Key points are not available for this paper at this time.
This work presents a regionalization of Castilla y León, an area located in the northwest portion of Spain, using 5-day rainfall. Daily data were available for 119 sites for the 6-year period 1990–1995, between the months of September and May. The region presents a special orography which affects the final results. Following similar studies (Van Regenmortel, 1995, International Journal of Meteorology 15, 313–323), Principal Component Analysis (PCA) (VARIMAX rotation) has been used together with the Cluster Analysis (Ward and average-linkage) as methods of regionalization. The causes of rainfall in Castilla y León, synoptic systems that affect the region, are obtained by using the PCA. As a result of a considerable reduction in the number of data, the properties of the different homogeneous rainfall areas are easier to define. This reduction permits a simulation of rainfall with a small number of variables having only a slight loss of information (80.5% of the variance is explained). Copyright © 2000 Royal Meteorological Society
Andr�s et al. (Sat,) studied this question.