NATURE|Vol 438|8 December 2005 BRIEF COMMUNICATIONS ARISING

PLANETARY SCIENCE Are there active on Mars? Arising from: J. W. Head et al. Nature 434, 346–351 (2005)

Head et al.1 interpret spectacular images water come from? Possibilities on Mars from the Mars Express high-resolution include melting and dewatering of a stereo camera as evidence of geologically hydrous compounds3–7 by heating due to recent rock glaciers in Tharsis and of a pied- increased geothermal gradients, magma A mont (‘hourglass’) at the base of a 3- intrusions or impacts. The long-term stability B km-high massif east of Hellas. They attribute of shallow groundwater and at the latitude growth of the low-latitude glaciers to snow- of the hourglass craters is debatable. But geo- fall during periods of increased spin-axis logically recent gullies heading part way up C obliquity. The age of the hourglass glacier, slopes have been interpreted as evidence for considered to be inactive and slowly shrink- shallow deposits (at depths of less than 0.5 ing beneath a debris cover in the absence of km) of groundwater or ice at similar latitudes, modern snowfall, is estimated to be more including locations nearby8. Moreover, aufeis than 40 Myr. Although we agree that the feeding of glaciers is not directly related to maximum glacier extent was climatically climate fluctuations, raising the possibility N controlled, we find evidence in the images to that some of the glaciers identified by Head et support local augmentation of accumulation al.1 remain active. b from snowfall through a mechanism that We find support for the aufeis hypothesis in does not require climate change on Mars. a Mars Orbital Camera high-resolution image Head et al.1 identify an accumulation area of the hourglass glacier (Fig. 1c), which shows for the hourglass glacier in an ‘alcove’ above a band of light-coloured, crevassed ice that is its upper crater (Fig. 1a). However, the arcu- within about 250 m of the uphill edge of the ately banded ice that originated from the glacier. It seems to be flowing from the slope canyon draining most of the alcove is topo- at the southern edge of the crater to the west graphically lower than, and appears to be of the canyon. Downslope, the crevassed ice pinched out by, flow from either side (http:// becomes progressively covered by dark debris. esamultimedia.esa.int/images/marsexpress/ The Mars Orbital Camera image cited by 180-170305-0451-6-an-01-Hourglass.jpe). Head et al.1 shows only the distal, lightly Hence, the greatest ice flux must have come cratered, inactive part of the glacier. Because N from small areas at the foot of the adjacent ice on Mars sublimes rapidly until it is cov- slopes and not from the larger area drained ered, the exposed and crevassed ice in the by the canyon, as would be expected for a proximal glacier must be active. The presence snowfall-fed glacier. This suggests that the of relatively bare, subliming ice and the c most recent source of ice was not associated absence of an accumulation area raise the with snowfall in the accumulation area, even possibility that water, or ice, is erupting on the Crevassed ice though earlier it may have been. martian surface today. What is the source of the most recent ice? We conclude that martian glaciers could be Talus slope Insight may come from another apparently fed from above or below. This could be by ice-filled double crater (Fig. 1b) that lies on the snowfall during climatically favourable times, flat plain only 35 km away and at the same ele- as argued by Head et al.1, or by extrusion of vation as the distal crater, according to Mars groundwater, as we argue here. These two Orbiter Laser Altimeter data (http://ltpwww. mechanisms are not mutually exclusive and Figure 1 | Images interpreted to show aufeis on gsfc.nasa.gov/tharsis/mola.html). Despite the their relative contributions may differ spatially Mars (257.3 W, 39.5 S). a, High-resolution absence of potential accumulation areas, sim- and vary temporally: during climates that sup- stereo camera (HRSC) image (about 20 m per ilarly banded and presumably debris-covered port snowfall, both may operate; at other pixel) of the upper basin of the ‘hourglass’ glacier icy flows still fill the craters and point to a times, diminished glaciers may still be and the mouth of the canyon (location A) draining the largest part of the proposed source within each crater. The observation supported by aufeis. We consider the relative accumulation area (not shown); inferred aufeis that these glaciers lack distinct accumulation contribution of snowfall and extrusion mech- source areas are near locations marked B and C. areas, together with the absence of glaciers anisms to the growth and extent of glaciers as Images are taken from Mars Express orbit 451. in many comparable locations elsewhere on an outstanding question in martian . The overlay (white outline) is part of higher- the surrounding plain, argues against a solely Alan R. Gillespie, David R. Montgomery, resolution (4.2 m) Mars Orbital Camera (MOC) climatic origin. Amit Mushkin image, M1102128. Black box, area enlarged in c. These observations are consistent with the Quaternary Research Center and Department of b, A second ‘hourglass’ glacier from the same glaciers having been formed by groundwater Earth and Space Sciences, University of HRSC image, about 35 km north of the area erupting from exposed aquifers, or from frac- Washington, Seattle, Washington 98195-1310, USA shown in a. c, Enlargement of the area in a black box in the MOC overlay in a. The reticulate e-mail: [email protected] tures in an icy regolith, and then freezing near pattern is interpreted as a field below the the surface (aufeis). This idea was proposed by talus slope. Arrow showing flow direction spans Garlick2 for a palaeoglacier that originated 1. Head, J. W. et al. Nature 434, 346–351 (2005). 2. Garlick, G. D. Geol. Soc. Am. Cordilleran Section abstr. 20, the light-coloured, relatively bare ice that from a fissure on Arsia Mons. 162 (1988). becomes increasingly covered farther north. Where could such large amounts of ground- 3. Sharp, R. P. J. Geophys. Res. 78, 4073–4083 (1973). Scale bars: a, 2 km; b, 5 km; c, 300 m.

E9 BRIEF COMMUNICATIONS ARISING NATURE|Vol 438|8 December 2005

4. Luchitta, B. K. Icarus 45, 264–303 (1981). 7. Montgomery, D. R. & Gillespie, A. Geology 33, 625–628 morphologically similar to the floor deposits 5. Komatsu, G. et al. Lunar Planet. Sci. XXXI abstr. 1434 (2005). across the entire frame; we interpret its rela- (CD-ROM) (Lunar and Planetary House, Houston, 8. Malin, M. C. & Edgett, K. S. J. Geophys. Res. 106, 2000). 23429–23570 (2001). tive brightness to be due to a dust cover that 6. Kargel, J. Mars — A Warmer, Wetter Planet was formed on the margin of the encroach- (Springer–Praxis, Chichester, UK, 2004). doi:10.1038/nature04357 ing talus slope. Nearby craters with similar scales, hour- glass-like shapes, and lobate flow-like deposits (Fig. 1b of ref. 1) can also form from local PLANETARY SCIENCE crater-wall accumulation zones. In these cases, the crater interior walls are optimum locations for snow and ice accumulation, resulting in Head et al. reply glacier-like, lobate flow down the walls and out Replying to: A. R. Gillespie, D. R. Montgomery & A. Mushkin Nature 438, onto the crater floor, a phenomenon that has doi:10.1038/nature04357 (2005) been well documented in mid-latitude crater interiors5,6. Alcoves along crater walls create an Gillespie et al.1 concur with our interpre- lie in shadow, producing colder temperatures. environment favourable to the collection of tation that certain lobate equatorial and The accumulation zone for the hourglass snow and ice, and may provide steep cliffs that mid-latitude features on Mars are due to deposit on Mars was thought to be the entire act as sources for debris that, together with ice debris-covered glaciers formed largely dur- alcove region east and southeast of the hour- sublimation, create debris-covered glaciers, ing past periods of increased spin-axis obli- glass deposit2 (Fig. 2a, b, d of ref. 2). Snow and much like those that occur in the Mars-like quity, when climate regimes favoured snow ice accumulating in this entire alcove region Antarctic Dry Valleys (Fig. 3a, b of ref. 2). and ice accumulation and glacial flow2. They probably formed distinct glaciers that moved On the basis of these considerations, suggest that the ‘hourglass’ deposit, dated at downslope and converged to form multiple groundwater-fed ‘glaciers’ do not seem to be more than 40 Myr old2, could be active today lobes, the remnants of which are seen in the required in these locations. We agree that this owing to an additional mechanism that upper part of the hourglass. mechanism should be investigated, particu- supports “local augmentation of accumula- Changes in ice velocity, as might occur with larly for deposits formed earlier in the history tion from snowfall” without climate change changes in ice thickness related to deepening of Mars, when thermal gradients were such on Mars. This mechanism requires the pre- of bedrock topography, compressed and that groundwater may have been much closer sent, or very recent, release of groundwater folded the ice, eventually merging all lobes to the surface. to the surface to form aufeis (groundwater- together, then to flow through the neck of the J. W. Head*, G. Neukum, R. Jaumann, fed ‘glaciers’) where the groundwater is gen- hourglass. When climatic conditions changed, H. Hiesinger, E. Hauber, M. Carr, P. Masson, erated by dewatering of hydrous compounds the distribution of snowfall and snow accu- B. Foing, H. Hoffmann, M. Kreslavsky, S. Werner, or melting by magmatic or impact-generated mulation probably also changed, resulting in S. Milkovich, S. van Gasselt, The HRSC heat. We assess whether this suggestion the cessation of ice accumulation and flow; the Co-Investigator Team applies to the deposits in question — it was ice and snow deposits disappeared, and dry *Department of Geological Sciences, Brown previously proposed for much older deposits mass-wasting dominated, producing the talus University, Providence, Rhode Island 02912, USA in other areas of Mars3,4. We make particular piles superposed on top of the glacial deposits e-mail: [email protected] reference to the key relationships in the at the base of the slope (Fig. 2a–d of ref. 2, and accumulation zones. Fig. 1a, c of ref. 1). 1. Gillespie, A. R., Montgomery, D. R. & Mushkin, A. Nature Glacial accumulation zones are areas char- We therefore interpret the configuration 438, doi:10.1038/nature04357 (2005). 2. Head, J. W. et al. Nature 434, 346–351 (2005). acterized by a yearly net addition of ice, and and morphology of the lobes, as well as the 3. Carr, M. J. Geophys. Res. 100, 7479–7507 (1995). are separate from zones, which abrupt contact between the base of the slope 4. Lucchitta, B. K. J. Geophys. Res. 89, suppl.B 409–418 (1984). undergo a yearly net loss of ice. In active and the glacial deposits, to be a natural con- 5. Howard, A. Lunar Planet. Sci. XXXIV abstr. 1065 glaciers and in the source regions for broader sequence of both the shape of the initial (2003). 6. Marchant, D. R. & Head, J. W. 6th Int. Mars Conf. abstr. 3991 piedmont glaciers, the accumulation zones are accumulation zone and its postglacial (2003). commonly centred on steep-sided alcoves that evolution. The candidate aufeis in the high- provide broad traps for wind-blown snow and resolution image (Fig. 1c of ref. 1) appears doi:10.1038/nature04358

E10