Spatial patterns of daily rainfall in central Kenya: Application of principal component analysis, common factor analysis and spatial correlation

Abstract

Spatial patterns of daily rainfall in central Kenya were investigated using principal component analysis (PCA) and common factor analysis (CFA) of covariance matrices, together with spatial correlation analysis. Data consisted of daily values from 42 stations for the period January 1971 to November 1979. All statistical analyses were made both on the whole data set, and by month.The number of components/factors to retain for rotation could not be unambiguously determined by means of scree tests. Therefore, various numbers of components/factors were rotated, mainly by using the Direct Oblimin procedure. It was not possible to obtain simple structure solutions from the PCA, which explained 50-70 per cent of the total variance. CFA explained 30-40 per cent of the total variance and was easily rotated to simple structure. It is concluded that this difference is due to the large portion of unique variance in the data. Because of this, maximum likelihood CFA is judged superior. Also the scale-independence of the method is considered advantageous.The PCA and CFA correctly located the limit between the region of two pronounced rainy seasons, and the region of more complex and less distinct rainfall seasonality.Spatial correlation analyses suggest two seasonal rainfall regimes, one of 'random' convection during December-March, and another of rainfall organized in approximately NE-SE direction. The latter is mainly attributed to orographic influence, but possible links to fluctuations in the general SE-monsoonal flow and the East African low-level jet are briefly discussed.KEY WORDS Monsoon rainfall Rainfall regime

Daily rainfall affinity areas in Mediterranean Spain

“…The PCA is designed not merely as a data reduction technique, but as a method which ensures that only the fundamental modes of variation of the data are considered. Several criteria have been suggested for deciding how many PCs to retain in order to separate 'signal' from 'noise' (Jolliffe, 1986;Bärring, 1987;Preisendorfer, 1988), but a clear-cut number of PCs is certainly rare. In this study, the simple scree test of Cattell (1966a) was adopted.…”

Synoptic patterns associated with wildfires caused by lightning in Castile and Leon, Spain

“…The PCA was applied in T-mode with a correlation matrix, using the time steps as variables and the grid points as observations. Several authors (Bärring, 1987;Preisendorfer, 1988;Jolliffe, 2002) have used different criteria to decide the number of PCs to retain so as to discriminate between signal and noise. In this case, we have considered only the most important components that account for at least 90% of the total variance.…”

Daily rainfall patterns in the Spanish Mediterranean area: an objective classification

“…Several criteria have been suggested for deciding how many PCs to retain in order to separate 'signal' from 'noise' (Jolliffe, 1986;Barring, 1987;Preisendorfer, 1988), but a clear-cut number of PCs is certainly rare. The simple scree test of Cattell (1966) was adopted.…”