The differentiation of thaws in connection with slushflow occurrences.

Abstract

Classification of meteorological conditions accompanying slushflow occurrences is attempted. Days with mean daily air temperature above O°C have been chosen for the analysis. Each day has been described with a set of meteorological parameters. Multivariate statistical methods have been used to derive rules for discrimination of "slushflow" and "non-slushflow'' situations based on these parameters. The linear discriminant analysis and Bayes' rule have been applied to data for the Khibini Mountains in Russia and mountainous regions of Norway. The results obtained show that snow depth and total water supply have a statistically sufficient difference between the two situations to enable them to be used for relatively reliable recognition of such situations. Classification accuracy is better than 75% for both types of situation. Transformation of the classification method into a forecast method is discussed.