Radar evidence of subglacial liquid water on Mars

We wish to thank the authors of the letter for the opportunity to clarify the points raised by their comments. We will reply by quoting each sentence from the letter in their original order.

> "The widely publicized interpretation (Orosei et al. 2018) of a synthetic aperture radargram as evidence of liquid water at the base of the martian south polar layered deposits (SPLD) draws on two lines of argument:"

This sentence contains some factual inaccuracies. The paper is based on the analysis of data from 29 orbits, in which the subsurface was probed simultaneously at two different frequencies (see Table S1 in the Supplementary Materials). Our identification of liquid water as the most plausible explanation for the observed strong basal echoes in an area about 20 km across centered at 193°E, 81°S, is based on data from nine intersecting orbits, for a total of 18 radargrams in which the strong basal reflector was consistently observed at different frequencies, altitudes and illumination conditions. Also, quoting from the Supplementary Materials, "data processing on Earth consisted of range compression and geometric calibration to compensate for altitude variations. In our analysis, SAR (synthetic aperture radar) processing was not performed because of the smoothness of the SPLD in this area, which causes surface echoes to originate solely from the specular direction; in this case, SAR processing would be reduced to a simple moving average of nadir echoes."

> "the brightness of the reflection itself, and a model by the authors of this letter that argues for the plausibility of briny sludge lubricating the base of the North PLD (Fisher et al. 2010, J. Geophys. Res. Planets 115, E00E11)."

This is an incorrect representation of our arguments. The Fisher et al. (2010) model was not used in the interpretation of radar data. We refer to such paper in the final remarks, to support the statement that perchlorates are ubiquitous on Mars and that they can strongly depress the freezing point of water (see for example Kounaves et al., 2010). Removing the citation from the paper would not affect the interpretation and the conclusions. Our analysis and data interpretation are based on radar signal features and on the computation of the permittivity, which is the physical quantity that can discriminate between dry and wet soil, sludge or pool of liquid water (Guéguen and Palciauskas, 1994). The approach we used to analyze and perform data interpretation has been extensively tested and widely accepted in detecting and mapping subglacial lakes and basal melting in Antarctica and Greenland (see for example, Carter et al., 2007; Ashmore and Bingham, 2014).

Our work is based on the discovery and processing of strong radar echoes coming from the bottom of the SPLD, and on the thorough discussion of their possible origin from an electromagnetic point of view. We conclude that they are most likely caused by the presence of an extended basal feature possessing a dielectric permittivity higher than the one of dry natural materials, thus inferring the presence of water. The work is solely based on experimental radar data that are analyzed by means of an electromagnetic model accounting the propagation within the SPLD. Theoretical models of temperature conditions within the Martian polar caps have been only mentioned when discussing the implications of our findings.

> "Without the model the radar evidence does not stand alone,

We strongly disagree with this statement. For example, Oswald and Gogineni (2008) have identified subglacial water beneath the Greenland ice sheet using solely the intensities of basal echoes in radar data, and have advocated the application of their method for all available observations independently from the specific radar characteristics. Their Figure 7 shows histograms of basal echo power that are surprisingly similar to panel B of Figure 4 in our paper: the difference in mean echo power between dry and wet areas in both figures is nearly the same, being about 15 dB. A signal power of ?10–20 dB higher than the surrounding bed regions is considered to be indicative of basal water (see, for example, Palmer et al., 2013). Furthermore, Rutishauser et al. (2018) have recently used eight radar transects to identify two subglacial hypersaline lakes in the Canadian Arctic: their relative reflectivity is 10 to 15 dB higher than the surrounding terrain. Their study also proves that liquid water can exist in unfavorable geothermal conditions and at low basal temperature (in the case of these lakes about -15°C to -18°C) if the water has a very high salt content. We regret that the Rutishauser et al. (2018) discovery was announced after the final version of our article was accepted for publication.

> "as a bright reflection below a stratified medium like the SPLD can derive from any number of scattering phenomena.

This statement is too generic and no detailed justification is presented to support it. The authors of the letter should explain what is meant by "scattering phenomena". As mentioned above, the paper hinges on the analysis of all propagation and reflection effects that might conceivably explain the observed strong basal echoes. Using quantitative arguments, we have gone to great lengths to demonstrate that the stratification of the SPLD cannot explain the observed strength of basal echoes (see for example figures 4, S4, S5 and S6 in Orosei et al. 2018 and its accompanying Supplementary Materials). Dismissing this work with such an unqualified statement is in our view unacceptable, and we request that proper consideration is given to our quantitative analysis. In particular, we expect that any present or future criticism to the conclusions of the paper needs to be based on a plausible alternative explanation arising from the quantitative electromagnetic analysis of the radar echoes, which, for the considered study, have the following properties:

- The scattering phenomena, including constructive interference from plane parallel layers within the SPLD, are strongly dependent on frequency and geometry of observation, as it can be seen for example in Figure S5 of the Supplementary Materials. Therefore, a particular layer arrangement producing an enhanced basal echo at 3 MHz could be almost ineffective in doing the same at 5 MHz (Figure S5, panel C). By contrast, strong basal echoes at 193°E, 81°S are observed in all nine orbits intersecting over the area, for a total of 18 radargrams acquired at three different frequencies (3, 4 and 5 MHz), and a variety of altitudes and illumination conditions (Figure S3);

- Figure S5 illustrates the strongly oscillatory trend of basal echo power enhancements produced by coherent scattering effects. A layer arrangement capable of enhancing basal echo power by a similar amount at three different frequencies is nearly impossible to find, leading to the prediction that the enhancement of basal echo strength cannot follow a simple trend with frequency. In contrast to such expectation, we observe that basal echoes become weaker as frequency increases both inside and outside the bright reflector, as it can be deduced from the distributions of basal echo power values shown in panels A, B and C of Figure S4. This trend implies that attenuation within the SPLD increases with frequency, which is the expected behavior of most natural materials including dust-laden water ice.

> "Unfortunately, the application of our model to the SPLD observation has not been done correctly. Orosei et al. assume a surface temperature of 160K and the postulated perchlorate brine located 1.5 km below the surface requires a temperature >200K. This implies a thermal gradient of at least 26.7K/km, nearly twice the value of 15 K/km used in our model.

In our paper, we have not attempted to model the thermal structure of the SPLD. Geothermal models of Mars are not supported by any in situ measurement and therefore they are poorly constrained (see, for example, Wieczorek, 2008).

On Earth, where heat fluxes measurements have been extensively performed and the different parameters entering the heat equation are much better constrained, the geothermal models can still be quite inaccurate, especially under the ice sheets in Antarctica and Greenland. Geothermal flux is the least well-known component of the heat equation, and it constrains the temperature at the bottom of the ice sheet (see for example, Llubes et al., 2006). Indeed, recent direct measurements of geothermal flux at the base of the West Antarctic Ice Sheet (Fisher et al., 2015) showed that regional geophysical and glaciological models might fail at the local scale, as this study measured values of the geothermal flux that were three times larger than those predicted.

Physical models should be changed if they disagree with experimental measurements. Experimental measurement do not need to be validated by models. Radio Echo Sounding measurements performed on both Antarctica and Greenland to detect subglacial lakes and meltwater have been widely used to compute the local and regional geothermal flux and to better constrain thermal models (see for example, Schroeder et al., 2014).

> "In calculating dT/dz = G/k, we took the heat flow G to be 0.03 W/m2 and the thermal conductivity k to be 2.0 W/m-K. The resulting value of 15 K/km is itself an upper limit, as a more recent global mean estimate for G is 0.020 W/m2, with regional variations unlikely to exceed ± .006 W/m2 (Parro et al. 2017, Sci. Rep. 7: 45629), and k is likely to be closer to 3.2 W/m-K, the value for pure ice over the range 160-200K.

We argue that the assumption for the SPLD of a uniform thermal conductivity equal to that of pure water ice may be inadequate. For example, Putzig et al. (2005) found that the surface of the SPLD has a very low thermal inertia, implying the presence of a very porous layer that could act as a thermal insulator and increase significantly the temperature at the base of the SPLD. Mellon and Phillips (2001) have assumed a value of 0.045 W m-1 K-1 for a dry, unconsolidated soil in their study of the melting of subsurface ice as a source of liquid water for Martian gullies. Although this is not proposed as a possible explanation for the presence of water beneath the SPLD, we note that such value of thermal conductivity would produce a temperature increase of 40 K at a depth of less than 100 m for a geothermal flux of 0.020 W m-2. Also, Whitten and Campbell (2018) find that SHARAD data over the study area present a diffuse echo pattern they name "fog", blurring subsurface reflections and originating from an unknown surface or volume scattering mechanism at the wavelength of SHARAD (which is shorter than those of MARSIS). This is suggestive of the presence of fractures, voids and other irregularities within the SPLD that do not show in MARSIS observations, likely because they are small compared to MARSIS wavelength, but that would very likely lower thermal conductivity.

> "Periodic climate change can exaggerate the gradient, but the temperature will not change by more than a few degrees at 1.5 km depth. Thus, while the conclusion is tantalizing, the Orosei et al. argument does not "hold water" from a thermophysical perspective.

An experimental measure does not have to be in agreement with a theoretical model, while any model should take into account experimental data. Therefore, we maintain that theoretical arguments against the presence of liquid water cannot be used to dismiss the experimental evidence in favor of it. We are open to criticism and to the discussion of our findings, but any future debate needs to start from a clear understanding of our work, which is focused on a careful quantitative electromagnetic analysis of MARSIS radar echoes.

In closing, we wish to thank again the authors of the letter for giving us the opportunity to better illustrate our methods, in what we perceive as a common effort to understand the nature of the Martian subglacial environment.

References

Ashmore D. W., Bingham, R. G. 2014. Antarctic subglacial hydrology: current knowledge and future challenges. Antarct. Sci. 26, 758-773.

Carter, S. P., Blankenship, D. D., Peters, M. E., Young, D. A., Holt, J. W., Morse, D. L. 2007. Radar-based subglacial lake classification in Antarctica. Geochem. Geophys. Geosyst. 8, Q03016.

Fisher A. T., Mankoff, K. D., Tulaczyk, S. M., Tyler, S. W., Foley, N. 2015. High geothermal heat flux measured below the West Antarctic Ice Sheet. Science advances 1, e1500093.

Guéguen Y., Palciauskas, V. 1994. Introduction to the Physics of Rocks. Princeton Univ. Press.

Kounaves, S. P., Stroble, S. T., Anderson, R. M., Moore, Q., Catling, D. C., Douglas, S., McKay, C. P., Ming, D. W., Smith, P. H., Tamppari, L. K., Zent, A. P. 2010. Discovery of natural perchlorate in the Antarctic Dry Valleys and its global implications. Environmental science & technology, 44, 2360-2364.

Llubes, M., Lanseau, C., Rémy, F. 2006. Relations between basal condition, subglacial hydrological networks and geothermal flux in Antarctica. Earth and Planetary Science Letters 241, 655-662.

Mellon, M. T., Phillips, R. J. 2001. Recent gullies on Mars and the source of liquid water. Journal of Geophysical Research 106, 23165-23180.

Oswald, G. K. A., Gogineni, S. P. 2008. Recovery of subglacial water extent from Greenland radar survey data. J. Glaciol. 54, 94-106.

Palmer, S. J., Dowdeswell, J. A., Christoffersen, P., Young, D. A., Blankenship, D. D., Greenbaum, J. S., Benham, T., Bamber, J., Siegert, M. J. 2013. Greenland subglacial lakes detected by radar. Geophysical Research Letters 40, 6154-6159.

Putzig, N. E., Mellon, M. T., Kretke, K. A., Arvidson, R. E. 2005. Global thermal inertia and surface properties of Mars from the MGS mapping mission. Icarus 173, 325-341.

Rutishauser, A., Blankenship, D. D., Sharp, M., Skidmore, M. L., Greenbaum, J. S., Grima, C., Schroeder D. M., Dowdeswell J. A., Young, D. A. 2018. Discovery of a hypersaline subglacial lake complex beneath Devon Ice Cap, Canadian Arctic. Science advances 4, eaar4353.

Schroeder, D. M., Blankenship, D. D., Young, D. A., Quartini, E. 2014. Evidence for elevated and spatially variable geothermal flux beneath the West Antarctic Ice Sheet. Proceedings of the National Academy of Sciences 111, 9070-9072.

Whitten, J. L., Campbell, B. A. 2018. Lateral Continuity of Layering in the Mars South Polar Layered Deposits From SHARAD Sounding Data. J. Geophys. Res. Planets 123, 1541-1554.

Wieczorek, M. A. 2008. Constraints on the composition of the martian south polar cap from gravity and topography. Icarus 196, 506-517.

The widely publicized interpretation (Orosei et al. 2018) of a synthetic aperture radargram as evidence of liquid water at the base of the martian south polar layered deposits (SPLD) draws on two lines of argument: the brightness of the reflection itself, and a model by the authors of this letter that argues for the plausibility of briny sludge lubricating the base of the North PLD (Fisher et al. 2010, J. Geophys. Res. Planets 115, E00E11). Without the model the radar evidence does not stand alone, as a bright reflection below a stratified medium like the SPLD can derive from any number of scattering phenomena. Unfortunately, the application of our model to the SPLD observation has not been done correctly. Orosei et al. assume a surface temperature of 160K and the postulated perchlorate brine located 1.5 km below the surface requires a temperature >200K. This implies a thermal gradient of at least 26.7K/km, nearly twice the value of 15 K/km used in our model. In calculating dT/dz = G/k, we took the heat flow G to be 0.03 W/m^2 and the thermal conductivity k to be 2.0 W/m-K. The resulting value of 15 K/km is itself an upper limit, as a more recent global mean estimate for G is 0.020 W/m^2, with regional variations unlikely to exceed 0.006 W/m^2 (Parro et al. 2017, Sci. Rep. 7: 45629), and k is likely to be closer to 3.2 W/m-K, the value for pure ice over the range 160-200K. Periodic climate change can exaggerate the gradient, but the temperature will not change by more than a few degrees at 1.5 km depth. Thus, while the conclusion is tantalizing, the Orosei et al. argument does not "hold water" from a thermophysical perspective.

Michael H. Hecht, MIT Haystack Observatory, 99 Millstone Rd., Westford, MA 01886
David A. Fisher, University of Ottawa, Dept. of Earth Sciences, Ottawa, ON K1N 6N5, Canada
David C. Catling, University of Washington, Box 351310, Dept. Earth & Space Sciences, Seattle WA 98195. USA.
Samuel Kounaves, Tufts University, Dept. of Chemistry, Medford MA 02155.