WEATHER CATEGORIZATION BASED ON THE AVERAGE LINKAGE CLUSTERING TECHNIQUE - AN APPLICATION TO EUROPEAN MIDLATITUDES

The computer-assisted classification of weather at Prague-Clementinum used the average linkage clustering technique. Since the results of th e clustering exhibit the snowballing effect, the usual methods of dete rmining the threshold aggregation level (i.e. the level at which the c lustering procedure is to be terminated) appeared to be inapplicable. A new method based on Monte Carlo simulations of the means was develop ed. Its key idea is the termination of the clustering procedure at dif ferent aggregation levels in different parts of the data set. This ens ures that the number of resultant clusters is reasonable, while minimi zing the numbers of very small clusters and unclustered days. The weat her categorization resulted in 44 clusters for 14 winters in the perio d 1965-1978. Thirty-one of the clusters had sizes of 5 or more days. T he Monte Carlo scores, comparing the means and variances of the cluste rs with those of a large number of subsets chosen randomly, indicate t hat all the resulting clusters represent meaningful weather types. Thi s study may provide a basis for nucleated clustering, which will enabl e us to deal with longer data series and to study the long-term trends of the properties of weather types.