Impact of small-scale variability of soil pH on the vegetation composition in a mixed hardwood stand, East-Flanders Belgium

Abstract

1 Soil pH of the forest stand varies widely for the different depths and is very important in determining the forest vegetation types. Main aim of the study was to examine the correspondence between forest vegetation composition and small-scale heterogeneity in pH by several statistical methods. 2 Ordination using DCA (Detrended Correspondence Analysis) found clear distinction in the vegetation structure and pointed pH differences as the main factor explaining the variability (Eigenvalue= 0.76 for the first three axis). 3 Clustering analysis by TWINSPAN resulted into four major vegetation groups. Most of these divisions were explained by differences in pH. 4 The spatial dependence of the pH at ectorganic and top mineral layers was examined by geostatistics analysis. All variograms showed a similar well-defined structure. The structure of variograms accounted for 95 - 100 % of the spatial dependence for the pH at all depths. All variograms represented the same range of 50 m for the pH at ectorganic horizons, but the distance of 65 and 72 m was calculated respectively for the pH at 0 - 5 and 5 -15 cm depth. All variograms increased to the maximum lag over which they were calculated. 5 The raster image of pH variability resulting from Kriging interpolation coincided significantly with the raster image of vegetation composition. It was concluded that it is possible to predict the pH value based on the vegetation composition or vice versa. Predicting the vegetation type from pH value is only possible for the range of the distance of the pH variogram.