Two modes of origin are suggested by Rapp (1959,40). The first is depositional in nature. The corestone acts to brake debris being transported by avalanche which then rolls over the top to be deposited on the distal side. In the second the corestone protects debris from erosion so that in time the tail protrudes above the surface of the boulder tongue. The identification of a two-tailed feature led Rapp to favour the former mechanism. Potter (1969,161) considers the work of Bagnold (1941,190) on sand shadows in desert environments to be instructive. His model for the deposition of sand by wind is shown below.
Not everyone agrees with the depositional hypothesis. Luckman (1977,40) states 'Evidence from the Canadian Rockies suggests an erosional origin because of their smooth graded shape (which would not be expected from avalanche deposition, Potter 1969), parallelism to the direction of avalanche movement, their position in the erosional parts of the track (Gardner, 1970), and, in some cases obvious lichen trimming on the header boulder.
Luckman's account appears to be confused on several points. Potter explicitly states that in his opinion debris tails are of a depositional origin. On page 161 he writes `It is clear that the erosional mechanism of origin....is not satisfactory because it does not explain why double tails are sometimes found.' Development parallel to the avalanche direction will occur for either mechanism, hence the analogy to crag-and-tail moraines and sand shadows. Gardner's study remains the only piece of work that provides reasonable evidence for an erosional origin.
RESULTS
Morphometric and particle size data were collected from thirty two debris-tails on boulder tongues on the side of Sentinelnosa, Spitsbergen located at approximately 78o50' north, 16o20' east. It was found that the proximal deposit was of a much greater length than expected and the distal deposit was smaller. The corestone appeared to exert a much closer control on the length of the proximal deposit, the corestone height was the dimension most closely related to proximal deposit length. This deposit appeared to develop at a constant average angle of deposition of some 8.6o, much less than both the angle of repose of talus and the angle of the boulder tongues themselves. Except for isolated cases there was a lack of significant difference in, particle size and shape between the background samples, the proximal deposit and the distal accumulation.
A MODEL FOR DEBRIS TAIL FORMATION
An alternative model for debris tail formation was proposedon the basis of these results. The proximal deposit was believed to come about primarily through the action of dirty wet-snow avalanches. The relatively large size of this deposit was attributed to the greater relative frequency of wet avalanches compared to dry in the high arctic. Wet avalanches are the more important in geomorphological terms as their high frictional resistance with the ground surface results in a greater propensity for debris entrainment.
An attempt was made to explain the low average angle of deposition. Research upon the propagation of plastic waves in snow, pressures following snow impact and the rate of deposition of material following avalanching was used to calculate the average angle of deposited talus for various angles of snow accumulation in front of the corestone. The table below displays the results obtained. While these figures are less than 8.6o, they are much closer to this figure than the angle of repose of talus fragments (over 45o). Because the effect of compaction by additional collisions was excluded and the fact that simple calculation procedures were used a large error was to be expected. It was felt that the figures below provided a fair representation of how a low angle could result from the action of wet-snow avalanches.
| Angle of repose of snow (degrees) | 45 | 50 | 60 | 70 | 80 | 85 |
| Average angle of deposition (degrees) | 0.686 | 0.754 | 0.969 | 1.265 | 2.789 | 5.545 |
Talus transported by wet avalanches was also believed to contribute to the distal deposit. Following Rapp (1959) this was felt to be due to material passing over the corestone. While some evidence was found for this mechanism, the sorting in the distal deposits where found, was across the deposit, with smallest particles found in the centre. Lateral deposition would thereforeappear be more important than `overtopping'. The sorting was presumed to result from smaller particles being able to maintain buoyancy in the turbulence behind the corestone at a lower velocity.
Recourse to further work upon sand shadows since Bagnold's pioneering study led to the idea that horseshoe vortices may develop around the corestone during avalanching. These turbulences form around an object protruding into the flow of any current with a vertical velocity gradient such as the wind, or a river flow. Snow avalanches also exhibit this vertical gradient. They cause enhance accumulation of material behind the obstacle and erosional zones in the immediate surround when operating upon fine material such as sand. it is not known if they are of sufficient force to disturb talus-sized debris, measurement during an event would be required to ascertain if this is the case.
CONCLUSION
The model proposed here for the formation of debris tails provides an explanation for many observations made both in this study and elsewhere. The greater relative length of the proximal deposit can be explained by the greater relative frequency of wet avalanches with a high debris content to clean dry avalanches in the high arctic. Dry events with their greater velocity generate stronger vortices able to extend the distal deposit more efficiently. Within the framework of this model Gardner's observations, (Gardner 1970) are not inconsistent with the depositional hypothesis. It would be quite possible for an area of the boulder tongue to be eroding and yet for deposition to occur behind the corestone due to the vortices generated bypassing avalanches. The complex nature of turbulences would mean the corestone's control upon the distal deposit would be much less simple to determine than for the proximal. This was the case when the morphometric data was analysed.
To support the model proposed here more knowledge is required on the dynamics of snow avalanches, in particular the possible generation of horseshoe vortices and the methods of debris entrainment and deposition. This study supports the depositional origin of debris tails originally proposed by Rapp(1959). The sand shadow model employed by Potter (1969) was found wanting in its original form, but by considering the action of vortices about the corestone, an explanation can be provided form any of the issues raised.
This work has shown the usefulness of simple techniques in studying landform processes and made a contribution to the study of these little known landforms.
BIBLIOGRAPHY
Bagnold R A (1941) 'The Physics of Blown Sand and Desert Dunes.'
New York. William Morrow & Co.
Gardner J S (1970) 'Geomorphic Significance of Avalanches in the Lake Louise District.'
Arctic and Alpine Research. vol 22 p.135-144.
Luckman B H (1977) 'The Geomorphic Activity of Snow Avalanches.'
Geografiska Annaler vol 59A (1-2) p.31-48
Potter N (1969) 'Tree Ring Dating of Snow Avalanche Tracks and the Geomorphic Significance of Avalanches, Northern Absaroka Mountains, Wyoming.'
Geological Society of America Special Paper 123 p.141-165
Rapp A (1959) 'Avalanche Boulder Tongues in Lappland.'
Geografiska Annaler vol 41 p.34-48