Validation of a priori forest type classifications to predict floristic composition.

Forest type classifications based on field or satellite data collection have been used to identify conservation priorities, and thereby support decision-making, zoning and conservation planning. For these reasons, the capacity of different methods of forest classification to predict floristic composition is a crucial topic and needs to be examined. Here, three predictions are tested, considering how floristic composition is consistent with different forest type classifications: (1) forest type classifications are valid for floristic inferences (2) biogeography-based forest types perform better than stand structure-based forest types; (3) the efficiency in predicting floristic composition is depending on which factors affect floristic patterns: biogeographic or anthropogenic drivers. Species presence-absence of all vascular plants, sampled in the Italian Network Level I (CONECOFOR), are analysed to determine the best floristic classification using cluster analysis and non-metric multidimensional scaling. Analyses of similarity are used to test for differences between the National Inventory of Forests and Carbon (INFC), the European Forest Type Categories (EFTC) and the Corine Land Cover 2006, and factors affecting floristic patterns are tested.