ROCK ON JAMES ROSS ISLAND,

Jorge A. Strelin1, Toshio Sone2

1. Instituto Ant‡rtico Argentino and Centro Austral de Investigaciones Cient’ficas, Av. Malvinas Argentinas s/n¼, (9410) Ushuaia, Tierra del Fuego, Argentina e-mail: [email protected]

2. Institute of Low Temperature Science, Hokkaido University, Sapporo 060, Japan e-mail: [email protected]

Abstract

Lack of cover in north-western James Ross Island, favours the development of a number of periglacial landforms. Ice-cored rock glaciers, protalus lobes, and recently discovered protalus ramparts are some of the most conspicuous cryogenic features. The ice-cored rock glaciers appear in a complex and genetically related landform system. Besides their mor- phological characteristics, these landforms are also differentiated by their dynamic behaviour. Mechanisms of ice and debris flow and debris extrusion are discussed in order to ascertain the initial age of the main rock glac- ier formation. Protalus lobes and protalus ramparts, formed at the base of slopes and ephemeral snow patches with no relation to former glaciers, are also typical features of this environment. All these landforms were probably formed after the third Neoglacial, 1300-1000 years BP.

Introduction

Since 1990, the joint Argentine (Instituto Ant‡rtico Argentino) - Japanese (Institute of Low Temperature Science) Group ÒCriolog’aÓ has focused its research on cryological and geomorphological topics in the northern Antarctic Peninsula area. In this paper, we present results of a study in the NW part of James Ross Island (Figure 1).

About 80% of James Ross Island is ice-covered, and most of this is due to the large Mount Haddington . This ice cap stretches over an area 40 km in dia- meter, reaching the highest point of the island at 1628 m (a.s.l.).

Most of the ice-free land is located in the NW sector of the island, where it is isolated from the main ice cap. This area corresponds to a former glacial landscape carved in friable Mesozoic sedimentary rocks covered by Cenozoic volcanics. The latter, mainly basalt and pyroclastic breccias, are preserved as 300 to 900 m (a.s.l.) high remnant plateaus, separated by wide va- lleys. The steep slopes that surround the volcanic plateaus are affected by large , glacier ero- sion, , and snow-debris avalanches. Uninterrupted and intense shattering leads to rock fall, roll, slide, and creep. Figure 1. Location and geomorphological map of the NW sector of James Ross Island.

Jorge A. Strelin, Toshio Sone 1027 The climatic conditions are polar arid to semiarid, and sorted . Most of these cryogenic fea- the location of the island within the area of seasonal tures were morphologically described by Strelin and sea-ice results in maritime influences during the sum- Malagnino (1992). mer and a more continental winter season. In the study area, the mean annual air temperature at sea level is ca. The present work focuses on the morphological and -6.5¡C and the annual precipitation, mostly snow, is morphodynamic aspects of ice-cored rock glaciers estimated to be around 200 mm water equivalent. This (Potter, 1972), protalus lobes (Whalley and Martin, low annual accumulation results in large snow-free 1992) and protalus ramparts (Bryan, 1934; Ballantyne areas during much of the year. The area is affected by and Benn, 1994). The first are partially channelled in cold and wet southwesterly winds (Schwerdtfeger, short valleys and the last two are present at the foot of 1975) and warm and dry west to northwesterly valley slopes. winds (fšhn). The first description of rock glaciers in Antarctica was Small ice caps and glaciers develop respectively at the for southern Victoria Land (Mayewski and Hassinger, top and foot of the volcanic plateaus. The equilibrium 1980). On James Ross Island, close to the present study lines in this sector of the island are at a mean altitude of area, a similar landform was described and alternative- 200 m (a.s.l.). However, variation in factors such as ly called a debris-covered polar glacier and a rock gla- insolation, wind control on snow deposition, exposure cier (Chinn and Dillon, 1987). to fšhn, katabatic winds, orographic precipitation, etc., results in a remarkable equilibrium line variability Lachman II (from 0 to 500 m). The environment described above favours the development of periglacial landforms and Six ice-cored rock glaciers develop at the east foot of deposits such as talus slopes, ice-cored rock glaciers, Lachman Crags (Figure 1). Among these, Lachman II protalus lobes, protalus ramparts, stone banked ter- rock glacier is analysed here in detail. This rock glacier races, nivation hollows, mixed snow and debris is a component of a complex geomorphic system that avalanche deposits and several types of sorted and non- has different temporal and spatial related parts. Two

Figure 2. Geomorphological map of the south-east sector of Lachman Crags.

1028 The 7th International Conference main zones are distinguished (Figure 2): accumulation The comprise recessive ice-cored ridges that and ablation zones. become smoother in the direction of the front of the glacier. ACCUMULATION ZONE This zone is subdivided into a main and a secondary An active ice-cored rock glacier (Potter, 1972) extends accumulation zone. The first one involves ice and snow downvalley of the morainic sector and is obstructed by accumulation on a series of small ice caps situated on a rock glacier of a previous stage that is still active or in the top of Lachman Crags: Norte, Central and Sur ice a steady state. The moraines and rock glaciers, which caps. The second corresponds to a regenerated glacier, enclose this morphological system at its front, consti- principally nourished by ice and debris avalanches and tute the passive ablation zone. wind-drifted snow, situated at the foot of the crag along a 3 km wide front. MORPHOLOGY OF LACHMAN II ROCK GLACIER The following characteristics were observed in the ABLATION ZONE ablation zone of the ÒLachman II glacier-rock glacier Three ice lobes, partially separated by moraines, con- systemÓ (ice tongue, ice-cored moraines and rock gla- stitute the main ablation zone. The central glacier lobe ciers) (Figures 2, 3A and 3B). is the most extended, showing a frontal sector placed much lower than its surrounding moraines. In the main ablation zone, the glacier tongue surface has an average slope of 7 to 8¼. The ice foliation is

Figure 3. (A) Vertical aerial photograph (November 1992), (B) topographic map, (C) morphodynamic map (interval 1992 - 1995), and (D) morphometric map of Lachman II rock glacier.

Jorge A. Strelin, Toshio Sone 1029 Table 1. Parameters measured in the Lachman II rock glacier domain

clearly marked, enabling its photointerpretation and The debris cover of the rock glacier is usually 0.60 to reconnaissance in the field. Sometimes the ice gives 0.80 m thick, but at the outer limit it exceeds 1 m in way to the emergence of till. The flow pattern of the thickness. Permafrost, rather than the , was glacier can be traced through mapping this ice foliation. observed 1.10 m beneath the ground surface, close to E34 (Figure 3C). The central ice tongue is depressed at its front by 10 to 15 m relative to the enclosing ice-cored moraines. The characteristics of the sediments that cover the Downvalley, the moraines lose their shape while the rock glacier depend on their origin. Shattered volcanic debris cover increases. rocks are supplied from the crags and move downslope principally by supra- and intraglacial transport, where- Ridges and furrows develop where the debris on the as Cretaceous, sandy to clayey, loose sediments are glacier ice is at least 0.60 m thick. These flow-like fea- extruded from the glacier sole. This allows identifica- tures mark approximately the transition to the rock tion of the position of the deep shear zones on the rock glacier domain. The rock glacier stretches approximate- glacier surface. The biggest volcanic blocks, that cover ly 700 m downvalley and is roughly 500 m wide. The or crop out from the rock glacier front, reach tens of upper surface dips 4 to 5¼ in the flow direction and ends cubic metres in volume, but the average size does not at a steep talus apron of 24 to 42¼. During the summer, exceed 0.15 m in diameter. Where platy, mostly a stream originating at the front of the ice tongue, dis- basaltic, debris stretches over the surface, an open fab- charges water through a steep channel cut in the ric of strongly-imbricated clasts (dipping about 75¡ moraines and in the central part of the rock glacier. The upglacier) is recorded. The matrix appears approxi- release of additional meltwater is accomplished by two mately 0.20-0.30 m below the ground surface and con- ephemeral creeks that drain off the steep sides of the sists of unsorted gravel and sand. Where the rock gla- rock glacier. Conical holes, sometimes occupied by cier surface is principally nourished by finer material melt water, are spread out on the surface of the rock (sand or tuff), patterned ground (sorted nets and cir- glacier. cles) and lobes develop. The inner fabric of these features shows imbricated gravel, suspended in a Sporadic outcrops of glacier ice are visible at steep finer, homogenous sandy matrix. slopes developed in the central channel, in some of the bigger conical holes, and at the marginal talus. The ice MORPHODYNAMICS OF THE ROCK GLACIER core has a marked foliation, steeply dipping up-valley, Flow and ablation rates measured between January with ÒblackÓ regelated ice layers, including till, alternat- 1992 and January 1995, allow four zones to be identi- ing with prevailing white glacier ice. These layers fied: the glacier tongue, the ice-cored moraines, the range between centimetres to more than a metre in active ice-cored rock glacier and the older rock glacier thickness. The ridges and furrows of the rock glacier stage (Figure 3C and Table 1). surface parallel this foliation trend. At the sides of the glacier tongue, the flow vectors act In the frontal sector, it is possible to distinguish a rem- subparallel to the ice foliation, but in the central and nant landform that possibly corresponds to an older frontal sectors, they are orthogonal. A similar pattern rock glacier advance. This feature attains a length of was observed between flow direction and the ice-cored about 200 m and stretches 40 m over the surrounding ridges. In the rock glacier domain the flow glaciofluvial landscape. An inflection point develops vectors are always normal to the ridge and furrow along the superficial flow pattern (ridges and furrows), trends (Figures 3A and 3D). where the younger and more active rock glacier over- rides the central part of the older one (Figures 2 and The annual ablation rates also fall into four categories: 3A). This emphasizes their different dynamic the ice tongue front (more than 1 m ablation), the ice- behaviour. cored moraines (around 0.5 m ablation), the main rock

1030 The 7th International Permafrost Conference glacier (0.2 m ablation), and the older rock glacier stage glacier movement. For example, if we assume that a (nearly 0.02 m ablation). climatic change produces a fall in the mean annual tem- perature, and no more ice or debris are added to the In this kind of rock glacier it is assumed that the flow system, then the regelation zone will migrate upvalley follows glacier ice deformation without basal slip and debris will be extruded at higher levels of the glaci- (Potter, 1972; Whalley and Martin, 1992). Thus, er tongue. Considering this, the formation of the LliboutryÕs (1965) empirical formula: younger main rock glacier is more recent and possibly postdates the Little Ice Age. The older rock glacier 1 [1] stage follows a previous glacial readvance, probably the 54 3 4 Hmyu= 13. 73 / · (o / sina ) third Neoglacial (1300-1000 years BP; Clapperton and Sugden, 1988). is applicable to calculate the approximate ice thick- ness (H in m). The analytical data, resulting from sur- Protalus lobes and protalus ramparts face velocity (uo in m/year) and inclination (a) in the ablation zone, are consistent with the thickness estima- Protalus lobes and protalus ramparts are common fea- ted by considering the rock glacier topography (Figure tures in the ice-free sectors of James Ross Island. These 3B) and the inferred subglacier valley floor shape. landforms seem to be genetically related and unlike the tongue-shaped rock glacier, they develop without direct

AGE OF THE ROCK GLACIER relation to glaciers. In James Ross Island, the environ- Judging from the landscape around the rock glaciers, ment required for their formation comprises steep talus both (main and older) stages follow several glacier slopes (34-40¡), enough debris supply, and the presence recessive advances that are not older than 2900 radio- of perennial or late-lying snow patches. Most of the carbon years (Fukuda et al., 1992) or 1600 radio-carbon frost-shattered materials are removed from the outcrops years (Rabassa, 1982). by rock slide, roll and fall and transported on the talus slope by frost creep, gelifluction, snow-debris avalan- The mean annual flow velocity of 0.20 m/y and the ches, debris slide and flow, etc. Debris and snow (later travel distance of about 400 m from the position where transformed into interstitial ice) accumulate at the base the debris cover of the rock glacier was generated, sug- of the talus slope producing protalus ramparts or, in the gest that the youngest stage of Lachman II rock glacier case of higher debris input and creep, protalus lobes. is approximately 2000 years old. Similar rock glaciers MORPHOLOGY OF PROTALUS LOBES are described in Iceland (Whalley and Martin, 1992; Two zones with active protalus lobes are shown in Whalley et al., 1995). Using present day surface veloci- Figure 1. They are placed at the north-west foot of ty data and length, these authors determined an age of Lachman Crags and Cerro Triple. Like the tongue- 2000 years for one rock glacier located in Tršllskagi shaped rock glacier, a steep talus apron that dips (Whalley et al., 1995). However, licheonometric data around 35¡, develops laterally and at the front of these show 200 years for the oldest part of that rock glacier. features. In this case the landforms extend much more The authors explain this inconsistency by considering laterally, perpendicular to the flow direction, than rock that the ÒwrongÓ age of 2000 years results from a glaciers: usually they extend a few hundred metres pa- markedly reduced velocity, due to a decrease in the ice rallel to the slope and 10 to 50 m downward. They are thickness of the rock glacier core, over the last 10 to 20 m higher than the surrounding landscape. The 200 years. slope angle of the upper surface declines rapidly downslope, becoming sub-horizontal at the front. This situation could also be the case in Lachman, but Some times a levŽe-shaped form with an open debris the resulting older age can be explained differently. fabric is recognized at the frontal part. Ridges and fur- Subpolar glaciers are frozen to their beds and rest over rows are poorly developed. Unlike the ice-cored rock permafrost in their frontal areas. Due to this obstruc- glaciers, the debris source areas (plateau or valley tion of flow, debris rises to the surface through shear slopes) of the protalus lobes are larger than their sur- zones. There is good evidence of this ice flow condition faces. No depressions of any type nor other evidence of in the Lachman II rock glacier area. If the same climatic ice core existence were detected, but it is probable that conditions persist, no change in the position of the rege- the lobes are ice-cemented. lation zone will occur, and the polar glacier will show emergence of debris along a restricted transverse zone. Except for the occasional occurrence of large blocks, In the case of climatic variation, the regelation front debris sorting and imbrication of the protalus lobes are (Figures 3C and 3D) will migrate and different parts similar to that of the tongue-shaped rock glaciers. along the ice tongue surface will be subject to debris Some of the protalus lobes are covered with moss and ÒextrusionÓ. Following this process, an extended ice lichens and appear to be inactive. area will then be covered by debris without significant

Jorge A. Strelin, Toshio Sone 1031 inclined 24¡ downhill and reached the inner slope of the 5 m below its top (Figure 4). Both sides of the rampart are very steep, sloping at about 40- 50¼. An incipient protalus lobe emerges from the outer slope. The upper 1.5 m of the ridge exhibits 0.10 to 0.30 m large angular volcanic fragments arranged in an open fabric. At a lower level, finer grained, unsorted matrix appears 0.10 to 0.20 m under the slope surface. As shown in Figure 4, the height of the combined land- forms reaches a maximum of 10 m above the valley floor.

AGE OF PROTALUS LOBES AND RAMPARTS No fossil protalus lobes nor ramparts were observed. Their activity seems to depend on abundant debris sup- ply and the presence of perennial or semi-permanent snow patches on the valley slopes. Normally, they occupy valley and crag sides, related to higher and older shaped landscapes than the ice-cored rock gla- ciers. Thus their initiation date is less accurately esti- mated than for rock glaciers. The similar weathering of some inactive protalus lobes and the older tongue- shaped rock glacier, indicates that they probably started to form at about the same time (i. e. after the third Neoglacial, 1300-1000 years BP; Clapperton and Sugden, 1988). Figure 4. Photograph of the protalus rampart of Cerro Triple showing the cross section AB sketched below (for scale, compare with sketch). Conclusions More precise characterization of these features will A Òglacier-rock glacier systemÓ (Lachman II) with dis- result from the determination of the surface flow pat- tinct morphological characteristics was identified on tern during future summer seasons by tracking marked James Ross Island, northern Antarctic Peninsula area. blocks accurately referenced to fixed stations. In the ablation zone, the glacial ice tongue, ice-cored moraines, a main rock glacier and an older rock glacier MORPHOLOGY OF PROTALUS RAMPARTS are morphodynamically distinguished. On James Ross Island, protalus ramparts are usually associated with protalus lobes. One of these features, Based on the of the landscape relat- located at 100 m a.s.l., stretches like a 150 m long belt ing to these rock glaciers, they were formed after two that parallels the NE hill-foot of Cerro Triple (Figures 1 recessive glacier advances occurred at 1300-1000 and and 4). It is usually overridden by debris derived from 200 years BP. respectively. the scree slope and sometimes it gives way, laterally and at its front, to incipient protalus lobes. Where the As shown by Whalley et al. (1995), an initial age of the gap between the scree slope and the rampart is overrun rock glacier formation obtained using the by debris, the feature changes its outline to a definite distance/superficial speed rate was incorrect. protalus lobe (cf. Whalley and Martin, 1992). Reduction of the ice thickness or migration of the rege- lation zone could explain this inconsistency. A cross-section of the protalus rampart is sketched in Figure 4. The talus slope that feeds this landform has Protalus lobes develop unrelated to glaciers at the foot gradients of 36-40¼ and is composed of unsorted angu- of talus slopes and appear to be ice-cemented. Protalus lar small blocks, gravel and very coarse-grained sand, ramparts are mentioned herein for the first time in the generated by shattering of pyroclastic breccia and tuff. Antarctic Peninsula area. Their genesis is related to protalus lobe formation. The rampart has a sinuous track and was separated 10 to 15 m horizontally from the talus slope by temporary In the future, an ice-debris balance, correct modelling snow patches. In mid-February 1997, one of the snow of the flow mechanisms, and a better knowledge of cli- patches showed a maximum thickness of as much as matic parameters, will help to clarify the behaviour of 2 m, but it did not exist during the very dry summer of these landforms and their susceptibility to climatic 1995. In the summer of 1997, the snow surface was change.

1032 The 7th International Permafrost Conference Acknowledgments

We thank the Instituto Ant‡rtico Argentino (Buenos Aires), the Centro Austral de Investigaciones Cient’ficas (Ushuaia), Argentina, and the Institute of Low Temperature Science, Hokkaido University (Sapporo), Japan, for the logistical and technical support of the Project: ÒThe Permafrost on the Antarctic Peninsula AreaÓ. We are specially grateful to Daniel Martinioni for revising the English version of this paper.

References

Ballantyne, C.K. and Benn, D.I. (1994). Glaciological con- Rabassa, J. (1982). Estratigraf’a de los dep—sitos glacigŽnicos straints on protalus rampart development. Permafrost and en el sector norte de la isla James Ross, pen’nsula Periglacial Processes, 5, 145 - 154. Ant‡rtica, Argentina. In 5¼ Congreso Latinoamericano de Geolog’a, Argentina, 4, pp. 237-251. Bryan, K. (1934). Geomorphic processes at high altitudes. Geographical Review, 24, 655-656. Schwerdtfeger, W. (1975). The Effect of the Antarctic Peninsula on the Temperature Regime of the Weddell Sea. Chinn, T.J.H. and Dillon, A. (1987). Observation on a debris- Monthly Weather Review, 103, 45-51. covered polar glacier ÒWhisky GlacierÓ, James Ross Island, Antarctic Peninsula, Antarctica. Journal of Strelin, J.A. and Malagnino, E.C. (1992). Geomorfolog’a de la , 33, 300-310. isla James Ross. In Rinaldi C.A. (ed.), Geolog’a de la Isla James Ross, Ant‡rtida, Instituto Ant‡rtico Argentino, Clapperton, Ch.M. and Sugden, D.E. (1988). Holocene Buenos Aires, pp. 7 - 36. Glacier Fluctuation in South America and Antarctica. Quaternary Science Reviews, 7, 185-198. Strelin J.A. and Sone, T. (1994). Rock Glaciers in Antarctica. IGCP Symposium and Field Meeting, The Termination of the Fukuda, M., Strelin, J., Shimokawa, K. Takahashi, N., Sone, Pleistocene in South America, Centro Austral de T. and Trombotto, D. (1992). Permafrost occurrence of Investigaciones Cient’ficas, Ushuaia, Argentina, pp. 36-38. Seymour Island and James Ross Island, Antarctic Peninsula region. In Yoshida et al. (eds.), Recent Progress in Whalley, W.B. and Martin, H.E. (1992). Rock glaciers: II mod- Antarctic Earth Science, Japan, pp. 745-750. els and mechanisms. Progress in Physical Geography, 16, 127 - 186. Lliboutry, L. (1965). TraitŽ de Glaciologie, 2. Masson & Cie., Paris (1040 pp). Whalley, W.B., Hamilton, S.J., Palmer, C.F., Gordon, J.E. and Martin, H.E. (1995). The Dynamics of Rock Glaciers: Data Mayewski P.A. and Hassinger, J. (1980). Characteristics and From Tršllskagi, North Iceland. In Slaymaker O. (ed.), significance of rock glaciers in southern Victoria Land, Steepland Geomorphology, pp. 129-145. Antarctica. Antarctic Journal of the United States, 15, 68-69. Potter, N. (1972). Ice-cored rock glacier, Galena Creek, north- ern Absaroka Mountains, Wyoming. Geological Society of America Buletin, 83, 3025-3057.

Jorge A. Strelin, Toshio Sone 1033