I n s t r u m e n t s

a n d

M e t h o d s

3. THE AUTOMATIC FORECASTINGCHAIN
   
   The acoustic system ARFANG lives in the present: it can-
   not forecast any avalanche. The best it can do is to detect
   and record continuouslyevery « acoustic event » that hap-
   pens in a given area and then transmit it to NXLOG.
   In order to perform the daily avalanche prediction, this
   system first runsits nearestcasesselectionprocedurewhich
   usesEuclidiandistancecalculationstofindthemost
   'similar' past daysregardingsnow- andweatherconditions.

ThenNXLOGreadstheavalancheactivitityobserved
duringthese 'similar' daysin its avalanche datafiles. Then
NXLOG runsits expertsystem. This procedureappliesrules
making inferencesfrom avalancheactivityobserved during
nearest cases as well as from present snow- and weather
situation and topographicalconditions.At the end of this
process,NXLOG delivers a probability of accidental ava-
lanche foreach slope of the defined area (typically a ski
resort).

Fig. 4: Special infrasound microphones ECHO, installed under the snow cover.

Fig. 5: Spectrographic analysis(1): released explosion, (2): avalanche.

131