Avalanche warning onChowkibal-Tangdhar axis (J&K):A hybrid approach

摘要

Conventional as well as numerical techniques are beingwidely used for the prediction of snow avalanches.The present approach combines both the techniquesand delivers avalanche danger warning for 24 h in advance. Initially different levels of avalanche danger have been decided by observing fresh snow of 24 hand standing snow from a snow-meteorological database of the past ten winters (1992-2002) along with adatabase of avalanche warning and occurrences. Finallythese levels have been characterized by a criticalrange calculated by using a discriminant function,which is a function of all the significant snow and meteorological parameters. The significant snow and meteorological parameters have been selected by correlationanalysis. For the selection of significant parametersand to calculate the critical range for each of the dangerlevels, a new term, i.e. index of avalanche has been introduced and its variation in different ranges of snow and meteorological parameters has been discussed. Forthe winter of 2003-2004, model outcome has beencompared with the actual avalanche occurrences. Out oftotal 122 days during winter, there were 27 avalanchedays and 95 non-avalanche days. The accuracy of themodel for avalanche occurrence is 67% and for non-occurrence it is 84%.