Lobate scarps and the Martian crustal dichotomy
Thomas R. Watters
Center for Earth and Planetary Studies, National Air and Space Museum
Smithsonian Institution, Washington, D.C.
Mark S. Robinson
Department of Geological Sciences, Northwestern University, Evanston, Illinois
Abstract. Landforms reflecting crustal shortening are found in the ancient highlands of
the eastern hemisphere of Mars. These structures, referred to as lobate scarps, are
interpreted to be thrust faults. Lobate scarps occur near and are oriented roughly parallel
to the Martian crustal dichotomy, a major geologic and topographic boundary that divides
the heavily cratered highlands from the relatively smooth, featureless northern lowlands.
The long- and short-wavelength topography of lobate scarps in the northern Terra
Cimmeria?Amenthes region have been analyzed using photoclinometry, Earth-based radar
altimetry, and Mars Orbiter Laser Altimeter data. The measured relief of lobate scarps in
this region ranges from ;110 to 1230 m, and they occur on gentle regional slopes that dip
both toward and away from the dichotomy. Estimates of the horizontal shortening across
the lobate scarps studied range from roughly 0.24 to 2.6 km (n 5 9), assuming fault
plane dips of 258. The displacement-length (D-L) relationships of thrust faults associated
with the lobate scarps are consistent with those observed for terrestrial fault populations.
The compressional strain in the heavily cratered highlands near the dichotomy,
determined using the D-L data for the lobate scarps, is estimated to be ;0.17%.
Topographic data indicate that the dichotomy in the northern Terra Cimmeria?Amenthes
region has a distinct topographic signature. The spatial and temporal relationship of the
lobate scarps to the boundary suggests that they are related to its formation, supporting
models for a tectonic origin of the crustal dichotomy.
1. Introduction
The highlands of Mars have landforms described as lobate
scarps that are generally one-sided, are often lobate, and occur
in linear or arcuate segments. They are morphologically
similar to lobate scarps observed on Mercury [Watters, 1993;
Watters et al., 1998]. The fact that many Martian and Mer-
curian lobate scarps clearly deform and offset crater floors
and walls supports the interpretation that these structures
are compressional tectonic features resulting from thrust
faulting (Figures 1 and 2) [Strom et al., 1975; Cordell and
Strom, 1977; Melosh and McKinnon, 1988; Watters, 1993;
Watters et al., 1998].
Although the Tharsis dominated western hemisphere is the
most prominent tectonic center on Mars, the eastern hemi-
sphere has also experienced major tectonic events. Lobate
scarps in highland materials of the eastern hemisphere record
significant compressional deformation of some of the oldest
terrain on Mars [see Tanaka, 1986] and account for ;18%
of the total cumulative length of compressional structures
on the planet [Watters, 1993]. Martian lobate scarps, like the
analogous structures on Mercury, appear to occur on at least
two different length scales, here described as moderate and
large scale. Examples are found in the heavily cratered high-
lands of Amenthes and northern Terra Cimmeria (Figure
1). Amenthes Rupes (Figure 2) is one of the best preserved
large-scale scarps on the planet. It exhibits greater relief and
is longer than the moderate-scale scarps found in northern
Terra Cimmeria (Figure 3). The lobate scarps in this region
are about 300 to 400 km southwest of the Martian crustal
dichotomy, a geologic boundary between the southern
heavily cratered highlands and the relatively featureless
northern lowlands. The orientations of the lobate scarps
parallel that of the steep structural and/or erosional scarp
that marks the crustal dichotomy (Figure 1). These lobate
scarps are also radial to the Isidis basin, and it has been sug-
gested that they are related to the formation of Isidis [Wich-
man and Schultz, 1989]. The close proximity and parallel ori-
entation of the lobate scarps to the dichotomy boundary,
however, strongly suggest that they may be related to the for-
mation of the crustal dichotomy and may thus be significant in
constraining models for its origin.
We present the results of a study of the short- and long-
wavelength topography of lobate scarps in the northern
Terra Cimmeria?Amenthes region through photoclinomet-
ric analyses, Earth-based radar altimetry, and Mars Orbiter
Laser Altimeter (MOLA) data. A kinematic model for the
formation of lobate scarps that involves thrust faults is used
to estimate displacement and horizontal shortening. The
displacement-length (D-L) relations of the faults associated
with the lobate scarps is determined and compared to D-L
data of terrestrial faults. The D-L data are also used to
estimate the compressional strain recorded by the lobate
scarps in the highlands near the dichotomy. Finally, the rela-
tionship between the lobate scarps in the northern Terra Cim-
meria?Amenthes region and the origin of the crustal dichot-
omy is discussed.
Copyright 1999 by the American Geophysical Union.
Paper number 1998JE001007.
0148-0227/99/1998JE001007$09.00
JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 104, NO. E8, PAGES 18,981?18,990, AUGUST 25, 1999
18,981
2. Background
2.1. Mars Orbiter Laser Altimetry
Early results from MOLA, an instrument on the Mars
Global Surveyor, have already greatly contributed to our
knowledge of the detailed topography of Mars [Smith et al.,
1998; Zuber et al., 1998]. MOLA determines the elevation of
the surface within ;160-m footprints [Zuber et al., 1992]. The
data have a maximum vertical resolution (precision) of ;30
Figure 1. Viking Orbiter mosaic of the northern Terra Cimmeria?Amenthes region of Mars. The steep scarp
that marks the Martian crustal dichotomy between the southern heavily cratered highlands from the northern
lowlands strikes NW-SE (right side of mosaic). The black boxes indicate the locations of lobate scarps shown
in Figures 2 and 3. The white lines indicate the locations of Mars Orbiter Laser Altimeter (MOLA) and
Earth-based radar altimetry profiles shown in Figures 7, 8, and 10. The solid arrow indicates the location of
the profile (lobate scarp H) shown in Figure 6. This mosaic was generated using images from NASA [1991].
Figure 2. Viking Orbiter mosaic of Amenthes Rupes, possi-
bly the largest thrust fault scarp on Mars, is over a kilometer
high and over 400 km long. The line indicates the location of
the photoclinometric profiles (A?A9) shown in Figure 9. The
location of this mosaic is shown in Figure 1. This mosaic was
generated using images from NASA [1991].
Figure 3. Viking Orbiter mosaic of a region in northern
Terra Cimmeria. The area is dominated by a series of moder-
ate-scale lobate scarps with orientations that roughly parallel
the trend of the crustal dichotomy boundary. Arrows indicate
the locations of photoclinometric profiles shown in Figure 4,
and the location of this mosaic is shown in Figure 1. This
mosaic was generated using images from NASA [1991].
WATTERS AND ROBINSON: LOBATE SCARPS AND THE DICHOTOMY18,982
cm, an absolute vertical accuracy of ;30 m, and along-track
spatial resolution of 300 to 400 m [Smith et al., 1998]. The
across-track shot spacing obtained during the nominal map-
ping mission will depend on the mapping orbit and will vary
with latitude [see Zuber et al., 1992]. Two of the available
tracks of data cross the northern Terra Cimmeria?Amenthes
region (Figure 1) [see Smith et al., 1998, Figure 1; Frey et al.,
1998].
2.2. Earth-Based Radar Altimetry
Another source of topography is Earth-based radar altim-
etry [Roth et al., 1980; Downs et al., 1982]. The elevation of a
point on the surface is derived within a footprint that varies in
dimension with the geometry of the observation. The foot-
prints or resolution cells for Goldstone observations made
during Mars oppositions that occurred between 1973 and 1982
[Roth et al., 1980; Downs et al., 1982; Simpson et al., 1993] are
on average ;10 km wide (0.168 in longitude) and ;120 km
high (2.08 in latitude), an area of ;1200 km2 [see Downs et al.,
1982]. The vertical uncertainty in the radar altimetry for the
1978 and 1980 oppositions ranges from 50 to 280 m (for indi-
vidual points in Chryse and Amazonis Planitia) with an average
of ;160 m [see Downs et al., 1982, Table 1]. Although these
data are not sufficient to resolve the lobate scarps on Mars, the
radar altimetry does resolve the large-scale, long-wavelength
topographic variations in the highlands associated with the
lobate scarps.
2.3. Photoclinometry
Photoclinometry utilizes a function that describes the pho-
tometric properties of the surface to recover the slope and the
relative elevation between adjacent pixels of a digital image
[Wildey, 1975; Howard et al., 1982; Davis and Soderblom, 1984;
McEwen, 1991]. Photoclinometry is useful in obtaining short-
wavelength topographic data for relatively small landforms
with gentle slopes because the spatial resolution of the ex-
tracted topography is only limited by the spatial resolution of
the image and because the method is sensitive to small changes
in slope [Davis and Soderblom, 1984; Tanaka and Davis, 1988;
Watters and Robinson, 1997]. We used the Minnaert photomet-
ric function because it is well suited for the surface of Mars
[Binder and Jones, 1972; Thorpe, 1973; Davis and Soderblom,
1984; Tanaka and Davis, 1988; McEwen, 1991].
Critical parameters in the asymmetric photoclinometric
method used here are the scattered light value (SLV) (some-
times referred to as the haze value) that corrects for light
scattered in the atmosphere and from surfaces outside the
pixel area and the horizontal digital number (HDN) (some-
times referred to as the flat field value) that represents the
brightness value of a horizontal surface in an image [Davis and
Soderblom, 1984; Tanaka and Davis, 1988]. The SLV is chosen
by finding the brightness values of resolved shadows in the
image. The HDN is selected through photointerpretation and
is thus subjective. A more complete discussion of the SLV and
HDN parameters and errors that result from their misestima-
tions is given by Watters and Robinson [1997]. An independent
check on the accuracy of photoclinometric elevations can be
made by comparing them to elevations obtained by shadow
measurements. In a study of graben in Syria Planum, Tanaka
and Davis [1988] found that photoclinometrically derived ele-
vations are within 10?15% of elevations determined by shadow
measurements. Mouginis-Mark and Robinson [1992] also found
good agreement between elevations derived from shadow mea-
surements and photoclinometry in their study of the Olympus
Mons caldera.
3. Results
3.1. Topography
Elevation profiles across seven moderate-scale scarps in
northern Terra Cimmeria and the large-scale scarp Amenthes
Rupes were obtained using photoclinometry. Profile lengths
were selected to be the minimum necessary to span the full
width of the lobate scarp in order to reduce uncertainties
introduced by albedo variations and image calibration errors
that scale with profile length. The SLV was chosen by exam-
ining pixels in prominent shadows cast by the walls of impact
craters near the scarps. The HDN was determined by taking
the average of a 9 3 9 array (81 pixels) of pixels located near
the profile endpoints where the surface was judged to be
roughly horizontal. An array of pixels is used in order to in-
crease the signal-to-noise ratio (averaging reduces the effects
of random and digitization noise and round off errors in cali-
bration files). Our analysis of photoclinometric profiles across
lobate scarps in northern Terra Cimmeria reveals that a vari-
ation of 62 of the estimated HDN (which ranges from 775 to
923) does not significantly change the shape of the profiles or
result in a change in the direction of slope of the scarp face
(Figure 4). Gently sloping surfaces are, however, the most
sensitive to small variations in HDN [Davis and Soderblom,
1984; Watters and Robinson, 1997].
Photoclinometric results for scarps studied in northern
Terra Cimmeria (Figure 4) indicate that the measured relief
ranges from 112 6 9 m to 315 6 24 m (error estimates on
measured relief are based on a confidence in the derived ele-
vations of 67.5% [see Tanaka and Davis, 1988]). Although the
photoclinometric profiles were located in an effort to measure
the maximum relief on the lobate scarps, this may not have
been achieved in every case. The topographic data indicate
that lobate scarps have a simple morphology consisting of a
steeply sloping scarp face and a gently sloping back scarp
(Figures 4 and 5). Most of the scarp faces occur on the SW side
of the structures. This suggests that fault planes dip to the NE,
toward the boundary (Figure 5). The maximum slope on the
scarp faces ranges from ;38 to 98 (Table 1).
MOLA data track 36 crosses northern Terra Cimmeria near
the lobate scarps analyzed using photoclinometry and also
crosses a moderate-scale, north-northwest trending lobate
scarp located to the south (Figure 1). This scarp has a mea-
sured relief of about 327 6 1 m (where the profile crosses it)
(Figure 6), consistent with the range of relief determined for
other moderate-scale lobate scarps in the region studied using
photoclinometry (Table 1). The morphology of this lobate
scarp, as reflected by the MOLA data, is also consistent with
lobate scarps studied using photoclinometry. The maximum
slope of the scarp face is ;48, within the range of other mod-
erate-scale lobate scarps studied. Like the other moderate-
scale scarps in the region, the slope face is on the SW side of
the structure. Unfortunately, a direct comparison of colocated
photoclinometric profiles was not possible because a suitable
Viking Orbiter image could not be found.
Data from MOLA track 36 (trending roughly N-S) indicate
that the long-wavelength topography near the lobate scarps
studied using photoclinometry (Figure 3) is relatively flat (Fig-
ure 7). Earth-based radar altimetry profiles (trending E-W)
that cross northern Terra Cimmeria also indicate the topogra-
18,983WATTERS AND ROBINSON: LOBATE SCARPS AND THE DICHOTOMY
Figure 4. Elevation profiles derived through photoclinom-
etry across lobate scarps in northern Terra Cimmeria. Profiles
were generated for three different values of the HDN (shown
in legends) holding all other parameters constant. The vertical
exaggeration is 15:1 for profiles A-D and F and G. The vertical
exaggeration is 12;1 for profile E. For profile A the SLV is 439,
for profile B the SLV is 405, and for profiles C, D, E, F, and G
the SLV is 468. The locations of the profiles are given in Table
1 and shown in Figure 3.
WATTERS AND ROBINSON: LOBATE SCARPS AND THE DICHOTOMY18,984
phy is relatively flat and gently sloping (Figure 8). The lobate
scarp crossed by MOLA track 36 is about 1 km higher in
elevation than the other lobate scarps in region (Figure 7). The
regional slope of the long-wavelength topography in both the
MOLA and the Earth-based radar altimetry data is toward the
dichotomy (Figures 7 and 8).
Amenthes Rupes is over 400 km long with a measured relief
of 1225 6 92 m and a maximum slope of ;178 (Figure 9). It is
comparable in scale to the Discovery Rupes on Mercury
[Watters et al., 1998]. The scarp face of Amenthes Rupes, like
the moderate-scale scarps of northern Terra Cimmeria, is on
the SW side of the structure, distal to the scarp that marks the
dichotomy boundary. Again, this suggests the fault plane dips
to the NE, toward the dichotomy boundary. Earth-based radar
altimetry profiles across Amenthes (trending E-W) indicate
that the scarp is superposed on a gentle westward regional
slope (;0.28), away from the dichotomy (Figure 10) [also see
Smith and Zuber, 1996].
3.2. Estimates of Horizontal Shortening and Displacement
To estimate horizontal shortening across Martian lobate
scarps, we use a kinematic model that involves thrust faults
that propagate upward and break the surface (Figure 5). The
amount of horizontal shortening is estimated by assuming that
it is a function of the dip of the fault plane and the displace-
ment on the fault. The two variables in estimating the short-
ening in this way are the relief of the lobate scarp (h) and the
fault plane dip (u). The displacement (D) necessary to restore
the topography to a planar surface is given by D 5 h/sin u, and
the horizontal shortening (S) is given by S 5 h/tan u. Of the
two variables, the most significant error in estimating the hor-
izontal shortening is from the uncertainty in the fault plane
dip.
Faulting occurs at fault plane dips (u) for which the tectonic
stress is a minimum. This minimum occurs when tan 2u 5 1/ms,
Figure 5. Schematic cross-sectional view of a lobate scarp.
Lobate scarps consist of two morphologic features, a steeply
sloping scarp face and a gently sloping back scarp. The pro-
posed kinematic model for the formation of lobate scarps
involves deformation over a buried thrust fault that propagates
upward and eventually breaks the surface.
Figure 6. Mars Orbiter Laser Altimeter profile across a lo-
bate scarp in northern Terra Cimmeria. The approximate lo-
cation of the profile is shown in Figure 1. Data was extracted
from track 36 located at ;2348W longitude [NASA, 1998].
Elevations are relative to a reference potential surface, and the
vertical exaggeration is 15:1.
Table 1. Lobate Scarp Dimensions and Estimates of Shortening and Displacement
Index
Viking
Image
Resolution,
m/pixel Latitude Longitude Relief, m Slope
S
Range, m
S
u 5 258, m
D
u 5 258, m
Terra Cimmeria
A 629A21 260 5.78 S 237.28W 315 6 24 78 450?866 676 6 51 745 6 56
B 629A21 260 5.48 S 236.68W 185 6 14 58 264?508 397 6 30 438 6 33
C 629A22 261 6.08 S 234.08W 112 6 9 38 160?308 240 6 17 265 6 20
D 629A22 261 4.28 S 235.48W 217 6 16 88 310?596 465 6 35 514 6 39
E 629A22 261 3.28 S 237.28W 303 6 23 88 433?833 650 6 49 717 6 54
F 629A22 261 5.38 S 234.28W 165 6 12 68 236?453 354 6 27 390 6 29
G 629A22 261 5.88 S 235.58W 194 6 15 98 277?533 416 6 31 459 6 34
H z z z z z z 9.48 S 235.48W 327 6 1 48 467?898 701 6 2 774 6 2
Amenthes Rupes
379S46 243 1.58 N 249.58W 1225 6 92 178 1750?3366 2627 6 197 2899 6 217
The relief of lobate scarps A?G and Amenthes Rupes was determined using photoclinometry, and the relief of lobate scarp H was determined
with MOLA data. It should be noted that photoclinometric profiles were located in an effort to measure the maximum relief on the lobate scarps;
however, this may not have been achieved in every case. Also, the relief of the lobate scarp measured using MOLA data is the maximum relief
where the track crosses the structure. Error estimates on the measured relief, shortening (u 5 258), and displacement (u 5 258) for lobate scarps
A?G and Amenthes Rupes are based on a confidence in the photoclinometrically derived elevations of 67.5% [see Tanaka and Davis, 1988]. The
shortening range for these scarps is based on the measured relief in the derived elevations for a range in HDN of 62 at u 5 208 and 358. For
scarp H the estimated error on the measured relief and shortening is based on a precision of the MOLA data of 61 m [see Smith et al., 1998],
and the range in shortening is based on the measured relief for u 5 208 and 358. The location of profiles A?G are shown in Figure 3 and the
location of profile H is shown in Figure 1.
18,985WATTERS AND ROBINSON: LOBATE SCARPS AND THE DICHOTOMY
where ms, by analog with ordinary sliding friction, is defined as
the coefficient of internal friction [see Jaeger and Cook, 1979;
Turcotte and Schubert, 1982]. Data obtained from laboratory
experiments on the maximum shear stress to initiate sliding as
a function of normal stress, for a variety of rock types, are best
fit by a maximum coefficient of static friction of 0.6 to 0.9 (best
fit ms 5 0.85) [Byerlee, 1978]. These data suggest that thrust
faults will form with dips ranging from ;248 to 308, ;258 for
ms 5 0.85. This theoretical range of fault plane dips is in good
agreement with field measurements of u for terrestrial thrust
faults that typically range between 208 and 258 [see Jaeger and
Cook, 1979]. A notable exception is the large-scale Wind River
thrust fault with a average u of 358 that extends to a depth of
36 km [Brewer et al., 1980]. In this analysis we therefore con-
servatively assume that the thrust faults associated with lobate
scarps will fall within a range in u of 208 to 358 and that the
optimum u is 258.
It is also assumed, in the absence of any data to the contrary,
that fault plane dips are uniform (i.e., linear, not curved). This
assumption is reasonable because there are terrestrial thrust
faults with uniform fault plane dips that do not significantly
steepen or shallow with depth. The Wind River thrust [Brewer
et al., 1980] and other thrust faults that cut the Precambrian
basement of the Rocky Mountain Foreland in Wyoming [Gries,
1983; Stone, 1985] are examples. Some thrust faults in this
region exhibit consistent fault plane dip angles in the Precam-
Figure 7. Mars Orbiter Laser Altimeter profiles across the
dichotomy in northern Terra Cimmeria. More than 2.5 km of
relief is indicated over the transition from the heavily cratered
highlands to the lowlands. The approximate locations of the
profiles are shown in Figure 1. Data were extracted from tracks
36 and 20 located at ;2348W and ;2278W longitude, respec-
tively [NASA, 1998]. The long arrow indicates the approximate
location of lobate scarps A?G (Table 1 and Figure 3), and the
short arrow indicates the location of lobate scarp H (Table 1
and Figure 1). Elevations are relative to a reference potential
surface, and the vertical exaggeration is 100:1.
Figure 8. Goldstone Earth-based radar altimetry profiles
across northern Terra Cimmeria and the crustal dichotomy.
The latitudes of the profiles are given in the legend, and the
approximate locations of the western halves of the profiles are
shown in Figure 1. The arrows indicate the approximate loca-
tion of the lobate scarps A?G (Table 1 and Figure 3). Altim-
etry is relative to the 6.1 mbar surface which is offset to the
MOLA data reference potential surface by approximately
21600 m on average [Smith and Zuber, 1998]. The vertical
exaggeration is 100:1.
Figure 9. Elevation profiles derived through photoclinom-
etry across Amenthes Rupes. Profiles were generated for three
different values of the HDN (shown in legend) holding all
other parameters constant (Viking Orbiter image 379S46, im-
age resolution 243 m/pixel). The SLV is 436. The location of
the profiles is shown in Figure 2. Vertical exaggeration is 8:1.
Figure 10. Goldstone Earth-based radar altimetry profiles
across the crustal dichotomy in the Amenthes region. The
arrow indicates the approximate location of Amenthes Rupes
shown in Figure 2. The latitude of the profiles are given in the
legend and their approximate locations are shown in Figure 1.
Altimetry is relative to the 6.1 mbar surface which is offset to
the MOLA data reference potential surface by approximately
21600 m on average [Smith and Zuber, 1998]. The vertical
exaggeration is 100:1.
WATTERS AND ROBINSON: LOBATE SCARPS AND THE DICHOTOMY18,986
brian basement that steepen upward only where they cut Pa-
leozoic sedimentary sequences [Stone, 1985]. The dips on the
thrust faults cited, excluding the Wind River thrust, cut the
Precambrian basement at angles between 258 and 308.
From our new topographic data and the assumptions out-
lined above, the horizontal shortening was estimated using the
measured h of the lobate scarps and a range in u (208 to 358).
Our results for the moderate-scale scarps of northern Terra
Cimmeria indicate that the amount of shortening ranges from
a few hundred meters to just under a kilometer (Table 1), with
an average of ;490 m at u 5 258 (n 5 8). The bulk horizontal
shortening across the lobate scarp belt shown in Figure 3 is
estimated to be roughly 2.0 to 2.9 km (based on traverses
through the highlands that intersect four to six scarps and
assuming an average shortening of 490 m per scarp). The
amount of horizontal shortening across Amenthes Rupes is of
the order of 1.8 to 3.4 km (;2.6 km at u 5 258) (see Table 1).
Thus the shortening across the moderate-scale lobate scarps in
northern Terra Cimmeria (shown in Figure 3) is roughly com-
parable to the shortening across Amenthes Rupes, suggesting
that a similar bulk compressional strain was generated across
the region by the deformation event. It should be noted that
the estimates given above assume that overthrusting (where
the fault block is translated over the fault ramp onto the flat)
is not a significant component of the total horizontal shorten-
ing. This assumption is supported through observations that no
significant offsets occur in crater rims crosscut by scarps. If
overthrusting is significant, our estimates of horizontal short-
ening may be only lower limits.
3.3. Comparison With Terrestrial Faults
There is growing evidence based on field observations of
terrestrial faults that indicates that a positive correlation exists
between the maximum displacement on a fault (D) and the
length of the fault trace (L) [Cowie and Scholz, 1992; Gillespie
et al., 1992; Dawers et al., 1993; Cartwright et al., 1995]. This
relationship also seems to hold for planetary faults [Schultz,
1995; Schultz and Fori, 1996; Schultz, 1997; Watters et al., 1997,
1998, 1999]. The exact nature of the relationship between D
and L is still being debated. Cowie and Scholz [1992] suggest
that the D-L relationship for continental faults is linear such
that D 5 gL and g is determined by rock type and tectonic
setting (Plate 1). It has also been suggested that there is a
power law relationship D 5 cLn, , where c is a constant related
to material properties with n ranging from 1 to 2 [see Gillespie
et al., 1992]. Regardless of its exact nature, a consistent scaling
relationship between D and L holds for all the fault types (i.e.,
normal, strike-slip, and thrust) in a wide variety of tectonic
settings and a wide range of length scales [Cowie and Scholz,
1992]. The ratio of displacement to fault length (g) ranges
between 100 and 1023 for terrestrial faults [Cowie and Scholz,
1992]. Some of the scatter in the D and L data probably
reflects the growth of faults by segment linkage where the
scaling characteristics change at different stages of fault evo-
lution [Cartwright et al., 1995, 1996; Dawers and Anders, 1995;
Wojtal, 1996; Moore and Schultz, 1999].
The value of g for the Martian lobate scarps (obtained by a
linear fit to D-L data with estimates of D based on u 5 258) is
6.2 3 1023 (n 5 9) (Plate 1). This is consistent with the values
of g of terrestrial fault populations [see Cowie and Scholz,
1992]. The displacements on the faults associated with lobate
scarps are about an order of magnitude lower than terrestrial
thrust faults (Plate 1); however, they fall within the same range
determined for Mercurian lobate scarps [Watters et al., 1998,
1999]. This is likely a reflection of the difference in tectonic
setting. Most terrestrial thrust faults occur in foreland fold and
thrust belts located at convergent plate margins, as is the case for
the thrust faults plotted in Plate 1 that occur in the foreland thrust
belt of the Canadian Rocky Mountains [Elliott, 1976a, b].
Plate 1. A log-log plot of maximum displacement as a function of fault length for terrestrial faults and thrust
faults, Amenthes Rupes, and eight other Martian lobate scarps (see legend). The data for terrestrial faults are
from nine different data sets (includes data for 29 thrust faults) [see Cowie and Scholz, 1992].
18,987WATTERS AND ROBINSON: LOBATE SCARPS AND THE DICHOTOMY
3.4. Estimates of Compressional Strain
The compressional strain reflected by the lobate scarps in
the northern Terra Cimmeria?Amenthes region can be esti-
mated using the displacement-length scaling relationship for
these structures. If the D-L scaling relationship of a fault
population is known, the strain can be calculated using fault
lengths alone [Scholz and Cowie, 1990; Cowie et al., 1993]. The
strain for large faults (L $ the maximum depth of faulting) is
given by
? 5
cos~u !
A O
k51
n
DkLk (1)
where u is the fault plane dip, A is the size of the survey area,
n is the total number of faults, and D 5 gL [Cowie et al.,
1993]. The value of g for the lobate scarps studied here (n 5
9), obtained by a linear fit to the D-L data with estimates of
D based on u 5 258, is 6.2 3 1023. The lengths of lobate scarps
(n 5 23) were measured in the heavily cratered highlands
(Npld) in the northern Terra Cimmeria?Amenthes region, ad-
jacent to the crustal dichotomy boundary (from 2258W to
2558W). The compressional strain in the heavily cratered high-
lands is estimated to be between 0.175% and 0.153% for a
range in u of 208 to 358 (0.169% for u 5 258).
4. Discussion
4.1. Origin of the Crustal Dichotomy
The origin of the dichotomy that divides the heavily cratered
highlands in the southern hemisphere from the younger, lightly
cratered lowlands of the northern hemisphere is one of the
most important unresolved questions about the geologic evo-
lution of Mars. Although it is generally agreed that the dichot-
omy is a geologic boundary, Smith and Zuber [1996] argue that
the apparent topographic difference between the northern and
southern hemisphere reflects a center of mass?center of figure
offset and that there is no clear topographic depression in the
northern hemisphere. It is difficult to reconcile this interpre-
tation with the topographic data from Earth-based radar al-
timetry and from MOLA [see Smith et al., 1998, Figure 1; Frey
et al., 1998]. Earth-based radar altimetry profiles that cross the
boundary in Amenthes (Figure 10) and northern Terra Cim-
meria (Figure 8) show an average elevation change of ;2.5
km. Initial results from MOLA confirm these measurements
(Figure 7) and show that the regional elevation change is .2.5
km and up to 6 km in some areas [Frey et al., 1998; Smith et al.,
1998]. Thus in the northern Terra Cimmeria?Amenthes re-
gion, and elsewhere along the dichotomy, the northern plains
are significantly lower than the highlands, and the dichotomy is
a real and distinct topographic feature [Watters and Robinson,
1998; Frey et al., 1998].
A variety of models have been proposed for the origin of the
crustal dichotomy. These models fall into two groups, one
involving impact processes and the other involving internal or
endogenic processes. Impact models involve either one giant
impact [Wilhelms and Squyres, 1984; McGill, 1989] or multiple
impacts [Frey and Schultz, 1988]. These models require that the
dichotomy is very old (early or pre-Noachian), forming before
the end of the period of heavy bombardment. McGill and
Dimitriou [1990] cite late Noachian to early Hesperian fractur-
ing and faulting in the northern lowlands and along the dichot-
omy boundary as evidence of a younger age of formation. They
propose that the early northern crust was thinned and subsided
through delamination or crustal erosion by mantle convection.
The eroded crustal material is thought to have been globally
dispersed [McGill and Dimitriou, 1990].
There is evidence of fracturing and faulting along the bound-
ary in Amenthes and northern Terra Cimmeria (Figure 1). The
presence of lobate scarps in proximity to and oriented parallel
with the boundary suggests that the dichotomy in this region
has a distinct tectonic signature. The age of the lobate scarps is
difficult to determine through crater counts because the struc-
tures are relatively small in areal extent. However, a number of
scarps crosscut the walls and floors of large (.16 km diameter)
degraded impact craters, while smaller (,16 km diameter),
relatively fresh craters are superimposed on some of the
scarps. On the basis of crater ages and crosscutting relation-
ships, Maxwell and McGill [1988] suggest that scarp formation
in the highlands of Amenthes near the dichotomy boundary
occurred during the late Noachian. The fact that some of the
lobate scarps in northern Terra Cimmeria deform intercrater
plains estimated to be early Hesperian in age [Greeley and
Guest, 1987] suggests that the formation of the scarps may have
continued into the early Hesperian. Wilhelms and Baldwin
[1989] conclude that wrinkle ridges and scarps in this region
formed in the early Hesperian. Thus it is plausible that the
formation of the fractures and faults associated with the di-
chotomy and the formation of the lobate scarps were roughly
syntectonic, separated spatially by ;400 km, if the boundary
has always been close to its present position, as suggested by
McGill and Dimitriou [1990]. Models proposing that only litho-
spheric extension accompanied the formation of the crustal
dichotomy [e.g., McGill and Dimitriou, 1990] cannot account
for the observed compressional deformation.
4.2. Plate Tectonics
A plate tectonic model has been proposed for the origin of
the dichotomy where ancient crust was removed through hemi-
spheric subduction and the dichotomy boundary marks relic
plate margins [Sleep, 1994]. Hemispheric subduction may be a
viable hypothesis for the origin of the crustal dichotomy be-
cause it is the dominant process for crustal thinning on the
Earth [Turcotte, 1996], and it is consistent with the remarkably
flat topography of the northern plains revealed by the MOLA
data [Turcotte, 1999]. In his model, Sleep [1994] suggests that
the dichotomy boundary extending from the Tyrrhena region,
through Amenthes and Cimmeria, to Memnonia is a passive
margin. Watters and Robinson [1998] suggest that compres-
sional stresses due to flexure from an emerging spreading ridge
and ambient compressional stresses caused by subsequent
rapid spreading are one possible mechanism to account for the
formation of the lobate scarps and the fractures along the
dichotomy boundary in Amenthes and northern Terra Cimme-
ria. This model also suggests that the boundary in the northern
Terra Cimmeria?Amenthes region is a passive margin. The
analogy between the dichotomy boundary and passive margins
is further strengthened by MOLA data. Frey et al. [1998] report
that the regional topography of the dichotomy boundary zone
indicated in the MOLA data (a 2 to 4 km step between two
nearly flat surfaces over a distance of few hundred kilometers)
is consistent with some terrestrial passive margins.
There are, however, challenges to applying a plate tectonic
model to Mars. Pruis and Tanaka [1995] question many aspects
of the model proposed by Sleep [1994]. They argue that evi-
dence of tectonism and volcanism in areas predicted by the
WATTERS AND ROBINSON: LOBATE SCARPS AND THE DICHOTOMY18,988
Sleep [1994] model is sparse and that some structures that seem
to agree with the model often have incorrect orientations or
modes of origin. New topographic data, images, and magnetic
and gravity field measurements being returned from the Mars
Global Surveyor will allow an in-depth evaluation of the pos-
sible role of plate tectonics in Martian geologic history.
5. Conclusions
An analysis of lobate scarps in the northern Terra Cimme-
ria?Amenthes region using new topographic data indicates
that they reflect significant compressional deformation of the
heavily cratered highlands near the dichotomy boundary.
Crustal shortening due to thrust faulting is of the order of 2 to
3 km locally, and the regional compressional strain in the
highlands (Npld) adjacent to the dichotomy boundary is
;0.17%. The D-L relationships of the lobate scarps studied
(g 5 6.2 3 1023) is comparable to those observed for terres-
trial faults (g ranging from 100 to 1023). The formation of
fractures along the boundary appears to be roughly syntectonic
with the formation of the thrust faults responsible for the
lobate scarps. These new data and observations are consistent
with a tectonic origin for the dichotomy, involving both exten-
sional and compressional deformation. A successful model for
the origin of the crustal dichotomy must account for its distinct
topographic and tectonic signature. It must also explain the
presence of both the lobate scarps (compressional deforma-
tion) and fractures (extensional deformation) associated with
the dichotomy boundary.
Acknowledgments. We thank Kenneth L. Tanaka and Craig R.
Bina for their helpful reviews of the manuscript. We also thank Nor-
man H. Sleep for helpful discussions about the possibility of plate
tectonics on Mars. The PICS software used to acquire the photoclino-
metric data was provided by the USGS Flagstaff. This research was
supported by grants from National Aeronautics and Space Adminis-
tration?s Planetary Geology and Geophysics Program and the Mars
Data Analysis Program.
References
Binder, A. B., and J. C. Jones, Spectrophotometric studies of the
photometric function, composition, and distribution of the surface
materials of Mars, J. Geophys. Res., 77, 3005?3020, 1972.
Brewer, J. A., S. B. Smithson, J. E. Oliver, S. Kaufman, and L. D.
Brown, The Laramide orogeny: Evidence form COCORP deep
crustal seismic profiles in the Wind River mountains, Wyoming,
Tectonophysics, 62, 165?189, 1980.
Byerlee, J., Friction of rocks, Pure Appl. Geophys., 116, 615?626, 1978.
Cartwright, J. A., B. D. Trudgill, and C. S. Mansfield, Fault growth by
segment linkage: An explanation for scatter in maximum displace-
ment and trace length data from the Canyonlands Grabens of SE
Utah, J. Struct. Geol., 17, 1319?1326, 1995.
Cartwright, J. A., C. S. Mansfield, and B. D. Trudgill, The growth of
normal faults by segment linkage, in Modern Developments in Struc-
tural Interpretation: Validation and Modelling, edited by P. G.
Buchanan and D. A. Nieuwland, Geol. Soc. Spec. Publ., 99, 163?177,
1996.
Cordell, B. M., and R. G. Strom, Global tectonics of Mercury and the
Moon, Phys. Earth Planet. Inter., 15, 146?155, 1977.
Cowie, P. A., and C. H. Scholz, Displacement-length scaling relation-
ship for faults: Data synthesis and discussion, J. Struct. Geol., 14,
1149?1156, 1992.
Cowie, P. A., C. H. Scholz, M. Edwards, and A. Malinverno, Fault
strain and seismic coupling on mid-ocean ridges, J. Geophys. Res., 98,
17911?17920, 1993.
Davis, P. A., and L. A. Soderblom, Modeling crater topography and
albedo from monoscopic Viking Orbiter images, J. Geophys. Res., 89,
9449?9457, 1984.
Dawers, N. H., and M. H. Anders, Displacement-length scaling and
fault linkage, J. Struct. Geol., 17, 607?614, 1995.
Dawers, N. H., Anders, M. H., and Scholz, C. H., Growth of normal
faults: Displacement-length scaling, Geology, 21, 1107?1110, 1993.
Downs, G. S., P. J. Mouginis-Mark, S. H. Zisk, and T. W. Thompson,
New radar derived topography for the northern hemisphere of Mars,
J. Geophys. Res., 87, 9747?9754, 1982.
Elliott, D., The energy balance and deformation mechanisms of thrust
sheets, Philos. Trans. R. Soc. London, on Ser. A, 283, 289?312, 1976a.
Elliott, D., The motion of thrust sheets, J. Geophys. Res., 81, 949?963,
1976b.
Frey, H., and R. A. Schultz, Large impact basins and the mega-impact
origin for the crustal dichotomy on Mars, Geophys. Res. Lett., 15,
229?232, 1988.
Frey, H., S. E. Sakimoto, and J. Roark, The MOLA topographic
signature at the crustal dichotomy boundary zone on Mars, Geophys.
Res. Lett., 25, 4409?4412, 1998.
Gillespie, P. A., J. J. Walsh, and J. Watterson, Limitations of dimen-
sion and displacement data from single faults and the consequences
for data analysis and interpretation, J. Struct. Geol., 14, 1157?1172,
1992.
Greeley, R., and J. E. Guest, Geologic map of the eastern equatorial
region of Mars, scale 1:15,000,000, U.S. Geol. Surv. Misc. Invest. Ser.
Map, I-1802-B, 1987.
Gries, R., Oil and gas prospecting beneath Precambrian of foreland
thrust plates in Rocky Mountains, AAPG Bull., 67, 1?28, 1983.
Howard, A. D., K. R. Blasius, and J. A. Cutts, Photoclinometric de-
termination of the topography of the martian north polar cap, Ica-
rus, 50, 245?258, 1982.
Jaeger, J. C., and N. G. W. Cook, Fundamentals of Rock Mechanics,
3rd ed., 593 pp., Chapman and Hall, New York, 1979.
Maxwell, T. A., and G. E. McGill, Ages of fracturing and resurfacing
in the Amenthes Region, Mars, Proc. Lunar Planet. Sci. Conf., 18th,
701?711, 1988.
McEwen, A. S., Photometric functions for photoclinometry and other
applications, Icarus, 92, 298?311, 1991.
McGill, G. E., Buried topography of Utopia, Mars: Persistence of a
giant impact depression, J. Geophys. Res., 94, 2753?2759, 1989.
McGill, G. E., and A. M. Dimitriou, Origin of the Martian global
dichotomy by crustal thinning in the late Noachian or early Hespe-
rian, J. Geophys. Res., 95, 12595?12605, 1990.
Melosh, H. J., and W. B. McKinnon, The tectonics of Mercury, in
Mercury, edited by F. Vilas, C. R. Chapman, and M. S. Matthews, pp.
374?400, Univ. of Ariz. Press, Tucson, 1988.
Moore, J. M., and R. A. Schultz, Processes of Faulting in Jointed
Rocks of Canyonlands National Park, Utah, Geol. Soc. Am. Bull.,
111, 808?822, 1999.
Mouginis-Mark, P. J., and M. S. Robinson, Evolution of the Olympus
Mons caldera, Mars, Bull. Volcanol., 54, 347?360, 1992.
NASA, Mission to Mars: Digital image map [CD-ROM], USA_NA-
SA_PDS_VO_2004, NASA Planet. Data Syst., Pasadena, Calif.,
1991.
NASA, Mission to Mars: Mars Global Surveyor [CD-ROM], USA-
_NASA_PDS_MGS_0001, NASA Planet. Data Syst., Pasadena, Cal-
if., 1998.
Pruis, M. J., and K. L. Tanaka, The Martian northern plains did not
result from plate tectonics (abstract), Lunar Planet. Sci. Conf., XXVI,
1147?1148, 1995.
Roth, L. E., G. S. Downs, and R. S. Saunders, Radar altimetry of south
Tharsis, Mars, Icarus, 42, 287?316, 1980.
Scholz, C. H., and P. A. Cowie, Determination of geologic strain from
fault slip data, Nature, 346, 837?839, 1990.
Schultz, R. A., Gradients in extension and strain at Valles Marineris,
Mars, Planet. Space Sci., 43, 1561?1566, 1995.
Schultz, R. A., Displacement-length scaling for terrestrial and Martian
faults: Implications for Valles Marineris and shallow planetary gra-
bens, J. Geophys. Res., 102, 12009?12015, 1997.
Schultz, R. A., and A. N. Fori, Fault-length statistics and implications
of graben sets at Candor Mensa, Mars, J. Struct. Geol., 18, 373?383,
1996.
Simpson, R. A., J. K. Harmon, S. H. Zisk, T. W. Thompson, and D. O.
Muhleman, Radar determination of Mars surface properties, in
Mars, edited by H. Kieffer et al., pp. 652?685, Univ. of Ariz. Press,
Tucson, 1993.
Sleep, N. H., Martian plate tectonics, J. Geophys. Res., 99, 5639?5656,
1994.
18,989WATTERS AND ROBINSON: LOBATE SCARPS AND THE DICHOTOMY
Smith, D. E., and M. T. Zuber, The shape of Mars and the topographic
signature of the hemispheric dichotomy, Science, 271, 184?188,
1996.
Smith, D. E., and M. T. Zuber, The relationship between MOLA
northern hemisphere topography and the 6.1-Mbar atmospheric
pressure surface of Mars, Geophys. Res. Lett., 25, 4397?4400, 1998.
Smith, D. E., et al., Topography of the northern hemisphere of Mars
from the Mars Orbiter Laser Altimeter, Science, 279, 1686?1692,
1998.
Stone, D. S., Geologic interpretation of seismic profiles, Big Horn
Basin, Wyoming, I, East flank, in Seismic Exploration of the Rocky
Mountain Region, edited by R. R. Gries and R. C. Dyer, pp. 165?174,
Rocky M. Assoc. of Geol., Denver, Colo., 1985.
Strom, R. G., N. J. Trask, and J. E. Guest, Tectonism and volcanism on
Mercury, J. Geophys. Res., 80, 2478?2507, 1975.
Tanaka, K. L., The stratigraphy of Mars, Proc. Lunar Planet. Sci. Conf.
17th, Part I, J. Geophys. Res., 91, suppl., E139?E158, 1986.
Tanaka, K. L., and P. A. Davis, Tectonic history of the Syria Planum
province of Mars, J. Geophys. Res., 93, 14893?14907, 1988.
Thorpe, T. E., Mariner 9 photometric observations of Mars from
November 1971 through March 1972, Icarus, 20, 482?489, 1973.
Turcotte, D. L., Martian tectonics: Why the global dichotomy and the
Tharsis Uplift (abstract), Lunar Planet. Sci. Conf., XXVII, 1345?
1346, 1996.
Turcotte, D. L., Tectonics and volcanism on Mars: What do we and
what don?t we know? (abstract), Lunar Planet. Sci. Conf., XXX,
1187?1188, 1999.
Turcotte, D. L., and G. Schubert, Geodynamics: Application of Con-
tinuum Physics to Geological Problems, 450 pp., John Wiley, New
York, 1982.
Watters, T. R., Compressional tectonism on Mars, J. Geophys. Res., 98,
17049?17060, 1993.
Watters, T. R., and M. S. Robinson, Radar and photoclinometric
studies of wrinkle ridges on Mars, J. Geophys. Res., 102, 10889?
10903, 1997.
Watters, T. R., and M. S. Robinson, Lobate Scarps and the origin of
the crustal dichotomy on Mars (abstract), Lunar Planet. Sci. Conf.,
XXIX, 1475?1476, 1998.
Watters, T. R., M. S. Robinson, and A. C. Cook, Comparison of
Discovery Rupes, Mercury with terrestrial thrust faults: New esti-
mates of the decrease in radius of the planet due to global contrac-
tion (abstract), Lunar Planet. Sci. Conf., XXVIII, 1507?1508, 1997.
Watters, T. R., M. S. Robinson, and A. C. Cook, Topography of lobate
scarps on Mercury: New constraints on the planet?s contraction,
Geology, 26, 991?994, 1998.
Watters, T. R., M. S. Robinson, and R. A. Schultz, Comparative
studies of displacement-length relations of lobate scarps on Mercury
and Mars with terrestrial faults (abstract), Lunar Planet. Sci. Conf.,
XXX, 1826?1827, 1999.
Wichman, R. W., and P. H. Schultz, Sequence and mechanisms of
deformation around the Hellas and Isidis Impact Basins on Mars, J.
Geophys. Res., 94, 17333?17357, 1989.
Wildey, R. L., Generalized photoclinometry of Mariner 9, Icarus, 25,
613?626, 1975.
Wilhelms, D. E., and R. J. Baldwin, The role of igneous sills in shaping
the Martian uplands, Lunar Planet Sci., XIX, 355?365, 1989.
Wilhelms, D. E., and S. W. Squyres, The martian hemispheric dichot-
omy may be due to a giant impact, Nature, 309, 138?140, 1984.
Wojtal, S. F., Changes in fault displacement populations correlated to
linkage between faults, J. Struct. Geol., 18, 265?279, 1996.
Zuber, M. T., D. E. Smith, S. C. Solomon, D. O. Muhleman, J. W.
Head, J. B. Garvin, J. B. Abshire, and J. L. Bufton, The Mars
Observer Laser Altimeter investigation, J. Geophys. Res., 97, 7781?
7797, 1992.
Zuber, M. T., et al., Observations of the north polar region of Mars
from the Mars Orbiter Laser Altimeter, Science, 282, 2053?2060,
1998.
M. S. Robinson, Department of Geological Sciences, Northwestern
University, Evanston, IL 60209.
T. R. Watters, Center for Earth and Planetary Studies, National Air
and Space Museum, Smithsonian Institution, Washington, DC 20560.
(twatters@ceps.nasm.edu)
(Received November 12, 1998; revised April 13, 1999;
accepted April 15, 1999.)
WATTERS AND ROBINSON: LOBATE SCARPS AND THE DICHOTOMY18,990