Towards objectivity in vegetation classification: the example of the Austrian forests

We present a numerical classification of 2145 objectively sampled relevés from the entire forest area of Austria (Central Europe). The sample sites were selected by a combined method involving a systematic matrix and stratified random sampling. A TWINSPAN classification led to 32 clusters which are described in detail. Three main groups can be distinguished: (1) Alpine-dinaric coniferous forests on carbonate soils, (2) Coniferous forests on acid soils and (3) Deciduous forests. These groups correspond with accuracy to the classes Erico-Pinetea, Vaccinio-Piceetea and Querco-Fagetea in the traditional Braun-Blanquet system. Thus, the value of the Braun-Blanquet approach is supported by more or less objective sampling and numerical classification methods. The assumption of the objective existence of ecological species groups is strongly supported, too. Moreover, our results may help to solve some controverse points discussed in the European forest classification regarding the delimination between the three mentioned classes. This revised version was published online in August 2006 with corrections to the Cover Date.