Problemet Torne-Issjôn

Click to increase image sizeClick to decrease image size

Summary

The old conception of the relations between the water divide and the ice divide in the Scandinavian mountain range—the Scandes—was that they lay far apart during the melting stage of the inland ice. The big valleys, which, on the Swedish side of the Scandes, mainly face east or south-east, were therefore presumed to have held ice-dammed lakes. Traces of such lakes have been described by A. G. Högbom in 1892 and 1910, O. Sjögren in 1909, A. Gavelin in 1910, J. Frödin in 1914 and others.

The northernmost of all these supposed ice-dammed lakes was situated in the Torneträsk valley. Early in this century Otto Sjögren made a close study of the shore-lines all round Lake Torneträsk. This is at present 342.1 m (1 122 feet) above sea level, and Sjögren's highest shore-line is at the 653 m level. The peaks in the same area are between 1 200 and 1 700 m (about 3 400 and 5 600 feet) high. Sjögren referred to the highest shorelines as “nunatakk shore-lines”. He also distinguished a lower group of shore-lines marking “Glacial Lake Vassijaure”, and finally a third large group marking “Glacial Lake Torneträsk”.

To understand Sjögren's theory it is important to remember his idea that the ice receded from the west to the east in the Torneträsk valley. He thought that the ice tongue in retreating eastwards left a growing ice-dammed lake in front of itself. The first stage of this resulted in “Glacial Lake Vassijaure”. This drained through the Riksgränsen Pass (about 520 m above the sea) to the Atlantic. When the ice tongue had retired about 20 kilometres east of Riksgränsen, a new col for the drainage was ready north of Polno Bay, the western end of Lake Torneträsk. The canyon north of Polno Bay was according to Sjögren eroded by the outflow from a vast ice-dammed lake, which was extended towards the east as the glacier tongue retreated in the same direction. Finally the immense body of water collected in “Old Torneträsk” broke through the damming ice and established the present drainage at Tarrakoski (342 m above sea level).

Sjögren's theory, briefly summarized above, was attacked by A. Gavelin already in 1910. Gavelin considered that there had been “only one big ice-dammed lake of Torneträsk”, which had drained through the canyon north of Polno Bay. Thus the canyon was already cut when the glacier tongue receded from it.

In this paper I will try to show that there has probably never been any “Glacial Lake Torneträsk” at all. So far, I have spent three summers in the district in order to investigate the dissipation of the last ice sheet. The ancient shore-lines have served to elucidate first the order in which the glacier disappeared from the different valleys and secondly its rate of wasting, more particularly in the main valley of the area—the Torneträsk valley.

According to the old theory the whole area between the railway and the Vadvetjakko mountain near Riksgränsen was submerged. A lot of small lakes are now left, but the area is dominated by naked rock and small shallow swamps. It is thus possible to study large portions of the supposed bottom of the ice-dammed lake of Vassijaure, where one would expect to find deposits spread over the same area as the ancient lake. But I have found that lake terraces have only been built up at the mouths of the tributary valleys, and that there has been no sedimentation outside the terraces.

Further, I have seen kettles in the terraces corresponding to the outlet at Riksgränsen (map fig. 1), and many more in the terraces in higher levels. All the terraces, from the highest to the lowest, must therefore be regarded as what Flint calls “a series of ice-contact slopes” (Flint 1929). The same arguments apply to the eastern parts of the Torneträsk district.

The best evidence against the old theory will, however, be found in the lateral drainage channels. I have determined the gradients of these near the east end of the present Lake Torneträsk, and found them to vary between 1.2:100 m and 3:100 m towards the east. The slope ought to have been towards the west if the glacier tongue had moved into the Torneträsk valley from an easterly ice divide.

Several students have looked upon the canyon north of Polno Bay as incontrovertible evidence of a vast ice-dammed lake, the outflow of which must have cut it down in the rock. The first to mention this in the literature was Suess in 1888. In his view the insignificant brook at its bottom had been the real active factor. Twenty-one years later Sjögren considered that “Bardu Canyon”, as he called it, had been proved to be solely due to erosion by the outflow from “Glacial Lake Torneträsk” during the last deglaciation epoch. A. Gavelin (1910) and Wråk (1916) disputed this interpretation, but even modern students have in main accepted Sjögren's theory.

New possibilities for investigating the canyon have been offered by air photographs of this northern area. The attached anaglyphic picture shows the relative dimensions of the canyon. It is 5 km long, a little less than 300 m wide at the rim, and between 100 and 150 m deep. Two main directions of the crevices can be seen in the picture, one in the same direction as the canyon, i.e. N—S, and the other practically at right angles to this. The canyon is situated in the lowest part of an old rather wide and shallow valley between the mountains of Jalggisvare (west) and Riksochokka (east of the canyon), the tops of which are 6 km apart. The present water divide is about 1 200 m south of the canyon, and the small brook from this enters the canyon in a series of steep rapids. A string of narrow lakes occupies the bottom of the southern part of the canyon. Its southernmost part forms a 350 m long and 60 m wide “appendix”. I have walked all along the bottom of the canyon and then observed ice-polished rock in its east wall. Its west wall is mostly hidden by debris. The difference in the degree to which the western and eastern canyon walls are broken down is interesting. It is due to the mainly westerly winds having caused the snow to accumulate on the leeward side, i.e. on the eastern slopes of the mountains. Many patches of snow regularly remain on these slopes throughout the summer in these parts of Lapland. The meltwater from them will therefore ooze down towards the western wall of the canyon, keeping this wet and more easily broken up by frost. On the eastern side of the canyon, on the other hand, no snow remains after the melting season early in June.

The V-section of this deep and narrow valley has caused everybody to classify it as an entirely water-eroded gorge. But if the loose material were transported away, for instance by ice, a typical U-profile would be the result. This may also be seen in the anaglyphic picture, if one looks at the dip of the rock, which is practically vertical.

It is hardly possible to fix the real age of the canyon. Its “appendix” may be of late glacial origin, but its main part is probably a joint product of all the glaciation periods. The striation in this part of the Torneträsk region indicates that the glacier ice moved in a direction parallel to the canyon. This ought to have facilitated the removal of loose material from the canyon and to erode and polish its walls and bottom. The retiring glacier first left the mouth of the canyon, and its meltwater gushed down it, eventually reaching its southernmost part. The climatic conditions during the recession of the inland ice dissipated it rapidly. The yearly melting periods were long and intense, the full recession took comparatively few years. The time needed for the glaciers to recover the same area at the beginning of every glaciation period was, on the other hand, much longer, causing a remarkably large number of ablation periods. The erosion of the canyon may have been much greater during the initial than during the final stages of every glaciation period.

The name “Bardu Canyon” was originally chosen because it was mistakenly supposed to be part of Bardu Valley. But the canyon is the innermost part of Sördalen (the “South Valley”), a tributary to Bardu Valley which enters this 22 km north of the canyon. That is why I propose that it should be called the Sördalen Canyon.