Gradient analysis of dry grassland vegetation in Denmark

Abstract. We present a gradient analysis of 620 vegetation samples covering most of the floristic and environmental variation in semi‐natural grassland vegetation on well‐drained soils in Denmark. Vegetation was sampled using frequency in subplots. Explanatory variables were surface inclination, aspect, pH, geographical co‐ordinates together with indications of soil type. Detrended Correspondence Analysis revealed four floristic gradients that could be interpreted in ecological terms by measured variables supplemented with site calibrations based on weighted averaging of Ellenberg's indicator values. All four axes were interpreted using rank correlation statistics, and linear and non‐linear multiple regression of sample scores on explanatory variables. The first gradient was from dry calcareous to humid acidic grasslands; the second reflected an underlying gradient in fertility; the third reflected regional differentiation and the fourth was associated with variation in intensity of competition as indicated by association with calibrated Grime‐CSR values for the plots. We applied subset ordination to the data as a supplement to traditional permutation and correlation statistics to assess the consistency of ordination results. DCA axes 1 and 2 were consistent in space and time. This gradient analysis is discussed in a context of plant strategy theory and species diversity models. Ecocline patterns lend support to the view that grazing not only favours the ruderal strategy but also the stress‐tolerant strategy. The low rank of competition as an explanatory variable for the floristical gradients supports the notion that competitive effects play a subordinate role for species composition compared to microclimate and soil conditions in infertile semi‐natural grasslands.