[Current editors: Ladislav Mucina, Miquel De Cáceres]
There is a long history and tradition to the methods of classification of vegetation, which has its roots in classical plant geography of the 19th century (Pott, 2011). We divided here the history of vegetation classification into four periods.
According to Pott (2011), the period characterised by the physiognomic approach to vegetation started at the beginning of the 19th century with Alexander von Humboldt, the great master of plant geography. Von Humboldt focused his studies on plant distribution, formation of communities and plant growth forms. Purely physiognomic terms were created for the large-scale distribution of vegetation units or formations. Following von Humboldt, it is important to mention Joachim Frederik Schouw who made a point of focusing on the phenomena of plant distribution and community formation. August Grisebach took a similar approach in his important work ‘‘Vegetation of the Earth’’. These scientific pioneers, and others, were plant geographers who in their work paid particular attention to the phenomenon of the formation of plant communities in their work. For the most part, they applied the term formation as the basis for their study of vegetation. Grisebach (1838) defined formation as: "a community that can be identified through the 'growth form' of its dominant species".
The term formation is still used today when refering to a broad division of large areas of vegetation on the Earth. However, it does not suffice to describe a more detailed study of vegetation on a smaller scale. This realisation was already evident at the turn of the 19th century as the transition was made to attempting a more refined differentiation between vegetation units and classifying them as objectively as possible. Plant species became more important for classification purposes. Answering the question, which types can be found and in what quantity can they be found leads to determination of quality and quantity of individual species in the vegetation unit. This led to the development of plant sociology basis on the floristic approach.
According to Biondi (2011), the date of birth of phytosociology is set at the Third International Congress of Botany held in Brussels from 14th to 22nd May 1910, when, after discussion between plant sociologists and physiognomists, the following definition of association proposed by Flahault and Schöter (1910) was unanimously adopted: ‘‘An association is a vegetal grouping of determined floristic composition, presenting a uniform physiognomy that grows in uniform site conditions. The association is the fundamental unit of synecology’’. From this starting, several schools of phytosociology evolved in Europe during the 20th century, although the schools did not take concrete shape until the first half of the century. According to Pott (2011), their emergence can largely be attributed to the different regional conditions existing in the individual countries.
The Zürich-Montpellier school was started by Prof. Braun-Blanquet, and is considered to formally start when he published his book Pflanzensoziologie in 1928. In Braun-Blanquet's definition of association, the concept of characteristic species first appeared: ‘‘The association defines a vegetal grouping more or less stable and in equilibrium with the environment, characterized by a particular floristic composition, in which some exclusive or almost exclusive elements (characteristic species) reveal with their presence a particular and autonomous ecology’’. Braun-Blanquet's ideas were developed further by Prof. Rheinhold Tüxen in Germany, and both scientists carried on the development of the system well into the 1960's. Their influence was so important that methods of the Zürich-Montpellier school were the dominant classification standard in West, South and Central Europe for most of the XXth century. The Zürich-Montpellier school combined a standardized protocol for plot sampling, sorting of species-by-plot matrices, demarcation of community types, and their hierarchical ordering into a practical and efficient framework for the study of vegetation. However, prior to Ellenberg (1956) phytosociologists had no comprehensive account of the table-sorting method available, and indispensable ‘guidance by a master’ were listed as the major conditions of syntaxonomic practise (Mucina 1997). Other descriptions of the method can be found in several sources, including Braun-Blanquet (1964), Poore (1955), Mueller-Dombois & Ellenberg (1974), van der Maarel (1975), Dengler et al. (2008) and Kent (2012). Despite the number of sources, a number of aspects of the method remain unclear, even to experienced workers (Ewald, 2003).
The basic principles of vegetation science and analysis of plant communities in Nordic countries were formulated by von Post (1851). Much of the theoretical framework of the phytosociological approach, with papers on how to describe vegetation units and to measure them, was carried out by plant ecologists in Copenhagen (e.g. Raunkiær 1928; Sørensen 1948) and Uppsala (e.g., Sernander 1894; DuRietz 1921) which gave rise to ‘Scandinavian Schools’ in phytosociology. Characteristic features of these were comparatively small sample plots of fixed size, the demand for plot homogeneity and completeness of the species list, and the importance of dominant species and life forms for the classification of communities (Lawesson et al. 1997). The Scandinavian schools have never been as uniform as for example the Braun-Blanquet approach in Central and South Europe (Lawesson et al., 1997). Oksanen (1990) reviewed the history and recent situation of vegetation science in Finland in detail.
Synusiae are elementary one-layered floristically, physiognomically and ecologically homogeneous vegetation units, directly linked to uniform enviornmental conditions (Cain 1936; Lippmaa 1939; Barkman 1978). Synusiae are used in the sense of a concrete community and not as an abstract classification unit (Gillet & Gallandat, 1996).
According to Pott (2011), classification of vegetation underwent a distinct type of development in Russia and neighbouring countries. Broad expanses of country allowed a simplified form of vegetation description based on the dominant forms of growth. The approach favouring a more detailed floriostic description of the vegetation did not become popular until the recent years when it was used as a reliable method for recording biodiversity.
The transition to modern times in vegetation classification started in the 60', when computers were introduced into the processing of phytosociological data. Mucina (1997) calls the period between 1969 and 1974 the ‘Innovation period’, as it brought a series of changes in editorial policies as well as novelties in international cooperation in vegetation science. First, the period was marked by the foundation of the ‘Working Group for Data-Processing in Phytosociology’ originally aimed at providing a methodological background for a Prodromus of European vegetation units. From volume 29 onwards, the face of Vegetatio changed considerably when Prof. Eddy van der Maarel took over the journal’s editorial office. This step brought a decisive shift in the editorial policy of Vegetatio with an emphasis on theory and methods. These included methods such as cluster analysis, automatic table-sorting and ordination. The increase in numbers of papers using numerical methods in the period 1960-1986 was shown by Kent & Ballard (1988). The begin of numerical syntaxonomy - an integrative approach aimed at the derivation and characterization of vegetation types using numerical techniques - is set in 1964, according to Mucina & van der Maarel (1989). In the following years, many vegetation scientists, both within the traditional European approaches and on other continents, adopted numerical techniques and a wealth of papers, monographs and even handbooks were published. Another clear tendency was the theoretical development of classification and ordination as numerical techniques and the comparison and further development of resemblance measures and multivariate techniques. Two elaborate programs became available in the late 1970s. TABORD was a program for agglomerative clustering with relocation of relevés and clusters and subsequent one-dimensional arrangement of clusters in a structured table (van der Maarel 1978). TWlNSPAN, two-way indicator species anlysis, was a program for divisive clustering of relevé sets on the basis of indicator species differentiating relevé groups towards each other, followed by an arrangement of both relevé groups and species groups in a structured table (Hill 1979).
According to Chytry et al. (2011), since the 1990s, the field of vegetation survey and classification received new stimuli. First, there has been an increased demand from institutions in the field of nature conservation for comprehensive systems of vegetation/habitat/biotope classifications to serve as a tool for informed conservation planning and decision making. Second, there have been technological advances in the setting up and management of electronic databases.Third, availability of remote sensing data and geographical information systems (GIS) offer entirely new opportunities for survey and vegetation mapping of large or remote areas, so long as they are combined with detailed ground-based vegetation survey (Chytry et al. 2011).
Vegetation classification based on floristics (i.e. phytosociology) is still the mainstream vegetation classification scheme in Europe, as well as in several countries outside Europe, and has become increasingly popular worldwide from the 1990s onward (Dengler et al. 2008).
Mucina (1997) describes five trends in vogue in vegetation classification, which we briefly summarize here:
In paral·lel with methodological developments, there have been recent attempts to redefine the association concept and discuss the basis of floristically based vegetation classification. Particularly relevant for classification protocols is the role of diagnostic species in the defintion of vegetation units. Dengler et al. (2008) stress the importance of the total species composition and introduce diagnostic species later: "...Within the Braun-Blanquet approach, the concept of character and differential species is important for the recognition of previously defined [our emphasis] syntaxa". In this line, Willner (2006) suggested going back to the definition of association made by Flahaut & Schöter (1910), thus previous to the advent of diagnostic species, and complementing it with: "The floristic composition of a vegetation unit can only be considered to be 'definite', if each of its stands can be determined alone by its good diagnostic species". On the other hand, Biondi (2011) is less clear on the role of diagnostic species: "The association defines a system of vegetal organisms with a floristic composition that is statistically repetitive; it has a range of different features such as the structure, the ecological value (significant for different environmental parameters) and the quality of the dynamic and catenal relationships that it has with other communities. Especially pertinent for its definition and identification [again, our emphasis] is the characteristic specific complex, consisting of the preferring plants which are particularly linked to it in statistical terms and that are biogeographically and ecologically differential compared to similar synvicariant or geosynvicariant associations".