Previous Page Table of Contents Next Page

Avalanches - lessons in defence and control

Colin Fraser

Colin Fraser is the author of The Avalanche Enigma, a detailed history and study of avalanches and man's efforts to understand and control them. He has served as an avalanche consultant and worked with the Swiss Federal Institute for Snow and A Avalanche Research.

Although there is a long history of avalanche disasters, the development of effective techniques for controlling and preventing avalanches is relatively new. Switzerland and Austria have done pioneering work at considerable expense, yet newcomers to this field continue to ignore their lessons and to repeat old mistakes.

Snow avalanches, which can certainly be numbered among the great destructive forces of nature, have historically caused more death and damage in Europe than in other parts of the world. This is logical when it is considered that the European alps were the first snow-covered and mountainous region to become fairly densely inhabited. In recent years, however, many other mountainous areas of the world have been opened up as roads are built through them, or as mineral wealth is exploited, or as opportunities for winter tourism present themselves. And it seems safe to assume that the energy crisis will in future give rise to the construction of more and more dams in mountainous and snowy regions as nations are forced to exploit their latent hydroelectric resources. Which poses the question: what is known about avalanches and avalanche control ?

1. This triangular, 2-hectare patch of forest above the Swiss village of Andermatt was first preserved as an avalanche defence by legislation in 1397. In recent years snow support structures were built to supplement its effect.

A recent science

It is only in the last 30 to 40 years that avalanche control has progressed from a haphazard practice to an applied science founded on a basic understanding of snow mechanics. The pioneering work, however, proved to be considerably more expensive than anticipated. Among the countries where avalanche control has been studied Switzerland and Austria have had the most success. It has been particularly important for them because of the position they occupy in the centre of a dense European road system. Anyone wanting to learn about avalanche control could not do better than to study the experiences and techniques in these two alpine countries. It would be a good way to avoid repeating some expensive false starts and mistakes already made there and elsewhere. This is not to say that methods of avalanche control have now reached an optimum level of efficiency; more mistakes will doubtless be made even in the countries where the techniques are most advanced since it is still a relatively new technology. A brief resume of this experience may be helpful, but first, and in order to put avalanche control into perspective, one should have some idea of the nature and behaviour of avalanches themselves.

Johan Coaz

The processes that cause a mass of snow to break away and hurtle down a slope are exceedingly complex. It is true, however, that the numerous factors involved tend to combine to cause avalanches in certain locations - certain slopes or gulleys - more often than in others. For this reason, the Swiss forester Johan Coaz - who because of his observations and studies in the late nineteenth and early twentieth centuries has often been called the father of avalanche research-was able to report that there were 7 486 avalanche tracks that carried one or more avalanches each winter in the Swiss alps. This figure referred only to avalanches that affected the population in one way or another and excluded those at very high altitudes.

Obviously, an avalanche which recurs at least once a year down a certain track becomes a well-known local phenomenon and people steer clear of it. Unfortunately, however, avalanches are not always so regular in their behaviour; the complex of factors that go to create them may unite on some slopes only once in many decades, or even centuries. In the winter of 1950/51, a particularly severe avalanche period in the European alps, several houses which had stood untouched for more than 500 years were destroyed. And in January 1968, I was personally involved in a rescue operation after an avalanche hit some new houses in an area which had had no recorded avalanche since 13 March 1609

ln the middle ages this erratic and capricious behaviour of avalanches led to a fatalistic attitude among alpine populations: avalanches were often considered to be acts of God or the devil, and almost nothing was done to combat them.

Early legislation

Woodland was, however, recognized as affording protection, particularly if there were trees in the avalanche starting zone or if there was insufficient open slope above the tree line for an avalanche to gather the momentum necessary to destroy mature forest. For this reason, in many parts of the alps there were edicts protecting certain well-defined and strategically placed areas of forest. The first of these edicts were issued in the mid-fourteenth century and there were 322 of them still in force in Switzerland when the federal authorities took them over from the cantonal authorities in 1896. Not all the edicts were as carefully drawn up as they should have been; some failed to prohibit grazing by farm livestock in the woodland and, perhaps just as serious, the total ban on felling of trees prevented even the judicious work by foresters necessary to keep the woodland young and vigorous.

Early avalanche defences were usually mounds of earth and rock placed against the upslope side of a house, with the purpose of dividing the moving snow around it. Later, stone deflection walls were sometimes built, the first dating from the mid-sixteenth century after a disastrous avalanche had killed 61 people in the village of Leukerbad (Valais) in 1518. The wall was not sufficiently effective, however, because exactly 200 years later the avalanche struck again and killed 55 people.


The tendency is to build such structures in unbroken lines

Preventing avalanches is more expensive than it look.

Holding the snow

All early avalanche defences were passive in nature; they did nothing but attempt to reduce or eliminate damage from a mass of snow already in movement. The idea of going to the root of the problem, preventing an avalanche from even starting at all by supporting and holding the snow cover on the slope, came relatively late. Early attempts, beginning in the eighteenth century, used trenches and terraces dug across the slope. By the 1860s Johan Coaz was organizing the building of stone walls across the slope in avalanche starting zones. Neither of these systems were very effective: terraces were soon levelled out by snowfalls, whereas stone walls-because they completely break the wind flow-tended to cause drifting of snow so that the space behind them was quickly filled.

It was only in this century that the techniques for control in the starting zone were refined; the Austrians set a trend when, in the 1920s, they extended the stone walls protecting the Arlberg railway by using heavy fences supported by steel uprights. Fences of one sort or another were soon found to be cheaper and more effective than walls or terraces and by 1939 had entirely replaced earlier structures. Fences can be thought of as racks to support the snow cover, although they are also expected to halt any small slides that may start between them.

One could suppose that the design of support structures and their erection in the avalanche break-away zone would be relatively simple matters, but they have proved astonishingly complex. Up to the late 1930s all progress was based on empirical knowledge. There was no basis for calculating in advance the amount of stress the snow cover would impose on a fence; no one knew which materials were the most suitable or how the structures should be deployed on the slope to best advantage. The men responsible used their own judgement and any materials conveniently available - anything from railway sleepers and lengths of rail to steel girders and specially cut timber.

3. 1907 - Construction of avalanche defences over key roads and railway lines and to preserve mountainside vegetation is an old tradition in Switzerland as evidenced by this 1907 photograph. It shows stone snow walls on the slope above the Rhätische Bahn near Albula Grison.

4. 1957 - The same slope above the Rhätische Bahn 50 years later and following reforestation. Old snow walls on the upper levels are still useful, others among the trees have been allowed to fall into disrepair.

But in 1938 a group of scientists working in Davos, at what was to become the Swiss Federal Institute for Snow and Avalanche Research, produced a basis formula for determining the snow pressure on a fence, and this opened the way for the first calculated designs. So much further knowledge has been and is being accumulated that a complete guide to the design and installation of snow support structures is issued by the Swiss Federal Forestry Inspectorate. The handbook, compiled by the staff of the Avalanche Research Institute, runs to some 60 pages of diagrams, formulas, calculations and text.

5. Snow bridge

Everything is covered: there are formulas for determining the correct siting and spacing of the fences; specifications for materials and designs; instructions for calculating snow pressure on a given slope and at given altitudes; minimum permissible strength margins for different structures over and above the calculated stresses to be imposed on them; details of how to anchor the fences in the slope according to different geological circumstances; how to test the soil's ability to withstand the pressure imposed by the foundations, etc. It has been found of great value in the United States where it was translated into English. In Switzerland the provisions of the handbook are mandatory where subsidies are to be granted for avalanche control, because defence work is so expensive that no mistakes in planning or execution should be tolerated.

6. Snow rake

Modern fence-type support structures are usually 3 or 4 metres high and are of two basic kinds - snow bridges (Fig. 5) and snow rakes (Fig. 6). The snow rake is slightly superior because all the uprights can be embedded in the ground for extra strength. It also holds the snow better because it has more members running contrary to the stratification of the snow cover.

Influencing the slide

One of the main prerequisites of a snow support structure of either type is that it should influence the way a snowfall is deposited. That is to say, it should not cause drifting. To ensure this the bars of the structure must be well spaced, but not so widely that any quantity of the snow cover can slip between them. By a process of elimination it has been found that the compromise gap is 30 to 38 centimetres.

The early fences were always erected vertically, but this placed unfavourable stresses on the anchorage and also reduced their effective height. It was later found that it is much easier to provide a firm anchorage when the fence is given a slight downhill bias, and it is usual today to set them to make an angle of 100-105° with the slope on the uphill side (Fig. 7). This loads the foundations more favourably and makes full use of the structure's height.


Early snow-support structures (right) were set incorrectly. Effective height was lost and the pressure of snow tended to push the base of the fence up and away from the slope. Modern snow-support structures (left) correct these mistakes. The structure is set at an angle of 105 degrees to the slope to make better use of the height of the fence and to distribute the snow load more favourably for the foundation. In addition, stronger supports are used behind the fences.

A recent practice is to place the support structures in unbroken lines across the slope, but if an interrupted arrangement is used-as in the majority of schemes in existence-the maximum permissible gap between structures is 2 metres, unless a feature of the terrain definitely precludes an avalanche starting at that point. Attempts to cut costs by wider spacing of the units proved to be false economy; small avalanches were able to run down between the fences, tearing out the odd one here and there. Once a breach had been made larger avalanches tore more units out, and so it would have continued until, had the units not been replaced, the whole defence scheme would have been reduced to a litter of broken structures at the bottom of the slope. The maintenance costs on schemes that had been skimped in the first place were therefore very high. The current tendency is to spend more initially so as to eliminate the risk to the fences caused by small avalanches starting among them.

Spacing is important

The spacing of the fences up and down the slope is equally important. The supporting effect of a structure is only felt upslope for a distance of two to three times the depth of the snow. Using this knowledge, the slope angle, the degree of roughness of the ground, and several other factors, there is a formula for calculating the minimum spacing between the structure, up and down a slope. To exceed this spacing is dangerous because a small avalanche could start, pile up against a fence and so form a ramp for other avalanches. In theory, a succession of avalanches could form a ramp against fence after fence, rendering them ineffective and creating an avalanche slidepath.

The conclusion that there is no substitute for close spacing of the structures, both up and down and across the slope, was reached after some very costly experience.

The materials commonly used today for support structures are steel, aluminium and prestressed concrete. Wood has a relatively short life and is therefore not much used, except below the beeline and where replanted timber will be well enough established in 30 years to take over the structures' task of avalanche control. Reforestation is an integral part of all avalanche defence schemes which extend below the beeline.

The costs of avalanche control in the starting zone are very high; the Swiss now calculate that every metre of structure installed on a slope costs about 1250 Swiss francs (about US$400). Using the rough rule that a hectare of slope will normally require four lines of structures set 25 metres apart up the slope, and each 100 metres long, we arrive at a figure of about Swiss Fr.500 000 or $160 000 for every hectare of avalanche starting zone in which the snow cover is to be stabilized. Much of this figure is for labour, so costs therefore vary from country to country. In Italy, for instance, the same work costs about $100000 per hectare, which is still a high price to pay. But the cost of ignoring avalanche defence can be even more expensive. During the winter of 1950/51 avalanche damage in Switzerland was estimated at $5.5 million, and in Austria during the same period at $4.4 million.

It might be thought that European countries at least would have been quick to apply the same norms that the Swiss federal authorities insist upon if they are to grant subsidies for avalanche control, especially as the Swiss handbook is freely available and now exists in several languages. However, some European countries that have begun avalanche defence work relatively late are still going about it as though they were pioneering in an unknown area. In some instances they are wasting their resources.

Splitting wedges

The high costs of avalanche control in the starting zone make it necessary to look also at alternative control measures when possible. The medieval system of protecting individual installations from snow already in movement can sometimes be used. For example, splitting wedges built of concrete - but not of stones and earth as in past centuries - are very effective in protecting installations such as electricity pylons. Roads and railways are often protected by galleries or snow sheds which carry the sliding snow of an avalanche across them. In Europe such galleries usually cost from $4 000 to $5 000 per metre to build, but they are often cheaper as road protection than snow support structures high up in the avalanche starting zone. However, avalanche defence galleries have to be built with a solid engineering background in snow mechanics. I have seen some, built by a bilateral aid agency in a developing country, which were next to useless. The snow slid over the roof, which had no overhang, and landed on a flat area outside the gallery instead of moving on down the slope. Snow piled against the supporting pillars then slipped, or was blown by high winds, into the gallery, where it effectively blocked the carriageway for long periods each winter. It would have been sufficient to bulldoze away the flat area next to the gallery, or to provide the roof with a large overhang in order to ensure that the snow continued its slide down the slope.

The cheapest method of avalanche control for thoroughfares that can be closed for certain periods is the use of explosives or artillery. During or immediately after heavy snow storms, and with the thoroughfare closed, light artillery shells or charges are exploded in the avalanche starting zones with the purpose of provoking an avalanche; if none occurs, at least it is certain that the situation is safe until such time as there is a change in the weather and/or snow conditions. However, control with explosives calls for a great deal of knowledge and experience, and a fine sense of judgement and organization on the part of those entrusted with the decision as to when to close a thoroughfare and shoot or blast. For many developing countries, however, it would be a valuable technique to exploit, even if on an interim basis until avalanche defence systems can be built.

To newcomers in avalanche control I would like to repeat: countries that have not embarked on avalanche control hitherto-either because of lack of means or lack of need-and intend to do so in the foreseeable future would do well to examine the expensive and hard won experience gained in the European alps before beginning their own programmes.

Previous Page Top of Page Next Page