|
avalancheactivityarenot necessarily associatedwithvisual
observations. This can be accountedfor by local effects, as
the observations are not made in the recordingsite but in a
nearby ski resort.

4.2. EVOLUTION OF THE SEISMIC AVALANCHE

ACTIVITY DURING A SNOWFALL

As the date and time of each recordedevent is known, we
can follow up the evolution of the seismic avalanche ac-
tivity provided by theSARA system during asnowfall. This
allows to estimate the moments when the avalanchecrisis
starts and stops. The global height of snow, measured at
hourly intervals by the nearby automatic Nivôse weather
station of Les Ecrins, will be used as a comparison
parameter.
The evolution of both data during the January 21-24,
1995 period is represented on figure 4. This shows that a
significant avalanche activity starts as soon as the cumu-
lated fresh snow layer reaches about 40 cm at 3:00pm on
January 22. This avalanche activity stops rapidly three
hours later as the snowfall stops too. When the snowfall
starts again at 3:00am on January 23, the avalanche activ-
ity increases immediately and becomes very high (7 sig-
nals detected as avalanches on January 23). It definitely
stops soon after the snowfall has eventually stopped at
11:00pm on January 23.
Thus, the SDA system we have set up proves able to
follow up the naturalavalancheactivityin almost real time.
|
system for Seismic Detectionof Avalanches(named SARA)
has beenworking in pre-operational conditionssince Janu-
ary 1995. The SARA system is actually able to follow the
evolution of the avalanche activity in almost real time. It
could allow to:

* Supply anobjective estimation of the current avalanche
activity to help short-term forecastof the avalanchehaz-
ard,

* Supply a reliable data for analogous prediction meth-
ods,

* Improve the calibration of deterministic models for ava-
lanche prediction, such as the SAFRAN-CROCUS-
MEPRA system developed at Météo-France,

* Perform surveillance of roads or resorts threatened by
avalanches. Theobjectivity of the information provided
by a SDA station could officially legitimate decisions
to evacuate local population or to close roads.

* Given these promising results, we have developed an
operational, self-controlling prototype. The rejectionof
non-avalanchesignalsis performed in situ by the SARA
software implemented on a PC 486DX card inside a
LEAS FRANCE seismic station. Only the results interms
of seismic avalancheactivityare transmittedto the labo-
ratory at hourly intervals. This prototype is currently
being tested on the Saint-Christophe site. Two such
prototypes will be installed for Winter 96/97 on two
different sites.
|
 |