Introduction

The north coast of Ellesmere Island, from the mouth of Nansen Sound to Clements Markham Inlet (Fig.1), is fringed with ice, including first-year sea ice, multi-year sea ice and ice shelves. Many of the ice shelves, which are unique to this coast, grew as a result of thick sea-ice formation, which remained fast to the coast. The periodic disintegration of the ice shelves creates ice islands, which are generally carried with the pack ice towards the Beaufort Sea. With the likely exploration and development of offshore oil and gas reserves in the Beaufort Sea during the next decade, ice islands and thick sea ice fioes could be a hazard to offshore installations. In order to understand this hazard, it is beneficial to have accurate information about the ice conditions on the north coast of Ellesmere Island. In 1980, 1982, 1983, and 1984, we travelled extensively by snowmobile and aircraft along this coast, with the object of surveying and mapping recent ice-edge changes and the response of the landfast sea ice and ice shelves to these changes. This paper documents recent ice calvings and landfast ice growth and considers the relationship between the sea ice and the ice shelves, on the coastline between Yelverton Bay and C. Albert Edward.

Fig. 1. Map of northern Ellesmere Island, NWT. The largest remaining ice shelves (stippled areas) are shown, together with prominent capes, points, bays, inlets, and fiords on the north coast.

Recent Ice Edge Changes and Landfast Sea Ice Growth

1) Yelverton Bay

The most likely source of ice island T-3 is Yelverton Bay, where the former ice shelf disintegrated in about 1935 (Reference Hattersley-SmithHattersley-Smith, 1957). The bay is now almost entirely filled with 6 m-thick landfast ice, within which a few, small, shelf-ice fragments are trapped. About 85 km² of in situ ice shelf remains in this area: 1) 55 km² in a bay east of Hanson Pt. (Fig.2); 2) 20 km² at the west side of the bay (Fig.1); 3) 10 km² in the south-west of the bay (not illustrated). The ice front generally extends in a direct line Trom C. Evans to Alert Point (Fig.1), although minor changes occur as a result of periodic calvings. The most recent change occurred in 1984, when about 30 km² of ice broke away from Area 1 (Fig.2).

Fig. 2. Map of ice shelves and landfast ice (Areas 1, 2, 3 and 4) between Yelverton Bay and C. Fanshawe Martin, 1984. The question marks in Ayles Ford indicate that the boundary between the landfast ice and the fiord ice is unknown.

2) Cape Evans to Cape Egerton

These capes mark the mouth of Milne Fiord, which contains Milne Ice Shelf (Figs.2 and 3). The latter has an area of about 300 km² and is up to 100 m thick (Reference PragerPrager, 1983). In 1964, a small calving event, at the front of the ice shelf, created a 35 km² ice island, with a volume of about 2.5 km³ (Reference JeffriesJeffries, 1985). Since then, the area from which the ice island calved has been filled with an equivalent area of land fast ice that is up to 10 km thick (Area 2, Figs.2 and 3). North-east of Area 2, towards C. Egerton there is a narrow (~ I km wide) fringe of landfast ice along the front of the ice shelf.

Fig. 3. Oblique air photograph, looking across the mouth af Milne Fiord from C. Evans, July 1984. Note the difference in scale between the undulating topography of Milne Ice Shelf (MIS) and Area 2. (Source, MOJ.)

3) Cape Egerton to Cape Fanshawe Martin

The narrow, landfast-ice fringe, along the north-east front of Milne Ice Shelf, continues past C. Egerton to C. Bicknor and along the front of Ayles Ice Shelf (Fig.2). At some time during the interval 1959-74, Ayles Ice Shelf moved about 5 km out of Ayles Fiord and away from the east shore (Figs.2 and 4). During the same period, a 15 km² ice island, with a volume of 0.8 km³, calved from the ice shelf (Reference JeffriesJeffries, In press).

Ayles Ice Shelf now has an area of about 55 km² and the location from which the ice island broke away (Area 3, Figs.2 and 4) is filled with land fast ice. The movement of the ice shelf away from the shore created a space (Area 4, Figs.2 and 4), which now contains landfast ice, together with fragments of shelf ice. The landfast ice extends to the rear of the ice shelf, but it is not known how far. Further into the fiord, the ice cover is non-saline and contains numerous small icebergs, derived from a disintegrating ice tongue.

Fig. 4. Oblique air photograph, looking across Ayles Ice Shelf (foreground) and Area 4 to C. Fanshawe Martin (F), C. Richards (R) and M’Clintock Inlet (5), July 1984. The ice in Areas 3 and 4 has a similar topography to that in Area 2 (Fig.3) and contrasts with the larger-scale undulations on the ice shelf. (Source MOJ.)

5) Cape Richards to Cape Discovery

In June 1962, the outer 10 km of M’Clintock Inlet (Figs.4 and 5) were covered by the unbroken M’Clintock Ice Shelf, which had disintegrated by April 1966 (Reference Hattersley-SmithHattersley-Smith 1967). Between C. Richards and C. Discovery, there now exist only two small areas of in situ shelf ice: 1) 35 km², west of Bromley Island; 2) 6 km², in a small bay, north of C. Discovery (Fig.5). The mouth of M’Clintock Inlet is now filled with 10 m-thick landfast ice (Area 5, Fig.5), Numerous fragments of old shelf ice, with rocks and boulders on their surfaces, are trapped within the sea ice. South of Borup Point (Fig,5), the sea ice merges with the fiord ice, which also contains fragments of old shelf ice. The position of the ice front, between the capes, varies due to occasional small calvings. In 1982, the landfast ice extended up to 5 km north of Bromley Island, but a calving of up to 50 km² of ice, during the interval 1982-83, left the ice front as illustrated in Figure 5.

Fig. 5. Map of ice shelves and landfast ice (Areas 5, 6 and 7), between C. Richards and C. Albert Edward, 1984. The question marks in M’Clintock Inlet indicate that the boundary between the landfast ice and the fiord ice is unknown.

6) Cane Discovery to Cape Albert Edward

In August 1961 (Fig.5), Ward Hunt Ice Shelf covered an area of over 1000 km² and had a volume of 38 to 44 km³. Between August 1961 and April 1962, a massive calving caused an ice loss of almost 600 km², with a volume of 18 to 24 km³ (Reference Hattersley-SmithHattersley-Smith, 1963). By 1980, it is estimated that about 75 km of ice shelf had regrown, as a result of sea-ice accretion at its front (e.g. Area 6, Figs.5 and 6).

Fig. 6. Oblique air photograph, looking east along the front of Ward Hunt Ice Shelf, July 1984. The landfast ice of Area 6 has an undulating topography, similar to that of Areas 2, 3 and 4, but on a much smaller scale than on the ice shelf. Note the small ice island (1) at lower left. (Source, MOJ.)

The ice-shelf front remained quite stable until the 1980s. During the interval 1980-82, approximately 40 km² of shelf ice and fast ice grounded and/or calved at the extreme west end of the ice shelf, immediately north of the C. Discovery Ice Rise (Fig.5) (Reference JeffriesJeffries, 1982), In 1982-83, up to 40 km² of shelf ice and fast ice calved from the ice-shelf front, between C. Albert Edward and the Ward Hunt Ice Rise (Area 7, Figs.5 and 6) (Reference Jeffries and SersonJeffries and Serson, 1983), In spring 1984, Ward Hunt Ice Shelf covered an area of about 440 km² and had an estimated volume of 20 km³.

Of the sea ice which accreted after 1963, only the 45 km² Area 6 remains (Figs.5 and 6). However, in spring 1984, small areas of second-year sea ice had already accreted, west of C. Albert Edward (Fig.5).

Discussion

Over the past 50 years, there has been a considerable loss of shelf ice from the north coast of Ellesmere Island, the greatest losses having occurred at Yelverton Bay, M’Clintock Inlet, and Ward Hunt Ice Shelf. Ice calvings occur every third or fourth year, at apparently random intervals (Reference Sackinger, Shoemaker, Serson, Jeffries and YanSackinger and others, 1985), but it is difficult to predict the timing, location, and magnitude of calvings. Furthermore, although some calvings have been attributed to tidal and seismic events (Reference HoldsworthHoldsworth, 1971), little İs known about the causes of ice-shelf disintegration. From Yelverton Bay to C. Albert Edward (Fig.1), there are now about 920 km² of in situ ice shelf, 80% of which includes the Ward Hunt and Milne Ice Shelves. The same stretch of coast includes about 901 km² of landfast ice, most of which is found in Yelverton Bay and M’Clintock Inlet, The landfast ice has grown since each calving and is evidence that the ice shelves are quickly replaced. In cases such as Area 2 (Figs.2 and 3), Area 3 (Fig.2), and Area 6 (Figs.5 and 6), where landfast ice has accreted at the front of the ice shelves, the ice growth could be considered as ice shelf regeneration.

In Figures 3, 4, and 6, it can be seen that the landfast ice, like the ice shelves, has a similar, but smaller-scale, undulating topography, parallel to the coastline and prevailing winds. On the sea ice, the ridge tops are approximately 60 m apart and the troughs up to 1 m deep, in contrast, the ice-shelf ridges vary from 200 m to 300 m apart and the troughs are as much as 7.5 m deep. The origin of the ice-shelf rolls is unclear, but could have been due to pack-ice pressure and pressure-ridge formation, during early, sea-ice growth (Reference Hattersley-SmithHattersley-Smith, 1957; Reference ListerLister, 1962). Subsequently, as the ice thickened, the rolls deepened and became oriented parallel to the prevailing winds, which drove melt water along the troughs (Reference Hattersley-SmithHattersley-Smith, 1957; Reference CraryCrary, 1960; Reference ListerLister, 1962). At the present time, there is an adequate supply of melt water, due to the relatively warm summers and consequent negative, surface mass balance, which accumulates in depressions as melt ponds. Some areas οΓ the landfast ice are mechanically deformed and this would account for the ridges and troughs. However, most of the ice is undeformed and, in this case, the ridges and troughs probably develop as a result of wind action and melt-pond coalescence. As the ice thickens and ages, the ridges and troughs develop a sub-parallel pattern which is evident in Figures 3, 4, and 6.

The undeformed, landfast ice varies from 2 m to 10 m in thickness, with most of the ice being at the thicker end of the range. Undeformed, multi-year ice in the Arctic Ocean commonly has a steady-state thickness of 2.5 m to 5 m (Reference Maykut and UntersteinerMaykut and Untersteiner, 1971), but it has been shown that undeformed sea ice can reach thicknesses of at last 12 m, under present climatic conditions (Reference Walker and WadhamsWalker and Wadhams, 1979). Clearly, this is the case with the landfast ice common to the north coast of Ellesmere Island, which, as Reference Hattersley-Smith, Vasari, Hyvärinen and HicksHattersley-Smith (1972) suggested, grows up to the point where it can be regarded as incipient ice shelf.

The salinity and oxygen-18 composition of present landfast ice is evidence of a strong, fresh-water, run-off influence, which leads to fresh, brackish, or saline ice accretion, according to the prevailing meteorological and oceanographic conditions (Reference JeffriesJeffries, 1985). Present landfast-ice growth has been likened to the brackish, basement ice, described by Reference Lyons, Savin and TamburiLyons and others (1971), and is believed to have preceded the original, basement, ice growth (Reference JeffriesJeffries, 1985). The latter sea ice forms the core of Ward Hunt Ice Shelf, west of Ward Hunt Island, and is evidence of thick, sea-ice growth, which occurred about 4000 years ago, when the climate was cooler, with much less run-off than at present (Reference Jeffries and KrouseJeffries and Krouse, in press). A climatic cooling trend, continuing to the end of the present century, has been predicted by Reference Johnsen, Dansgaard, Clausen and LangwayJohnsen et al. (1970) and Reference Hibler, Langway and DunbarHibler and Langway (1977). The response of the ice shelves to this trend might be further expansion, as a result of lateral fast-ice accretion, which will increasingly resemble the original, basement ice. If this is so, the ice shelves might be expected to thicken and become more extensive, in the decades to come. Although the ice shelves might expand, it is likely that calvings will continue, creating large ice features, either of shelf ice or thick sea ice. Thus, it is suggested that Ayles Ice Shelf, already dislodged from its former, more secure position in Ayles Fiord, could break away completely and become an ice island.