|
Experiences with the Austrian Powder Avalanche Model
Horst Schaffhauser
fbva.aiatr@magnet.at
Abstract
Most of the disasters in the years of 1974, 1984 and 1988 were caused by flowering avalanche with extremely high powder components. Therefore in 1991 a common research project was started between WLV (Austrian Torrent and Avalanche Control System), the Austrian Institute for Avalanche and Torrent Research (AIATR) at the FBVA (Federal Forest Research Station) and the AVL (Research for Fire Engine) in Graz/Austria developing a three dimensional numerical powder avalanche model for optimizing the hazard zoning. On hand of two events the results are discussed.
Introduction and Objectives
The effects of the powder part influence during the event in St. Anton 1988, which caused seven victims and severe damages was the final impact to develope more detailed avalanche models, based on a 3-dimensional fluid dynamics.
In view of a practice-oriented improvement of the powder-snow avalanche model and for a technical and financial control the Federal Ministry for Agriculture and Forestry established a working group with representatives of the AVL, the WLV and the AITAR to optimize cooperation, the task were divided between those three institutes.
The tasks of the AVL were:
The task of the AIATR included:
Actual WLV projects with the results to be interpreted in cooperation with the representatives of the WLV. Expert opinions were in that case provided by the WLV.
The WLV was in charge of the following tasks:
According to this division of tasks two case studies were used to describe the applied methods and the use of the AVL-model in connection with planning fundamentals for the WLV.
Methodology
The three-dimensional AVL-powder avalanche model based on the fundamental
differential equation-system, which governs the conservation of mass,
momentum, snow concentration and the effects of turbulence are taken into
a two equation turbulence model (Brandstätter W., et al. 1992).
At the beginning there were difficulties regarding the one-phase numerical gas-dynamic simulation model. User-friendliness, in particular the determination of the avalanche starting zones, the generation of the calculation grid and the dimensional representation of the terrain, could be optimized only after numerous test runs. The direct entry of the delimitating coordinates of the avalanche starting zones, the automated establishment of the three-dimensional calculation grid (finite volume method), and the automated transfer of the digital terrain data helped to greatly reduce the time required for the preprocessing phase. Depending on the project requirements, 2 to 4 hours were needed. The digital terrain models are currently being developed at the FBVA in Vienna by the extended planning office (EPS) of the WLV. In its modified form, the routine "Fire" includes a preprocessing part, the main routine and a postprocessing part.
After having entered the preprocessing section, first the area of topographic influence of the potential avalanche, is starting zone over the limiting coordinates, thickness and density of the snow, and the snow mass are entered. With the help of a system utility the distances of the isohypses are modified, markers and the roughness and permeability of buildings (deflection dams, etc) are determined and inscribed using different key numbers (0-1). Before the calculation grid is delimitated geometrically, the center line of the avalanche has to be determined. After the sizes of the grids and the time-related data (variable running time of the simulation, time intervals between 0.1 and 0.4 sec, output interval in seconds) have been entered, calculation grids of up to 6x10 finite elements are automatically generated. The preprocessing phase has thereby been terminated and after securing of the project the programme is ready to run. Immediately after the start, the computing process and the input parameters (mass of snow, iteration steps, gas density) can be checked at any time.
In the post-processing part it is possible to call in the maximum and minimum total pressures, the velocity vectors, and the densities within the three space axes in longitudinal and cross sections as well as in the horizontal projection for each and every selected time interval and for any box number. The scaling is given in [Pa] and [tm-2]. Apart from those forms of representation, a subroutine offers the possibility to illustrate, for the entire time of the avalanche event, total pressure and distribution of density separately for each level of the grid and in the form of colour-coded isolines or iso surfaces. As to the scaling, it is distinguished between the local determination of minimum and maximum values along different longitudinal – cross sections, a global one (the highest and the lowest value of the entire area are selected), and freely selectable variable values. Depending on the number of grid boxes and the terrain relief, there are grid heights of up to 3 meters (finite elements), longitudinal sections of 4 meters, and grid widths of up 5o 5 meters. For the mapping, the scale can be changed at will.
Discussion
of the results
As before mentioned the disaster of St. Anton in March 1988 initiated the
AVL avalanche model. The release zone of the Wolfsgruben avalanche, with
an area of 37 hector is 800 meters above the valley (1284 m.a.s.l.)
situated. The lower part of the avalanche track is canalized. The length
of racture line and fracture width in the release area was determined by
terrestrial photogrammetry with an exactness within the meter range along
the crown line, relating to the fracture thickness within the one
decimeter range. The mean value of thickness was governed with 1.37 m,
with a density of 125 kg/m3 and the calculation grid was filled up with
60.000 mesh cells. Devided in 1000 time steps over the avalanche transit
time the calculation was timed in time steps of 0,20 seconds. The
dissolved degrees of grey represents in figure one the calculated
distributions of the dynamic pressure over the whole event. The computed
results with 13 kPa in figure 1 brought in a good line with the calculated
results (13-17 kPa, depending on the material data of the house) by civil
engineer for static (TSCHOM, H. 1988) on observed damaged buildings
(photo 1). In the same way as before
corresponds the mapped situation of the hazard indicators along the track
(destroyed protective forest by decree).
Before
the installation of avalanche control systems, a part of the ski-station
of Obertauern in Salzburg was threatened in the case of the release of
catastrophic avalanche with a high powder component. Permanent control
measures (avalanche-relanding mounds) in the run-off zone and temporary
systems (bomb trams) led to a protection against to the flowing part. As a
neuralgic zone remained the summit-slope of the Gamsleitenspitze with a
view to the snow-drift and geotechnical problems. Furthermore the
transition from the release area to the track is interrupted by a steep
grade changing of the slope profile by a rock stage. In the frame of an
avalanche control project of the WLV on this site was used at first time
in Austria a new type of avalanche designe, named "avalanche breaker"
(G. FIEBIGER, H. SKOLAUT, 1990)
That kind of snow-pack stabilizing structure – a steel concrete construction – get a vertical height of eleven meters so that the flowing part likewise at extreme events in the upslope snow-retarding basin is innocously deposited. In this respect the knowledge about the dynamic behaviour of the powder part is of an enormous importance for practize in control, hazard mapping and designe studies.
On the example of figure 2 follows plain the influence of the avalanche breaker relating to the fluid dynamic of the powder part. How is shown furthermore in figure two on the level one (three meters above the ground) the flux of the designe of the avalanche breakers and it appears deflected from the slope center line in direction of the left slope part. At level one behind the avalanche breaker it is expected a peak dynamic pressure of 29 kPA. In this manner the highway along the Obertauern pass (1739 m.a.s.l.) in this selection should be secured against the fluid dynamics of the powder part (1,5 kP).
Conclusions
For the development of the model were used whose initial parameter, such
as more thickness, the more density in the starting zone, the deposition
cubatures and understandable damaging physical impacts in the avalanche
track and run-off zone, i.e. in the sphere of action of eight avalanches,
had been exactly mapped. Studies of parameters from reconded avalanche
disasters where the gradual changes in the sphere of influence were
compared by the gradual reduction of the total mass in the release area in
ten% steps showed that the mass of fracture of 50 % comperative correspond
to the observed natural effects. In this connection the estimation of the
mass in the run off zone and the determination of the snow depth are on
special importance. Other important objectives are case studies using
Swiss (Hermann et al 1993) and French models (Bouvet et al 1995). It is to
be hoped that it will be possible in the framework of an international
project (EU-project "SAME" – Environment and Climate) to
measure the velocities of big artificial released powder avalanche for
verifying the AVL-model with the help of k/a-band Doppler radar.
International Conference "Avalanches and Related Subject", Kirovsk, Russia, September 2-6, 1996
| FieSy, 23/3/98 |  |
   |
|
