REGIONALIZATION OF PRECIPITATION IN SWITZERLAND BY MEANS OF PRINCIPALCOMPONENT ANALYSIS

The technique of principal component analysis and of cluster analysis has been applied to two sets of precipitation data in Switzerland, one containing 47 stations (1961-80), and the other 101 stations (1981-19 93), with the aim of understanding more fully the spatial distribution of precipitation regimes. Seven regions were highlighted in the first case and 13 in the second. The high spatial coherence which appeared is quite remarkable and confirms the usefulness of these techniques fo r the analysis of the spatial distribution of meteorological variables , even in a topographically complex area such as Switzerland. The two regional distributions obtained not only correspond fairly well to the large, well-known physical regions of Switzerland, but also go much f urther, separating the Swiss Plateau into 3 clearly differentiated reg ions, for example. Regional distributions such as those discussed here can have value for climate change issues, and in particular numerical modeling of climate or climate change impacts on forests.