Icarus 283 (2017) 138–145
Contents lists available at ScienceDirect
Icarus
journal homepage: www.elsevier.com/locate/icarus
Low-amplitude topographic features and textures on the Moon:
Initial results from detrended Lunar Orbiter Laser Altimeter (LOLA)
topography
Mikhail A. Kreslavsky a,b,∗, James W. Head c, Gregory A. Neumann d, Maria T. Zuber e,
David E. Smith e
a
Earth and Planetary Sciences, University of California, Santa Cruz, CA 95064, USA
MExLab, Moscow State University of Geodesy and Cartography (MIIGAiK), Moscow, Russia
c
Earth, Environmental and Planetary Sciences, Brown University, Providence, RI 02912, USA
d
Solar System Exploration Division, NASA Goddard Space Flight Center, Greenbelt, MD, USA
e
Department of Earth, Atmospheric and Planetary Sciences, MIT, Cambridge, MA, USA
b
a r t i c l e
i n f o
Article history:
Received 29 January 2016
Revised 26 June 2016
Accepted 24 July 2016
Available online 2 August 2016
Keywords:
Moon, surface
Volcanism
Tectonics
Data reduction techniques
a b s t r a c t
Global lunar topographic data derived from ranging measurements by the Lunar Oribter Laser Altimeter (LOLA) onboard LRO mission to the Moon have extremely high vertical precision. We use detrended
topography as a means for utilization of this precision in geomorphological analysis. The detrended topography was calculated as a difference between actual topography and a trend surface defined as a
median topography in a circular sliding window. We found that despite complicated distortions caused
by the non-linear nature of the detrending procedure, visual inspection of these data facilitates identification of low-amplitude gently-sloping geomorphic features. We present specific examples of patterns of
lava flows forming the lunar maria and revealing compound flow fields, a new class of lava flow complex
on the Moon. We also highlight the identification of linear tectonic features that otherwise are obscured
in the images and topographic data processed in a more traditional manner.
© 2016 Published by Elsevier Inc.
1. Introduction
Topography has always been an important tool in geomorphologic analysis, both terrestrial and planetary. During the last two
decades new topographic data with exceptionally high vertical
precision has become available: airborne lidar data (e.g., Tarolli,
2014 and references therein) for the surface of the Earth and orbital laser altimeter data for the Moon (Smith et al., 2010a, 2010b),
Mars (Smith et al., 2001), and Mercury (Zuber et al., 2012). These
data have revolutionized Earth and planetary geomorphology.
Despite great advances in computer data processing, visual
inspection remains the main analysis tool in geomorphology. Laser
altimeter data are, in a sense, too precise for easy visual perception. For example, the ranging precision of the Lunar Orbiter Laser
Altimeter (LOLA) (Smith et al., 2010a, 2010b) onboard the Lunar
Reconnaissance Orbiter (LRO) mission to the Moon is about 10 cm,
and the topographic amplitudes on the Moon reach ∼10 km, which
gives an impressive dynamic range of ∼105 . Since the human eye
∗
Corresponding author at: Earth and Planetary Sciences, University of California,
Santa Cruz, CA 95064, USA. Fax: +1 831 459 3074.
E-mail address: [email protected] (M.A. Kreslavsky).
http://dx.doi.org/10.1016/j.icarus.2016.07.017
0019-1035/© 2016 Published by Elsevier Inc.
distinguishes only a few tens of shades of grey, it is obvious
that without special effort the accuracy of the data could remain
unused. There are many approaches to topography visualization.
For example, a widely used approach, a combination of simulated
shadowing and color-coding, allows good visual perception of
both large-amplitude large-scale features and smaller steep topographic features; however, such visualization still leaves some
topographic information, most specifically, low-amplitude gently
sloping features, beyond the limits of visual perception. Another
class of approaches to the analysis of topography is based on data
filtration; such methods make features of certain spatial scale
visually perceptible at the expense of features of larger and/or
smaller spatial scales. Removal of information about large spatial
scales is referred to as “detrending”, and the result is a “detrended
topography”.
In order to fully actualize the great potential of the LRO LOLA
data, we apply a particular detrending algorithm to the LOLA
topographic data for the Moon. This technique provides important
new information on previously unrecognized lava flow textures
and "stealth" tectonic structures in the lunar maria. We found that
for features with the smallest spatial dimension of a kilometer
and larger, detrended topography supersedes analysis of images
M.A. Kreslavsky et al. / Icarus 283 (2017) 138–145
139
obtained at low Sun illumination (for example, near the lunar
terminator; e.g., Lloyd and Head, 1972; Head and Lloyd, 1971,
1972a, 1972b, 1973), which has long been used for analysis of
low-amplitude gently sloping morphologic features. The same
detrending algorithm has been successfully applied to martian
topography by Kreslavsky and Head (2002), Head et al. (2002),
and Head and Marchant (2003).
In this paper we first describe the source data used, the
detrending algorithm, and the rationale for its choice; we then
consider the nature of the product, and illustrate the types of
artifacts that can appear in the data and that, if not considered,
can lead to misinterpretation. We conclude by showing selected
examples of textures and features revealed in the detrended LOLA
data.
2. Source data
As source data for detrended topographic maps, we use LRO
LOLA data in the form of raster topographic maps available from
the Planetary Data System (PDS), the so-called Gridded Data
Record - GDR in the PDS terminology. The data are available in
simple cylindrical projection for the entire lunar surface, and in
polar azimuthal stereographic projection for both high-latitude regions. The PDS contains two versions of the data, where elevation
for each grid element (pixel) is represented by either integer (with
0.5 m quantization) or by floating-point number. We used the
latter version to fully employ the high ranging precision of LOLA.
Elevation at each pixel had been calculated as an average of all
individual elevation measurements within the pixel; pixels with
no measurements were derived by interpolation between nearby
pixels (Smith et al., 2010a).
There is a range of maps with different discretization (pixel
size) in the PDS. Higher spatial resolution, which requires finer discretization, is desirable for interpretation; however, pixels obtained
by interpolation become dominant at very fine discretization. For
the simple cylindrical projection map we found the 64 pixels per
degree product to have the optimal discretization, the finest discretization, under which the majority of pixels do have data (21%
of all pixels had been obtained by interpolation in the data release
we used). This discretization corresponds to a 474 m pixel size at
the equator.
3. Detrending algorithm
The detrended elevation at each pixel was calculated as a difference between its actual elevation and the median elevation of
all pixels within a circular window centered at this pixel. The window radius defines the spatial scale of the features that are considered as a global trend and thus are subject to removal. Small windows favor detection of the weakest small-size features, however,
the amount of information removed from the data is huge and the
amount of remaining information is tiny. For larger windows, the
retained amount of information is greater, but features of interest
may become obscured by clutter. We produced maps with a 5, 10,
and 15 pixel window radii, which is equivalent to about a 5, 10,
and 15 km diameter window at the equator. In our calculations the
window radius was the same in pixels for the whole map and thus
prone to map projection distortion. The impact of this distortion
on visual perception of the map is minor, except for high latitudes;
the high latitudes, however, are covered by the polar stereographic
maps that have small distortion.
In a sense, the filtering algorithm used here is similar to
widely used high-pass linear filters. The difference is that we take
the median, while the linear high-pass filter takes the weighted
arithmetic mean in the sliding window. Any reasonable detrending
algorithm, including both our median-based and the traditional
Fig. 1. Artificial topography consisting of seven similar parabolic domes rendered
as artificially shaded relief (a), illumination from upper left, and results of the application of the linear high-pass filter (b) and median detrending procedure (c) rendered in a grayscale map, brighter shades denoting higher elevations. The grayscale
stretch is the same in (b) and (c). White circle between (b) and (c) shows the detrending window.
linear filters, perfectly filters out topographic features much larger
than the window, preserves topographic features much smaller
than the window, and inevitably distorts features comparable to
the window in size. In a sense, for features somewhat smaller
than the window, the median-based detrending produces less
distortion, than the linear filter, as illustrated in Fig. 1. This makes
the result of the median-based detrending much better for visual
perception than the result of a linear high-pass filter. This was the
principal reason for our choice of the median-based detrending
approach.
The essential advantage of the median-based detrending is accompanied by several shortcomings. Due to the non-linear nature
of the procedure, distortions of window-scale features produced
by the median-based detrending are more complicated; for example, unlike in the case of the linear high-pass filter, in this case, a
gentle tilt of the whole surface may alter the appearance of some
features. The median is more computationally intensive than linear averaging, and computational feasibility requires careful implementation of the algorithm.
4. Key to interpretation of the detrending results
Artificial examples of the detrending results are shown in
Figs. 1c and 2. Fig. 3 shows location of several examples of detrended topographic maps on the Moon. The maps themselves
are shown in Figs. 4, 5, 7 and 8; Figs. 6 and 9 shows low-Sun
photomosaics for the areas shown in Figs. 5 and 8 for comparison.
Grayscale rendition is the most suitable approach for visual
analysis of the detrended topographic maps. We use a linearly
stretched grayscale with brighter shades denoting topography
above the trend surface, and darker shades denoting topographic
lows. The use of inverse, “negative” coding is also possible and
might be useful in order to improve visual perception for some applications, as well as a non-linear stretch.
140
M.A. Kreslavsky et al. / Icarus 283 (2017) 138–145
Fig. 2. Artificial example of the median detrending result. Upper pane, sample topography rendered as artificially shaded relief, illuminated from the upper left.
Model topographic shapes include conical shields, parabolic domes, linear and sinuous scarps, and linear depressions. Lower pane, the result of detrending. White
circle in the upper left corner shows the detrending window.
Detrended topographic maps reveal objects and textures undetectable by other means. This is, however, made at the expense
of removal of a significant part of the information from the data.
This inevitably leads to some complications in data interpretation,
first of all, due to the appearance of artifacts. One of the purposes
of detrending is to use the data up to their limit of precision.
Not surprisingly, imperfections in the source data become visible
in the detrended maps. In Figs. 7 and 8, areas are displayed with
the strongest stretch (20 m in elevation with respect to the trend
surface from saturated black to saturated white); in the eastern,
smoother parts of both scenes, north–south lineations are readily
seen. These lines reflect individual LRO orbits and are caused by inaccuracy in the knowledge of the LRO spacecraft position; typical
high errors are about a meter.
Gaps in data coverage also introduce complications. Small (hundred meters to a kilometer) impact craters have a steep sizefrequency distribution, therefore the number of undersampled
craters is very large. The detrended maps are peppered with black
dots and short horizontal dashes introduced by the presence of
such craters. The horizontal (normal to the orbit tracks) dashes
appear to be due to pixels obtained by interpolation. Sometimes
LOLA laser shots sample an elevated crater rim(s) and miss the
crater cavity. In this case a white dot appears in the map. Such
cases are infrequent in the relative sense, but since the number of
craters is very large, it is easy to find such examples. Detrended
topography is an excellent data set for systematic survey of small
gentle volcanic shields in the lunar maria (e.g., Head and Gifford,
1980); however, some undersampled craters mimic such shields,
and therefore a specific cross-check with visual wavelength images
is required. Since the gaps between the orbit tracks are irregular
and sometimes rather wide, the size threshold of local features
that cannot be missed can be large, up to a few km.
As we mentioned above, features whose shortest spatial dimension is much shorter than the window diameter typically are
well represented in the map, while features whose dimensions are
larger than the window are filtered out without any trace. To aid
in the visual interpretation we show the detrending window in
Figs. 4, 5, 7 and 8. The features whose shortest dimension is comparable to the window are strongly distorted, and introduce potential difficulties in interpretation. In simple cases where there are
isolated features of the potentially misleading size, it is possible to
detect the distortion and take it into account. Figs. 1c and 2 show
a few synthetic model examples of such distortion.
Ideal linear scarps of any profile would be filtered out completely and disappear in these maps. In the real data this never occurs, because natural features are never perfectly linear and there
is always some noise in the data. To avoid the disappearance of
the linear scarps in the lower left corner of Fig. 2, we added some
white noise to the source topography. In the presence of such
noise, the slope breaks outlining the scarp appeared in the detrended topography.
Typical lunar highlands have many craters in the potentially
misleading 5–20 km size range. The size of central peaks and the
width of the walls of larger craters are also in the same range.
As a result, the potentially misleading features are not isolated,
and distortions in the highlands are significant and complicated.
Therefore, detrending of topography of the lunar highlands is not
as useful an approach as it is in the regionally flatter maria. All
of our examples analyzed in the subsequent Figures are either
from the lunar maria, where potentially misleading features are indeed isolated and do not pose a significant problem, or are from
the Imbrium basin-forming impact ejecta (Fra Mauro Formation)
that are much smoother at the scale of interest than the typical
highlands.
The detrending procedure itself does not tend to produce linear artifacts, except for those adjacent and parallel to real linear
topographic features, as shown in Fig. 2. Therefore any narrow lineament longer than ∼2–3 window diameters is almost certainly
a real morphological feature. Even in the highlands, despite all
the distortions, linear tectonic structures are very well seen in the
maps.
As mentioned above, the detrended topographic maps are
most useful for low-amplitude, gently sloping topographic features
among generally smooth surfaces, such as the lunar maria. A large
number of such features have been detected and studied with
low-Sun (high incidence angle) images. Such images have a significant advantage: their spatial resolution can be much higher
than the resolution of LOLA-derived topography, therefore, much
smaller features could be detected, and much more detailed studies are possible. However, low-Sun imaging has limitations. One of
them is caused by the very low obliquity of the lunar spin axis.
Except for the polar regions, low-Sun illumination is only available
from the east or from the west, which produces an azimuthal bias:
gentle north-south striking linear features will be easier to detect
and study than those with an east-west strike. Another problem
is related to the high amplitudes of major topographic features
on the Moon. For example, the incidence angle at which mare infill of deep basins can be imaged is limited by shadows from the
basin rim. An additional problem is caused by the small size of
the Moon. The incidence angle changes rapidly over the surface,
M.A. Kreslavsky et al. / Icarus 283 (2017) 138–145
141
Fig. 3. LROC WAC image mosaic for the lunar nearside in simple cylindrical projection with locations of detrended topographic maps shown in Figs. 4, 5, 7, and 8.
Fig. 4. Portion of the detrended topography map centered at 22ºN, 35º and
stretched to +/− 13 m amplitude. The detrending window size is 10 km (shown
with a circle below the grayscale bar). Long arrows, Rima Brayley; short arrows,
two examples of wrinkle ridges; B, crater Brayley, E, crater Euler.
and orbital imaging under proper illumination conditions requires
careful planning of observations and cumbersome mosaicking. For
sufficiently large features, the detrended topography is free of all
these problems.
5. Structures in the detrended topography maps
In this section, we consider appearance of the most frequently
occurring structures in lunar maria: wrinkle ridges, sinuous and
linear rilles. Wrinkle ridges are abundant in the maria and present
in each of areas shown in Figs. 4, 5, 7, 8. They are usually interpreted as shortening structures (e.g., Watters, 1988). Wrinkle
Ridges often have a diverse kilometer- and subkilometer-scale
morphology (Fig. 4, southwestern short arrow); however, some
of them are smooth arches often many kilometers wide (Fig. 4,
northeastern short arrow) and are better seen in detrended topography than in low-Sun photographs. Some ridges are so subtle
that it is impossible to identify them as wrinkle ridges umambiguously. The detrended topographic maps are excellent for wrinkle
ridge inventorization and mapping, because such maps facilitate
uniform objective identification of the ridges due to the absence of
illumination-related orientation bias. However, the typical width
of wrinkle ridges is about 10 km, in the potentially misleading
range, and their topography is significantly distorted. Therefore,
despite being very useful for wrinkle ridge identification, the maps
are not useful for analysis of their detailed structure.
Sinuous rilles (e.g., Hurwitz et al., 2013 and references therein)
are very well displayed in the detrended topography. Their shortest spatial dimension, width, usually does not exceed a few km,
meaning that they are smaller than the detrending window and
are therefore not distorted. Their significant lengths make them
readily identifiable by visual inspection, even if they are narrow
and somewhat undersampled by LOLA measurements. An excellent
example is Rima Brayley (Fig. 4, denoted with long arrows). Sinuous rilles that are even narrower are heavily undersampled and are
not recognizable in LOLA topography data. Although sinuous rilles
are usually readily seen in the orbital photographs, the detrended
topography maps provide uniform, illumination-independent identification, similarly to wrinkle ridges.
Linear rilles (McGill, 1971; Solomon and Head, 1980; Head and
Wilson, 1993; Petrycki et al., 2004; Klimczak, 2014) interpreted as
tectonic graben are also well seen in the new maps. Although some
linear rilles are wide and therefore their floor topography can be
distorted, their walls are narrow and therefore well preserved in
the detrending procedure. The length and linearity of the rilles favor their detection even amid rough, steep and complex terrain,
142
M.A. Kreslavsky et al. / Icarus 283 (2017) 138–145
Fig. 5. Portion of the detrended topography map centered at 9.8ºN, 9.2ºE and stretched to +/− 30 m amplitude. The detrending window size is 10 km (shown with a circle
next to the grayscale bar). Two pairs of long arrows show two linear depressions. Wide arrow points to an en-echelon junction of two linear rilles (apparently, graben),
which is not apparent in the images. L, Lacus Lenitatis; M, crater Manilius; H, Rima Hyginus; A, Rima Ariadaeus. Two boxes outlines areas shown in Fig. 6.
because the detrending procedure never produces long linear artifacts. The majority of the linear rilles are readily seen in the
low-Sun photomosaics. We found, however, a handful of east-westoriented linear depressions that probably are surface manifestation
of graben but are not easily identifiable in the images. Two examples of such "stealth" graben are shown in Fig. 5 with two pairs of
long arrows. In the images (Fig. 6) only a short western segment of
the northern feature can be identified and all other parts are not
seen.
6. Volcanic features and textures in the detrended topography
maps
The most impressive features seen in the detrended topography maps are patterns of lava flows forming mare surfaces. Three
very extensive partly overlapping lobate lava flows are well seen in
the central part of Mare Imbrium (Y1, Y2 and Y3 in Fig. 7). They
have been previously identified by Schaber (1973) with low-Sun
images that revealed the subtle topographic steps outlining the
flows, with the exception of their east-west-oriented segments. In
Fig. 7 the outlines are seen very well, the emplacement sequence
(Y1, Y2 and Y3) can be established unambiguously, and it is clearly
seen that these flows are the last flows in the center of Mare Imbrium. They appear to represent some of the latest stages of Imbrium basin infill. In the southern part of the scene, flow Y3 is
crossed by three subparallel wrinkle ridges; the flow was not deflected by the substantial ridge topography, therefore the ridge formation postdates the flow emplacement. The northeastern tip of
flow Y2 is located next to a wrinkle ridge; the forward flow could
have been stopped by the ridge, which would mean that formation
of wrinkle ridges occurred both before and after the formation of
lava flows (Schaber, 1973).
In addition to the previously known flows Y1, Y2 and Y3, the
detrended topography map reveals a number of other flows. For
example, the arrow in Fig. 7 shows a narrower flow that predates
Y1. It has a linear depression along its axis; this depression is about
one kilometer wide and a few meters deep; we interpret it as a
lava channel. Such flows are typical on Mars and at much smaller
scales on the Earth. Detailed mapping of these lunar flows with
the use of the detrended topography map has been carried out by
Qiao et al. (2016).
Volcanic plains in the western part of the scene (X in Fig. 7)
have a very distinctive pattern. Its characteristic topographic amplitude is 1–3 m, and a characteristic spatial scale is on the order of
several kilometers and larger. The pattern seems to be formed by
a hierarchic dendritic system of subtle ridges that we interpret as
a lava distributary system forming so-called compound flow fields
(Head and Wilson, 2016a, 2016b; Wilson and Head, 2016; Qiao
et al., 2016). Regardless of the details of its geometry and origin,
this pattern reflects a lava emplacement style and therefore eruption characteristics that are different from the previously described
flows. Individual lobes of flows X are cross-cut by the boundary
with flows Y, therefore flow field X predates flows Y.
Fig. 8 provides another example of different mare-forming volcanic styles in the western part of Mare Serenitatis. Long arrows
show a boundary between a completely smooth unit in the eastern part of the scene and a unit with a specific pattern with an
∼1–2 m topographic amplitude in the western part. The pattern
details are not readily discernable. This pattern might be similar
to pattern X (Fig. 7) in Mare Imbrium, but differ by a shorter characteristic spatial scale and a slightly lower topographic amplitude.
However, we again see clear evidence for volcanic units of different emplacement style and therefore different eruption characteristics. The sequence of emplacement of the smoother and rougher
units is difficult to decipher in this case. A narrow dark lane in
Fig. 8 at the contact on its eastern side indicates that the uphill
direction is toward the western, rougher unit; however, this can be
explained by both western-unit-forming flows superposed over a
M.A. Kreslavsky et al. / Icarus 283 (2017) 138–145
143
Fig. 6. Two portions of the LROC WAC global mosaic showing northern and southern parts of the scene in Fig. 5 at a larger scale. Arrows are duplicated from Fig. 5. L, Lacus
Lenitatis; M, crater Manilius; C, crater Manilius C; K, crater Manilius K; Z, Hyginus Z; H, Rima Hyginus; A, Rima Ariadaeus.
pre-existing flat eastern unit, and by embayment of pre-existing
western units by thin smooth lava sheets forming the eastern
unit.
The detrended topography maps are good for systematic identification of low volcanic constructs, domes and shields, starting at
a few km in diameter. Again, the maps have the advantage of uniform, illumination-independent detection. Some shields have very
gentle slopes and are poorly seen in the images, but can be distinguished on the detrended topography maps. In the northwestern
part of Fig. 8, two pairs of arrows point to two large gently sloping volcanic shields. The centers of each of these two shields are
spatially associated with narrow linear rilles; images of their centers reveal a few small steep cones and pits that might represent
vents. Despite these similarities, the north-western shield is well
seen in the low-Sun image (Fig. 9), while the south–eastern one is
not detectable.
7. Conclusions
We demonstrated that the detrended topographic maps are
a useful tool for visualizing high-precision topographic data. Although such maps can only be used in conjunction with other
data sets and other forms of presentation of topographic data,
they permit the identification of gently sloping features of low
topographic amplitude in the lunar plains. In particular, the maps
reveal widely varying patterns of maria-forming lava flows and
several examples of "stealth" graben that have not been seen
previously in the images. The maps can assist surveys of small
144
M.A. Kreslavsky et al. / Icarus 283 (2017) 138–145
Fig. 7. Portion of the detrended topography map covering the southern part of
Mare Imbrium, centered at 35ºN, 19ºW, and stretched to +/− 10 m amplitude. The
detrending window size is 10 km (shown with a circle below the grayscale bar).
Arrow shows a linear depression, probably a lava channel. X, Y1, Y2, Y3 mark lava
flows discussed in the text. Craters: C, Carlini, H, Helicon, V, Le Verrier.
Fig. 9. Portion of low-Sun LROC WAC regional mosaic showing the north–western
part of the scene in Fig. 8. Two pairs of short arrows are duplicated from Fig.
8 and show two low shields. Only the north-western shield is distinguishable in this
image.
Acknowledgments
We acknowledge financial support from the NASA Lunar Reconnaissance Orbiter Lunar Orbiter Laser Altimeter (LOLA) experiment
(NNX09AM54G and NNX11AK29G to JWH) and the NASA Solar System Exploration Research Virtual Institute (SSERVI) grant for Evolution and Environment of Exploration Destinations under cooperative Agreement No. NNA14AB01A at Brown University. All work on
data processing and analysis was carried out by MAK at MIIGAiK
under Russian Science Foundation support, project 14-22-00197.
References
Fig. 8. Portion of the detrended topography map covering western half of Mare
Serenitatis, centered at 27ºN, 14ºE, and stretched to +/− 10 m amplitude. The detrending window size is 10 km (shown with a circle above the grayscale bar). Two
pairs of short arrows show two low shields. Long arrows mark the boundary between volcanic plane units with different textures. Box outlines area shown in
Fig. 9.
and large volcanic shields, wrinkle ridges, linear rilles (graben)
and sinuous rilles, with the advantage of globally uniform,
illumination-independent detection. We are currently completing
the global data set and will make this available for independent
analysis at http://www.planetary.brown.edu/html_pages/data.htm
upon completion of the project.
Head, J.W., Gifford, A., 1980. Lunar mare domes: Classification and modes of origin.
Moon Planets 22, 235–258.
Head, J.W., Lloyd, D.D. 1971. Near-Terminator Photography, Apollo 14 Preliminary
Science Report, NASA SP-272, 297–300.
Head, J.W., Lloyd, D.D. 1972a. Near-Terminator Photography, Apollo 15 Preliminary
Science Report, NASA SP-289, 25–5–25-101.
Head, J.W., Lloyd, D.D. 1972b. Low-Relief Features in Terrain of the Descartes Region
and Other Areas: Near-Terminator Photography, Apollo 16 Preliminary Science
Report, NASA SP-315, 29–7–29-103.
Head, J.W., Lloyd, D.D. 1973. Near-Terminator and Earthshine Photography, Apollo 17
Preliminary Science Report, NASA SP-330, 4–33-4-39.
Head, J.W., Marchant, D.R., 2003. Cold-based mountain glaciers on Mars: Western
Arsia Mons. Geology 31, 641–644.
Head, J.W., Wilson, L., 1993. Lunar graben formation due to near-surface deformation accompanying dike emplacement. Planet. Space Sci. 41, 719–727.
Head, J., Wilson, L., 2016a. Generation, ascent and eruption of magma on the
Moon: New insights into source depths, magma supply, intrusions and effusive/explosive eruptions (Part 2: Predicted emplacement processes and observations). Icarus doi:10.1016/j.icarus.2016.05.031.
Head, J., Wilson, L., 2016b. Mare basalt volcanism: Generation, ascent, eruption and
history of secondary crust on the Moon. In: Proc. 47th Lunar Planet. Sci. Conf..
The Woodlands, TX, Abstract #1189.
Head, J.W., Kreslavsky, M.A., Pratt, S., 2002. Northern lowlands of Mars: Evidence for
widespread volcanic flooding and tectonic deformation in the Hesperian Period.
J. Geophys. Res. 107. doi:10.1029/20 0 0JE0 01445.
Hurwitz, D.M., Head, J.W., Hiesinger, H., 2013. Lunar sinuous rilles: Distribution,
characteristics, and implications for their origin. Planet. Space Sci. 79, 1–38.
doi:10.1016/j.pss.2012.10.019.
Klimczak, C., 2014. Geomorphology of lunar grabens requires igneous dikes at
depth. Geology 42, 963–966. doi:10.1130/G35984.1.
Kreslavsky, M.A., Head, J.W., 2002. Fate of outflow channel effluents in the northern lowlands of Mars: The Vastitas Borealis Formation as a sublimation
residue from frozen ponded bodies of water. J. Geophys. Res. 107. doi:10.1029/
20 01JE0 01831, CiteID 5121.
M.A. Kreslavsky et al. / Icarus 283 (2017) 138–145
Lloyd, D.D., Head, J.W., 1972. Lunar surface properties as determined from earthshine and near-terminator photography. In: Proc. Third Lunar Sci. Conf.,
Geochim. Cosmochim. Acta, pp. 3127–3142.
McGill, G.E., 1971. Attitude of fractures bounding straight and arcuate lunar rilles.
Icarus 14, 53–58.
Petrycki, J.A., Wilson, L., Head, J.W., 2004. The significance of the geometries of linear graben for the widths of shallow dike intrusions on the Moon. Lunar Planet.
Sci. 35 abstract 1123.
Qiao, L., Head, J.W., Wilson, L., Kreslavsky, M.A., Xiao, L., 2016. Compound Flow
Fields in Southwest Mare Imbrium: Geomorphology, Source Regions and Implications for Lunar Basin Filling. In: Proc. 47th Lunar Planet. Sci. Conf.. The Woodlands, TX Abstract # 2038.
Schaber, G.G., 1973. Lava flows in Mare Imbrium: Geologic evaluation from Apollo
orbital photography. In: Proc. 4th Lunar Planet. Sci. Conf., pp. 73–92.
Smith, D.E., Zuber, M.T., Frey, H.V., et al., 2001. Mars Orbiter Laser Altimeter: Experiment summary after the first year of global mapping of Mars. J. Geophys. Res.
106, 23689–23722.
145
Smith, D.E., Zuber, M.T., Neumann, G.A., et al., 2010a. Initial observations from the
Lunar Orbiter Laser Altimeter (LOLA). Geophys. Res. Lett. 37, L18204.
Smith, D.E., Zuber, M.T., Jackson, G.B., et al., 2010b. The lunar orbiter laser altimeter
investigation on the lunar reconnaissance orbiter mission. Space Sci. Rev. 150,
209–241.
Solomon, S.C., Head, J.W., 1980. Lunar mascon basins: Lava filling, tectonics, and
evolution of the lithosphere. Rev. Geophys. Space Phys. 18, 107–141.
Tarolli, P., 2014. High-resolution topography for understanding Earth surface processes: Opportunities and challenges. Geomorphology 216, 295–312.
Watters, T.R., 1988. Wrinkle ridge assemblages on the terrestrial planets. J. Geophys.
Res. 93, 10236–10254.
Wilson, L., Head, J., 2016. Generation, Ascent and Eruption of Magma on the
Moon: New Insights Into Source Depths, Magma Supply, Intrusions and Effusive/Explosive Eruptions (Part 1: Theory). Icarus doi:10.1016/j.icarus.2015.12.
039.
Zuber, M.T., Smith, D.E., Phillips, R.J., et al., 2012. Topography of the Northern Hemisphere of Mercury from MESSENGER Laser Altimetry. Science 336, 217.