How can we predict and prevent avalanche
risks? What scientific facts can be taken as the starting point?
Some of the prevention methods used in the past - and still applied
today - have shown their limitations. One example is nivo-meteorological
forecasting, based on data obtained from nivology (the study of
snow) and meteorology; all too often evacuations planned with this
method have gone on to prove unnecessary. Avalanche maps, which
attempt to combine all the spatial and spatio-?temporal data for
a given site, are not much better and are of little use for forecasting.
Bridging the gaps
The EU's SAME (Snow Avalanche Modelling and Mapping in Europe)
project was set up with the aim of correcting these inadequacies.
It involved close international and interdisciplinary cooperation
between 14 partners over a 30-month period to explore three main
lines of inquiry: avalanche information systems, warning systems
and modelling, and sensor testing
'These lines of inquiry reflect the areas of expertise in the research
teams' explains project coordinator Gérard Brugnot, Natural
Risks Delegate at Cemagref in Grenoble (France). 'Some scientists
focus on measurements in the field, others on models, others on
These approaches also reflect national differences. 'That is no
doubt to do with the fact that management of this kind of natural
risk is closely linked to town and country planning and decision
support, and thus to a political perspective. Before the SAME project,
everybody was convinced that they had the right system. One of the
merits of this project is the way it enabled us to bridge the gaps
between one country and another. The exchanges between researchers
were very rewarding, each one recognising the value of the others'
Data, language and models
The first stage of the project was to collect an extensive variety
of data on avalanches, including site maps, incident files, procedures,
etc. Once gathered, they then had to be classified and harmonised.
At this stage, terminology proved a major problem. This was overcome
by creating a multilingual glossary of the terms used to model and
describe avalanches and to record all the relevant details.
The SAME project also made progress in the field of modelling:
an inventory of about 50 different models was compiled, giving their
characteristics and conditions of application. This allowed researchers
to construct a single computerised platform capable of using several
different avalanche models for risk forecasting.
Some of these were field-tested at five sites in France, Italy,
Switzerland, Spain and Norway - each with a well-documented avalanche
history. 'We can conclude that there is clearly no miracle model.
Each one of them reflects certain parts of the full picture on the
ground and this is why it was so interesting to be able to compare
them. The researchers were able to identify the complementary elements
- all of which will have to be taken into account in the future.'
Artificially triggered avalanches at specific test sites equipped
with measurement instrumentation, too, have allowed better measurement
and understanding of the internal dynamics of snowslides. A variety
of detectors, radar instruments and sensors of all kinds were also
tested at the same sites. This included precise analysis of a series
of parameters and testing the effectiveness of systems for warning
everyone in imminent danger.
'The accidents during the winter of 1999 were, in a way, expected.
But much more precise forecasting would have been needed to avert
them, and that is something we are not yet able to do,' says Gérard
Brugnot. 'To make progress, we need to know more about the mechanisms
that govern the build-up of snow - wind and precipitation - instability
within the snow cover, and the flow of avalanches. The models constructed
on the basis of this knowledge could then be checked by experiments
under controlled conditions and compared with field data provided
With further progress in mind, the Cadzie project has been started
to follow up SAME and address the new concerns raised by the February
1999 avalanches. The objective of this research, which is focusing
on exceptionally catastrophic avalanches, is to improve zoning and,
in parallel, make defences more effective and more reliable.
Avalanche mapping, model validation and warning systems (SAME)
Environment and Climate (ENV 2C)
Fax : +33-4-76762711
- Norwegian Geotechnical Institute, Oslo, Norway
- Laboratoire d'Instrumentation en Microinformatique et Electronique,
Université de Grenoble I, Grenoble, France
- Faculté de Géologie, Université de
Barcelone, Barcelone, Espagne
- Centre d'Etudes de la Neige, Centre National de Recherches
- Forstliche Bundesversuchasanstalt, Institut für Lawinen-
und Wilbachforschungf, Innnsbruck, Austria
- Instituto de Cartografia de Catalunya, Barcelone, Spain
- Ecole Polytechnolique Fédérale de Lausanne,
- Space Research Institute - Russian Academy of Sciences,
- Swiss Federal Institute for Forest, Sow and Landscape, Davos,
- Icelandic Meteorological Office, Reykjavik, Iceland
- Technische Universität Graz, Graz, Austria
- Università degli Studi di Pavia, Pavie, Italy
- Centro Sperimentale Valanghe e Difesa Idrogeologica, Arabba
- Università degli Studi di Torina, Turin, Italy
Simulation of the February 1999 MontRoc avalanche in the Chamonix
area (calculation and image production: Mohamed Naaim).