Volume 48, Issue 13 e2021GL093631
Research Letter
Free Access

Characteristics of the Basal Interface of the Martian South Polar Layered Deposits

Aditya R. Khuller

Corresponding Author

Aditya R. Khuller

School of Earth and Space Exploration, Arizona State University, Tempe, AZ, USA

Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, USA

Correspondence to:

A. R. Khuller,

[email protected]

Search for more papers by this author
Jeffrey J. Plaut

Jeffrey J. Plaut

Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, USA

Search for more papers by this author
First published: 16 June 2021
Citations: 12


We expand on previous studies of the South Polar Layered Deposits' (SPLD) basal interface using data acquired by the Mars Advanced Radar for Subsurface and Ionospheric Sounding (MARSIS) to obtain detailed maps of elevation, topography, and reflected radar power. Using these maps, we derive the thickness (ranging from 0 to 3.7 km) and volume of the SPLD (∼1.60 × 106 km3). While most basal interfaces reflect less power than the average SPLD surface, areas with basal echo power exceeding that of the surface are widespread throughout the SPLD, including at the location of potential subglacial water bodies in Ultimi Scopuli. The occurrence of these high basal echo power signatures appears to be largely frequency independent in MARSIS data. While the cause of the relatively high basal echo power values is uncertain, our observations suggest that this behavior is widespread, and not unique to Ultimi Scopuli.

Key Points

  • We present new, detailed (∼44,000 points) maps of South Polar Layered Deposits' basal elevation and topography using Mars Advanced Radar for Subsurface and Ionospheric Sounding (MARSIS) data

  • We report the reflected basal echo power from three MARSIS frequencies, showing large spatial variability but minimal dependence on frequency

  • Areas with high basal echo power are widespread, and not limited to the location of potential subglacial water bodies in Ultimi Scopuli

Plain Language Summary

Using radio waves, we investigated the properties of materials lying below the surface of Mars' south polar regions (South Polar Layered Deposits; “SPLD”). We made maps of the buried surface below the SPLD (“basal interface”) and measured the radar energy reflected by this interface. From these maps, we measured the SPLD volume as 1.6 × 106 km3. Additionally, we found that there are multiple areas throughout the south polar region where the energy reflected from the basal interface is unexpectedly higher than that of the surface. Previous analyses of one such region suggested that these stronger reflections could be caused by the presence of an underground lake.

1 Introduction

The South Polar Layered Deposits (SPLD) are several kilometer-thick stacks of layered H2O ice-rich deposits extending outward from the Martian south pole. These deposits are thought to have been emplaced over the last 10–100s of million years in the form of atmospherically deposited H2O ice, dust, and minor amounts of CO2 ice (Becerra et al., 2019; Cutts, 1973; Herkenhoff & Plaut, 2000; Koutnik et al., 2002; Plaut, 2005). The layers within the SPLD are potentially linked to changes in Mars' obliquity and orbital eccentricity over time, and therefore might hold millions of years of Mars' recent climatic history.

Initial radar sounding of the SPLD revealed the presence of the basal interface, where the H2O ice-rich SPLD lie in contact with the substrate (Plaut et al., 2007). The locations of this basal interface were mapped to obtain subsurface topography, and provide estimates of the thickness and the volume of the SPLD (Plaut et al., 2007). In some cases, this basal interface showed anomalously bright reflections that were brighter than the surface return. The presence of subsurface liquid water was initially ruled out due to the cold temperatures expected at the base of the SPLD (Plaut et al., 2007). However, recent analyses at a few locations of bright basal reflectors suggest there may be a liquid water component at the interface (Lauro et al., 2021; Orosei et al., 2018), although the presence and stability of liquid water is under debate (Arnold et al., 2019; Sori & Bramson, 2019). Thus, key questions about the properties of the basal interface below the SPLD remain unanswered. In this work we build on previous studies by making use of the wealth of data acquired by the Mars Advanced Radar for Subsurface and Ionospheric Sounding (MARSIS) since 2005 over the south polar region to further assess the characteristics of the SPLD basal interface.

2 Methods

2.1 Mars Advanced Radar for Subsurface and Ionospheric Sounding (MARSIS)

MARSIS is a multi-frequency synthetic aperture orbital sounding radar onboard the European Space Agency's Mars Express spacecraft (Jordan et al., 2009; Picardi et al., 2004). MARSIS operates simultaneously at two of four frequency bands (1.8, 3.0, 4.0 and 5.0 MHz) with a 1 MHz bandwidth. After on-board processing, MARSIS′ along-track resolution is 5–10 km, with a cross-track footprint of 10–30 km. MARSIS data have a vertical resolution of ∼150 m in free space, and about 50–100 m in typical geological materials.

2.2 3D Imaging Volume

In this work, we use a compiled south polar data set from MARSIS data collected since 2005, consisting of over 2,000 orbits taken in the three higher frequency bands (3, 4 and 5 MHz) to generate a 3D radar imaging volume (Text S1; Gim et al., 2018; Plaut et al., 2018; Plaut, 2020). Note that MARSIS data cannot be acquired poleward of 87°S. Key features of this 3D radar imaging volume are: (a) voxel (volume pixel) dimensions of 1.5 × 1.5 km (horizontal) × 50 m (depth), (b) depth correction is applied in the subsurface using a wave speed in pure water ice (real dielectric constant ε = 3.1), (c) overlapping echo frames from different orbits are averaged, (d) empty voxels are filled with horizontally applied nearest neighbor interpolation in the volumes used for interface detection, and (e) slices are extracted for all vertical and horizontal planes in each volume for individual study and animations. By assuming that the wave speed is equal to that in pure water ice, reflectors are repositioned in an approximately correct geometry to facilitate identification of interfaces. In areas known to contain lenses of CO2 ice (Phillips et al., 2011), some distortions in reflector position can occur. However, this effect is evident in only a small fraction of the 3D volume (described in Section 12.).

2.3 Basal Interface Mapping and Power Characteristics

We identified and marked the basal interface reflector for all slices where it was discernible in the volume in RGB color composite images (red, 3 MHz; green, 4 MHz; blue, 5 MHz). The basal reflector was identified as the deepest linear feature below the surface return that was distinct in one or more frequency bands and contiguous with the internal SPLD structure above it. Figure 1 shows an example of this basal interface detection. The thickness of the reflector can vary slightly, and is typically 2–3 voxels thick (equivalent to 50–100 m uncertainty in depth). In some cases, multiple candidate basal interfaces were present, in which case a selection was made based on the elevation, context and relative brightness in the subject slice and in its neighbors. We also extracted the reflected radar power of the basal interface at each frequency for all mapped basal interface points, and report it relative to the surface reflected power (Text S2). This relative basal reflected power allows for a quantitative assessment of basal dielectric properties and compensates for any ionospheric attenuation or other observation-specific radiometric variations (Lauro et al., 2010; Orosei et al., 2018).

Details are in the caption following the image

Top: Vertical slice through the Mars Advanced Radar for Subsurface and Ionospheric Sounding 3D volume, equivalent to a radargram. The vertical dimension is converted to distance, assuming a wave speed in water ice for all points in the subsurface. Color assignments are: red, 3 MHz; green, 4 MHz; blue, 5 MHz. Middle: Same as top image, with basal interface detections marked in red. Bottom: Horizontal slice taken at an elevation between the typical positions of the basal interface and the South Polar Layered Deposits surface, with Mars Orbiter Laser Altimeter hillshade added for context. White line indicates the position of the slice in the top panels. This image is rotated 90° counterclockwise from the orientation of later map figures. Image center is 84.1°S, 90.0°E.

We determined the elevation of each detection of the basal interface relative to the International Astronomical Union (IAU; Seidelmann et al., 2002) reference ellipsoid, using the water ice depth correction (e.g., Phillips et al., 2008). Then, the elevation of the Mars Orbiter Laser Altimeter (MOLA; Smith et al., 2001; 256 pixels per degree gridded radius data in polar stereographic projection) surface overlying each subsurface interface was extracted to find the thickness of the SPLD at each point. We interpolated the basal elevation points and the elevations at the unit boundary to obtain a map of basal interface topography using the Natural Neighbor algorithm in ArcGIS (Childs, 2004). This interpolated basal interface topography map was then subtracted from the MOLA surface topography to obtain a thickness map of the SPLD. Points along the SPLD unit boundary were constrained to have zero thickness.

3 Results

3.1 Basal Interface Topography

We mapped about 44,000 points on the SPLD basal interface (Figure 2a), representing a 25X improvement over previous work (Plaut et al., 2007). Most of the basal interface detections lie at elevations 3–4 km above the reference ellipsoid. However, unusually low elevation detections are present in Ultima Lingula (which also contains high elevation detections), and within some craters. The highest elevation of the basal interface is about ∼4.7 km above the ellipsoid. Note that we have omitted the unusually deep depressions in the near-polar region previously mapped by Plaut et al. (2007) because their regional context suggests that they represent a distinct unit that lies below the base of the SPLD (Figure S1).

Details are in the caption following the image

(a) Elevations of the South Polar Layered Deposits (SPLD) basal interface, relative to the Mars International Astronomical Union reference ellipsoid, based on Mars Advanced Radar for Subsurface and Ionospheric Sounding (MARSIS) measurements. (b) Interpolated topography of the SPLD basal interface based on MARSIS measurements, with Mars Orbiter Laser Altimeter data shown outside the SPLD. The SPLD unit boundary is outlined in black on both maps.

Most basal interfaces appear relatively flat in the 3D volume. However, in some cases, likely subsurface depressions (corresponding to impact craters or other basins) are observed. Near 87°S, 90°E, we observe an unusually high elevation region near the polar cap. This feature is just south of a previously identified subsurface dome near 87°S, 49°E, thought to represent a local region of enhanced H2O ice accumulation (Whitten et al., 2017). Basal interface topography near the SPLD boundary often appears to extend off the PLD, indicating that the deposits were emplaced on complex, pre-existing topography.

3.2 Basal Interface Time Delay and SPLD Thickness and Volume

Figure 3a shows one-way time delay measurements of the SPLD basal interface, and Figure 3b shows the resultant interpolated thickness map of the SPLD. The thickest portions of the SPLD (∼3.7 km) lie below the residual polar cap near 0°E, whereas regions of moderate thickness 1.5–2 km) are distributed throughout the interior of the SPLD. Outside of the residual cap area, the SPLD material is concentrated asymmetrically toward the eastern hemisphere, with numerous clusters of moderate thickness present near the three chasmata (Chasma Australe, Promethei Chasma, and Ultimum Chasma) and Ultima Lingula. These regions of moderate thickness are approximately concentric around the south pole, although there are a few exceptions. For instance, while the distal areas of the SPLD are generally low in thickness (<∼1 km), the anomalous low elevation region in Ultima Lingula (near 73°S, 140°E) coincides with a region of moderate thickness. An unusually high thickness region near 81°S, 166°E is a partially buried crater filled with SPLD materials. Around 75°S, 150°E, a regional trend of alternating structures of high and low elevation appears to be present under the SPLD (Figure 3a). These high-low elevation sequences are oriented parallel to thrust fault-related landforms including Thyles Rupes, which is centered at 69.3°S, 132.3°E (Klimczak et al., 2018).

Details are in the caption following the image

(a) One-way time delay measurements between the surface and the South Polar Layered Deposits (SPLD) basal interface. (b) Thickness of the SPLD, based on interpolated Mars Advanced Radar for Subsurface and Ionospheric Sounding measurements and Mars Orbiter Laser Altimeter surface topography. Points along the margin of the SPLD (black outline) were constrained to have zero thickness.

While our study utilized an areally more extensive SPLD unit outline (based on Tanaka et al., 2014) that might increase the derived volume over previous studies, we also excluded the anomalously deep areas in the near-polar region that were mapped previously (Plaut et al., 2007). Thus, we derive a SPLD volume of ∼1.60 × 106 km3, corresponding to a global equivalent water layer thickness of ∼11.1 m, which is consistent with previous mapping (Plaut et al., 2007).

3.3 Basal Interface Reflected Power

A map of the basal interface reflected power relative to that of the average SPLD surface using 4 MHz data (Text S2) is shown in Figure 4a. Areas with high values (warm colors) indicate regions where the basal interface power exceeds that of the surface. Although most basal interface values are lower than that of the surface, as expected, there are numerous regions where the reflected power of the basal interface exceeds that of the surface. The most prominent example of this behavior is below the south polar residual cap, that is, the thickest portion of the SPLD, between 340°E and 0°E. Other bright basal reflectors appear to be scattered throughout the SPLD, suggesting that the process(es) causing this behavior may be widespread. Figure 4b illustrates the distribution of bright basal reflectors, showing regions where basal interface power relative to the surface is greater than 4 dB and the interface is deeper than 1 km (to exclude shallow basal reflectors). While the regions where there is potential evidence for liquid water (near 81°S, 193°E; Orosei et al., 2018; Lauro et al., 2021) do indeed have bright basal reflectors in our data, the signature is not unique to this area.

Details are in the caption following the image

(a) Basal interface reflected power relative to that of the average South Polar Layered Deposits (SPLD) surface for 4 MHz Mars Advanced Radar for Subsurface and Ionospheric Sounding (MARSIS) data. (b) Regions where the basal interface reflected power exceeds that of the average SPLD surface by values greater than 4 dB (in 4 MHz MARSIS data) and the interface is deeper than 1 km. Grayscale map shows SPLD thickness (same as Figure 3b). (c) Two-dimensional histogram of basal interface reflected power at 4 MHz, relative to that of the average SPLD surface, versus SPLD thickness. Colors indicate the fraction of observations per bin relative to the total.

Figure 4c shows a 2D histogram of basal interface reflected power relative to that of the average SPLD surface, versus SPLD thickness (Text S2). The scatter of points does not follow what might be an expected pattern of decreasing basal reflected power with increasing thickness. This suggests that absorption in the bulk SPLD medium is not a major control on the basal reflector power at these frequencies. Instead, there is a mostly symmetrical, main concentration of values centered near ∼1.5 km and −5 dB (warmer colors in Figure 4a and green regions in Figure 3). A secondary cloud of points displays high relative power at large thicknesses (>2 km), including the thickest near-polar regions. Other regions of high relative power are typically in areas where the SPLD is thin, possibly resulting from lower attenuation. 2D histograms using 3, 4, and 5 MHz data show virtually identical relationships between relative basal interface power and SPLD thickness (Figures S2 and S3).

4 Discussion

4.1 Areas Lacking Basal Reflectors

There are several large regions of the PLD where MARSIS does not detect basal reflectors (e.g., 85°S, 30°E; 85°S, 75°E; 85°S, 135°E; 80°S-85°S, 270°E; see Figure 2a). The absence of apparent basal reflectors could be due to the following reasons: (a) loss of the MARSIS signal between the surface and the basal interface, due to scattering and/or absorption, (b) a lack of contrast in dielectric constant at the basal interface, or (c) unfavorable geometry of the interface, such as roughness or slopes leading to scattering away from the nadir look direction. For example, the gap in reflectors around 30°E corresponds to the expected position of the Prometheus impact basin rim, which may produce rough buried topography that inhibits a strong reflection at the basal contact (Byrne & Ivanov, 2004; Whitten et al., 2017). The gap around 75°E is not so easily explained, as it presumably corresponds to the flat interior of the impact basin, Promethei Planum. Scattered gaps in the distal lobes toward 180°E may be related to rough heavily cratered highlands terrain. Several gap regions poleward of 85°S do not bear any clear relationship with exposed topography, but are found in regions previously noted to have unusually deep reflectors (Plaut et al., 2007), which we now interpret to lie at the base of an earlier, pre-SPLD deposit (see below).

4.2 Possible Evidence for Sub-SPLD Dorsa Argentea Formation Reflectors

Plaut et al. (2007) reported a series of buried depressions that appeared to be present in the basal topography from MARSIS measurements between 84°S and 87°S, from 95°E to 295°E. As mentioned above, we interpret these depressions to be a separate, distinct interface from the rest of the SPLD basal interface (Figure S1). This depression-hosting interface likely represents a second, deeper unit. In some cases, this deeper unit appears to merge with subsurface detections of the Hesperian-aged Dorsa Argentea Formation (DAF; Plaut et al., 1988; Kress & Head, 2015). Radar evidence for the DAF extending under the Amazonian-aged SPLD has been reported previously (e.g., near the Prometheus basin; Whitten et al., 2020), suggesting that the DAF forms a continuous unit beneath the SPLD. Our observations are consistent with this suggestion, and seem to indicate that the unit that hosts the buried depressions is a portion of the DAF. However, this sub-SPLD DAF unit is not visible throughout the SPLD, perhaps for similar reasons discussed above for the lack of basal interface detections at some SPLD locations. The presence of a sub-PLD paleo-polar deposit in Planum Australe is analogous in some ways to the “basal unit” of Planum Boreum in the north (Fishbaugh & Head, 2005; Malin & Edgett, 2001). Like the north polar basal unit, the sub-SPLD unit identified here is penetrated by MARSIS frequencies, but typically not by the higher frequency SHARAD (Shallow Radar; Seu et al., 2007) (Nerozzi & Holt, 2019; Selvans et al., 2010). Where exposed, the north polar basal unit differs morphologically from the DAF, but the stratigraphic position and radar properties of the two units are similar.

4.3 Buried CO2 Ice

Subsurface CO2 ice deposits have been detected in some parts of the SPLD using SHARAD data, with an estimated total volume of 16,500 km3 (Bierson et al., 2016; Phillips et al., 2011; Putzig et al., 2018). While the known volume of buried CO2 is large relative to Mars' total atmospheric mass, it only represents ∼1% of the total SPLD volume. Because the dielectric constant for CO2 ice (∼2.1) is lower than that used for the depth correction in our MARSIS 3D volume (3.1), some distortion (e.g., internal layers not parallel) due to deposits of CO2 ice can occur (Figure S4). However, the thickest portion (∼1 km) of these deposits is present in a part of Australe Mensa that is outside the available MARSIS data limit of 87°S (Figure S4). Other scattered areas with CO2 ice are present around Australe Mensa, but the thickness of these deposits is relatively small (<500 m), which causes a small distortion of underlying layers and interfaces. For example, the presence of a 500 m thick CO2 ice slab would lead to a ∼90 m vertical misplacement of underlying features under our assumption of a pure H2O ice column. This value is comparable to our uncertainty in marking the position of the basal reflector. Assuming this distortion is present in all the regions where CO2 ice has been detected previously, we estimate the errors in the reported SPLD elevations and total volume due to CO2 ice as ∼100 m and <0.5%, respectively.

We find no evidence for significant “hidden” CO2 deposits elsewhere in the SPLD (e.g., postulated in the NPLD in Broquet et al., 2020, and “reflection-free zones” in Phillips et al., 2011), although the sensitivity to such deposits in our method would make their detection difficult. Our assumption of a water ice wave speed results in a plausible geometry of the basal interface both regionally and locally. This observation, using the improved resolution and coverage over previous mapping, reinforces earlier conclusions that the load of the SPLD has resulted in little to no deflection of the lithosphere (Ojha et al., 2020; Plaut et al., 2007).

4.4 Potential Presence of Subglacial Water Bodies

Recent analyses of MARSIS data at Ultimi Scopuli suggest that bright subsurface basal relative power returns (up to ∼5 dB at 4 MHz) from a few ∼20 km regions at this location indicate the presence of multiple subglacial water bodies (Lauro et al., 2021; Orosei et al., 2018). These bright regions are visible in Figure 4a, with a value of 6.3 ± 1.7 dB at 4 MHz for their relative power (Text S2), similar to the values obtained by Orosei et al. (2018) and Lauro et al. (2021). Note that those studies present basal power relative to the median of the surface power along each orbit, whereas we normalize basal power relative to the overall average SPLD surface power. The presence of stable, subsurface liquid water may require an anomalously high geothermal heat flux at these locations (e.g., Hecht et al., 2018; Sori & Bramson, 2019). However, it has been proposed that liquid water solutions with magnesium/calcium perchlorates might be stable at these locations in a super-cooled state (Lauro et al., 2021). We deem it unlikely that the bright reflectors under the thickest SPLD areas are due to liquid water, as they commonly extend close to the surface where the mean annual surface temperature is too low to permit even super-cooled perchlorate brines to remain liquid.

Our observations suggest that the process(es) causing these anomalously bright subsurface echoes may be widespread across the SPLD. If the presence of liquid water in the form of subglacial water bodies is widespread under the PLD, then our improved map of basal topography can be used in hydrogeological models (Arnold et al., 2019) to assess the hydraulic potential of the basal interface at these locations of bright subsurface basal power return.

5 Conclusions

  1. We have expanded on previous work to characterize the basal interface of the SPLD using a large compilation of MARSIS radar data. By mapping ∼44,000 points, we generated new, detailed maps of elevation, topography, and reflected radar power of the basal interface.

  2. We found that the thickness of the SPLD ranges from 0 to 3.7 km, and the total volume is about 1.60 × 106 km3. While most basal interfaces appear relatively flat, numerous regions have subsurface depressions, likely caused by impact craters or other basins. We observe a group of unusually deep radar interfaces that we interpret to be part of a distinct, separate unit from the rest of the basal interface. These deeper interfaces appear to merge with previous subsurface radar detections of the Hesperian-aged DAF, suggesting that the DAF forms a continuous unit under the SPLD.

  3. Areas with relatively high basal echo power are widespread throughout the SPLD. The process(es) that are causing these anomalous bright basal reflections may not be limited to regions with potential subglacial water bodies in Ultimi Scopuli. The signatures of these bright basal echo power regions are independent of frequency in MARSIS data. Although the presence and stability of liquid water at these locations is debated, our improved maps of basal topography can help assess the hydraulic potential of these bright basal reflector locations.


The authors would like to thank Andrew Dombard, Jennifer Whitten, and an anonymous reviewer for formal reviews that significantly improved the manuscript. Additionally, the authors thank Daniele Bellutta and Yonggyu Gim for assistance in constructing the 3D volumes of MARSIS data, and Than Putzig, Isaac Smith and Aaron Russell for providing the locations of the CO2 ice deposits detected by SHARAD. The authors are also grateful to Sarah Rogers and Alex Huff for very useful feedback and advice. Some of this work was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration.

    Data Availability Statement

    MARSIS data are available at the ESA Planetary Science Archive (https://archives.esac.esa.int/psa/#!Table%20View/MARSIS=instrument) and at NASA's Planetary Data System (https://pds%2Dgeosciences.wustl.edu/missions/mars%5Fexpress/marsis.htm). Data from the figures in the main text and the supporting information are available in Khuller and Plaut (2021) (repository: https://doi.org/10.5281/zenodo.4653741).