In this article, we discuss the relationships between plant sociology and ecology and highlight the potential of plant sociology for ecological modelling and environmental assessment. The classification criteria for plant communities (characteristic combination of species, specific relationships with the physical environment, particular chorological and dynamical features) assign strong indicator value to phytosociological units. Moreover, the modern approach of plant sociology, which spans from individual communities to vegetation series and geosigmeta, offers the opportunity to interact under a landscape ecological perspective. Within this general context, we particularly refer to the field of ecological land classification and present a research project on the ecoregions of Italy, which is based on the National Map of Vegetation Series and on bioclimatic, biogeographical, lithological, and geomorphological data.

This paper aims to discuss the role of plant sociology in the framework of ecology and biodiversity conservation, and to present an experimental approach that promotes the integration of vegetation science with landscape ecology. A new model of deductive plant sociology at the landscape level is proposed, which is rooted in the traditional approach but relates to the landscape ecology of the ‘European school’ for its emphasis on landscape classification. Using the case study of the Municipality of Rome, a methodological process for defining and mapping vegetation series, geosigmeta and landscape units is proposed, based on the overlay of climatic, lithomorphological, floristic and vegetation information, the latter in its syntaxonomical and dynamical aspects. At the landscape level plant sociology can significantly contribute to formalize ecological models based on complex isofunctional units and helpful in environmental management and planning.

This paper focuses on the statistical properties of wild-land fires and, in particular, investigates if spread dynamics relates to simple invasion model. The fractal dimension and lacunarity of three fire scars classified from satellite imagery are analysed. Results indicate that the burned clusters behave similarly to percolation clusters on boundaries and look more dense in their core. We show that Dynamical Percolation reproduces this behaviour and can help to describe the fire evolution. By mapping fire dynamics onto the percolation models the strategies for fire control might be improved.