sensing Data for the
Martian Volcano Ascraeus Mons. Ph.D. Thesis, Arizona State Univ., Tempe, AZ.
Zimbelman, J.R., 1985a. Surface properties of Ascraeus Mons: Dust deposits on a
Tharsis volcano. Lunar Planet. Sci. 16, 934?935.
Zimbelman, J.R., 1985b. Estimates for rheological properties for ?ows on the
martian volcano Ascraeus Mons. J. Geophys. Res. Suppl. 90, D157?D162.
Zimbelman, J.R., Edgett, K.S., 1992a. Volcanic and modi?ed landforms on the Tharsis
Montes, Mars. Lunar Planet. Sci. 23, 1581?1582.
Zimbelman, J.R., Edgett, K.S., 1992b. The Tharsis Montes, Mars: Comparison of
volcanic and modi?ed landforms. Proc. Lunar Sci. Conf. 22, 31?44.
Zimbelman, J.R., Greeley, R., 1983. Surface properties of Ascraeus Mons obtained
from IRTM data. Lunar Planet. Sci. 14, 881?882.
Zimbelman, J.R., McAllister, R., 1985. Surface morphology on the martian volcano
Ascraeus Mons. Lunar Planet. Sci. 16, 936?937.
Zimbelman, J.R., Mouginis-Mark, P.J., 1987a. A possible volcanic component in the
?ne-grain materials near Alba Patera, Mars. Lunar Planet. Sci. 18, 1126?1127.
Zimbelman, J.R., Mouginis-Mark, P.J., 1987b. Channels on Alba Patera, Mars:
Evidence for polygenic eruptions. Lunar Planet. Sci. 18, 672?673.
Zimbelman, J.R., McBride, K.M., 1989. Possible pyroclastic deposit near Elysium
Mons, Mars. Lunar Planet Sci. 20, 1243?1244.
Zisk, S.H., Pettengill, G.H., Catuna, G.W., 1974. High-resolution radar map of the
lunar surface at 3.8-cm wavelength. Moon 10, 17?50.
1030 J.K. Harmon et al. / Icarus 220 (2012) 990?1030