Coarse-scale substrate mapping using plant functional response types

Abstract

Due to the fine-scale variation of substrates, spatial information about soils must be generalized for mapping. The process of generalization is conventionally based on landscape controls of soil mosaics (like terrain and climate), and not on the real fine-scale patterns. In order to base generalization on real patterns, the present study aims to implement an indirect method for the spatial aggregation of substrate types. It relies on the effects of substrate occurrences on frequency spectra of plant functional response types. These spectra have been used to derive types of soil-landscapes. The area of investigation is Germany. Plants have been affiliated to response types using indicator values for nutrient supply, water supply and soil pH. These values rank species according to their realised optima on environmental gradients. The analysis was based on proportions of indicator groups in grid-cells of the German floristic survey. It was not based on absolute frequencies like in earlier studies. Features of the frequency spectra that could not be related to generalizable trends in the data have been omitted by a Principal Component Analysis (PCA). The PCA-scores have been used for a classification of soil-landscapes that could be enhanced afterwards by salinity indicators. The geographical breakdown resembles a spatially alternating dominance of various controls (rocks, climate, terrain) on substrate patterns. Unlike other coarse-scale geographical landscape classifications, both the varying weights of differentiating controls and coarse-scale spatial patterns are inductive outcomes of the analyses. These outcomes have the potential to help in the selection of controls used for predictive models of soil-landscapes.