Special Reports

Accumulation and Erosion of Mars' South Polar Layered Deposits

See allHide authors and affiliations

Science  21 Sep 2007:
Vol. 317, Issue 5845, pp. 1715-1718
DOI: 10.1126/science.1144120

Abstract

Mars' polar regions are covered with ice-rich layered deposits that potentially contain a record of climate variations. The sounding radar SHARAD on the Mars Reconnaissance Orbiter mapped detailed subsurface stratigraphy in the Promethei Lingula region of the south polar plateau, Planum Australe. Radar reflections interpreted as layers are correlated across adjacent orbits and are continuous for up to 150 kilometers along spacecraft orbital tracks. The reflectors are often separated into discrete reflector sequences, and strong echoes are seen as deep as 1 kilometer. In some cases, the sequences are dipping with respect to each other, suggesting an interdepositional period of erosion. In Australe Sulci, layers are exhumed, indicating recent erosion.

A marked similarity between Mars and Earth is the presence of thick polar ice caps, which potentially provide a wealth of data about recent climate and atmospheric history (1, 2). The martian polar plateaus are as much as 3.5 km thick (3) and composed of layers overlain by a residual (i.e., perennial) ice cap that, according to recent analyses, is no more than a few tens of meters thick (46). The polar residual ice in the south is predominantly CO2 ice containing trace amounts of dust and water ice (5). Compared to the north, the residual ice covers only a small portion of the south polar layered deposits (SPLD), which are offset from the pole by ∼2° and are asymmetrically distributed between latitudes 70°S and 80°S (3). Most plateau surfaces are buried under an optically thick layer of sand or of seasonal frost accumulations (7). The layers are composed of water ice mixed with dust, with layer thicknesses and ice/dust ratios possibly controlled by orbital cycles (8). They apparently formed by deposition of ice condensates or precipitates and atmospheric dust (9, 10). The polar plateaus are considered to be very young. On the basis of crater counts, the surface age of Planum Australe is estimated to be on the order of tens of millions of years (11, 12).

Planum Australe was probed previously with the Mars Advanced Radar for Subsurface and Ionospheric Sounding (MARSIS) onboard the European Space Agency's Mars Express (6, 13). The MARSIS radar signals penetrated to depths estimated to be greater than 3.7 km (6). Subsurface echoes are frequently very strong, indicating minimal attenuation of the signal through the SPLD. Although plausible solutions to models of layer composition based on the dielectric losses are unconstrained (14), the SPLD may be composed of nearly pure water ice (6). With a vertical resolution of 150 m in free space, however, MARSIS is not suited to study the structure of the deposits, because individual layers are a few meters thick (15).

SHARAD (Shallow Radar) is a synthetic-aperture, orbital sounding radar on the Mars Reconnaissance Orbiter (MRO) (16). SHARAD was designed to complement MARSIS by trading penetration depth for improved vertical resolution (17, 18) and is better suited for mapping 10-m-scale stratigraphy in the polar plateaus. Data are collected preferentially on the night side of Mars, where the Sun-excited ionosphere has a low electron density and plasma frequency, and radar echoes are thus nearly unaffected by distortions caused by propagation through a frequency-dispersive medium. SHARAD operates at a center frequency of 20 MHz, with a 10-MHz bandwidth that provides a free-space range resolution of ∼15 m after pulse compression. Resolution in the across-track direction ranges from 3 to 6 km, and in the along-track direction after synthetic-aperture processing is 0.3 to 1 km. The transmitted signal is a 10-W, 85-μschirped (linearly frequency-modulated) pulse emitted from a 10-m-long dipole antenna. After downlink to Earth, coherent azimuth sums on ∼1500 pulses taken at a pulse repetition frequency of 700 Hz result in a signal-to-noise ratio of 30 to 40 dB for a relatively smooth surface (17, 18). Waves that are transmitted into the subsurface reflect from dielectric discontinuities and return to the instrument at greater time delay than the nadir surface echo. The primary obstacle to the identification of subsurface echoes is interference from off-nadir surface echoes (known as “surface clutter”) that arrives at the radar receiver with the same time delay as that of subsurface echoes. The data described here were acquired in January and February 2007, on orbits 2202, 2413, 2624, and 2835. The ground tracks of the radar are located within Planum Australe's Promethei Lingula region (Fig. 1). Here we report on the SPLD stratigraphy as mapped by SHARAD in Promethei Lingula and interpret this information in terms of the depositional and erosional history.

Fig. 1.

Topography of the Promethei Lingula region. Promethei Lingula is the extension of the Planum Australe plateau delimited by Chasma Australe to the top, Ultimum Chasma to the bottom, and Promethei Planum to the right. The ground tracks of the MRO spacecraft during SHARAD data acquisitions are shown in red for orbit 2202, in yellow for orbit 2413, in green for orbit 2624, and in blue for orbit 2835. The inset shows the location in the south polar area of Mars.

Roughly half of Promethei Lingula is dissected by canyon systems. These include deep canyons radial to the Promethei Lingula distal margin (chasmata) and smaller, curvilinear canyons perpendicular to the chasmata commonly referred to as “spiral troughs” (Fig. 1). The origin of the chasmata has been attributed to dissection by katabatic wind (12, 19, 20) or to basal melting (21), where as development of curvilinear canyons has been attributed to aeolian erosion (20, 22) or to ductile deformation (23) of the SPLD.

SHARAD data were acquired during four closely spaced (<5 km), nearly parallel orbit segments (Fig. 1). These data (Fig. 2) are all characterized by a multilayered internal structure with dozens of internal reflections. Many of these reflections are traceable over distances of up to 150 km over the smooth area extending from the outer edge to the central part of the Planum Australe plateau (from right to left in Fig. 2). There are several bands (packets) containing many thin reflectors running parallel or sub-parallel to the surface, separated by thicker dark bands. To determine if they arise from subsurface interfaces or come from surface topographic features at the same time delay, we ran a high-fidelity model of the expected contribution of off-nadir topographic clutter using Mars Orbiter Laser Altimeter (MOLA) data (7, 24, 25). Results show no visible surface topographic feature that explains the observed bands; hence, we conclude that they are subsurface reflectors. This banded structure is similar, and certainly related, to the layering visible in optical images, and radar reflections are likely caused by intralayer compositional variations in dust content or density. Sets of reflectors can be correlated among the four radargrams, giving us confidence that the same layered structure extends laterally (across track) for several kilometers. Reflections are seen down to depths of about 1 km, consistent with the very low attenuation through the SPLD observed by MARSIS. However, such reflections are at a shallower depth than the deepest reflection in the MARSIS radargrams over the same area, which are found at 1.3 km and correspond to the local plateau height relative to the surrounding plains (6). We deem that the lack of any deeper reflectors is more likely due to attenuation or scattering of the SHARAD radar signal than to the character of the plateau basal interface.

Fig. 2.

Radargrams for orbits 2202, 2413, 2624, and 2835. A radargram is a two-dimensional cross section of the surface and subsurface in which brightness corresponds to echo power. The vertical axis is the two-way pulse travel time, and the horizontal axis is the distance of the spacecraft along its ground track. Similar subsurface layering structure is seen in these four closely spaced, parallel tracks.

The distinct subsurface layered structure noted above is not present along the entire length of the four SHARAD ground tracks. In some areas, these dielectric interfaces pinch out or disappear entirely. Possible reasons for these changes in the echo behavior include (i) tilting of the layers, which can move the specular reflection from a smooth interface out of the Doppler bandwidth of the processed data; (ii) increased surface roughness that decreases the effective amount of radar signal transmitted into the subsurface; or (iii) changes in the geologic structure or physical state of the materials that diminishes their dielectric contrast.

From analysis of the radargrams, it is possible to identify four main boundaries (colored lines in Fig. 3) that separate different dielectric sequences. The first sequence, limited at the base by the yellow boundary and located in time between 1.7 and 5.3 μs of depth below the surface, is dipping toward the right. Several of the upper layers in this sequence terminate at the surface, which is nearly horizontal in this segment. Below this reflector a second one (in white) is detectable at a time depth of about 4.3 to 5.6 μs. Beneath such an interface a second well-defined sequence of layers is clearly visible up to a time depth of about 5.6 to 7.1 μs where the third main reflector (in orange) closes the series. Finally, below it another multilayered structure, limited by a reflector (in light blue) located at a depth of about 11.0 to 12.9 μs, is also visible.

Fig. 3.

Enlarged view of data for orbit 2202. Four main reflectors, which separate different dielectric sequences, are marked with lines of yellow, white, orange, and light blue. Red ellipses highlight echo features identified as surface clutter through radar surface echo simulation (see text).

To convert echo time delays to depths requires assumptions about the material that the radar pulse is passing through. As mentioned above, the SPLD may consist mostly of water ice (6). We thus convert time to depth by assuming a real part of the relative permittivity of 3.5, corresponding to that of water ice with a few percent of dust. The four reflectors shown in Fig. 3 are located, from top to bottom, at depths of about 130 to 420 m, 340 to 450 m, 450 to 570 m, and 890 to 1030 m. Because the depths scale by the inverse of the square root of the dielectric constant, small variations in material (i.e., dust content) will not have a profound effect on calculated depths.

Above the first reflector in Fig. 3, the layers are continuous and generally of even thickness laterally. Below the second reflector the same observation applies. Between the two, the layer sequence goes from thick (about 240 m) to nearly zero. The layers become difficult to follow within this wedge-shaped region. We propose two possible interpretations for this morphology: (i) The pinching out of the layers in the wedge-shaped region is an erosional feature (an unconformity); or (ii) the shape of this region is a primary feature of the deposition, i.e., the thickness of some layers varies laterally. The first interpretation is consistent with conclusions from a study of exposed layers in Promethei and Ultimum Chasmata (Fig. 1) (19). However, because there is no clear image of truncation of the bedding along a dipping boundary in the wedge-shaped region, we are unable to rule out interpretation (ii).

We interpret the dipping layers that are truncated at the surface within Australe Sulci as due to erosion of the plateau surface rather than due to lateral variations in SPLD accumulation rate, because this last process would be unable to produce the observed constant thickness of the layers. This interpretation is consistent with the observation of erosional surface features in Australe Sulci, whose shape and direction may be controlled by the wind, most likely by off-pole katabatic winds (12). The local stratigraphic evolution can thus be interpreted as consisting of at least four stages: (i) essentially uniform deposition of a lower package; (ii) erosion at an angle, or laterally variable deposition; (iii) renewed deposition on the angled surface; and (iv) truncation of the angled package to the present horizontal surface as result of erosion.

Despite the age uncertainties of the polar deposits, the strength of CO2 ice is insufficient to support the long-term topography of the polar caps (26). SHARAD data are consistent with a mechanically strong plateau material, as we see no evidence, such as folding or faulting, for flow in the bulk of the plateau's layered structure. However, flow resulting in broad-scale changes in layer thickness or flow at the very base of the deposits cannot be ruled out by these observations. The continuity of layers, spanning distances of up to 150 km, is evidence that the depositional processes that formed Planum Australe were regional in scale, as already inferred from layer correlation studies of exposures in Australe Mensa (27) and Promethei Lingula (19). Spectral and albedo observations of the surface indicate that the plateau surface is covered by an optically thick layer of dust or rocky material (7), but we do not see evidence in our data that can be interpreted in terms of an ice-free surface layer, perhaps because of its small thickness compared to SHARAD vertical resolution.

References and Notes

View Abstract

Navigate This Article