Skip to the main content

Review article

Phytocoenological Research in Forest Ecosystems at the Beginning of the 21st Century

Igor Dakskobler ; Biološki inštitut ZRC SAZU, Regijska raziskovalna enota Tolmin

Full text: croatian pdf 79 Kb

page 53-62

downloads: 1.140



Phytocoenology (phytosociology) studies interactions bet ween plant communities. It researches the dependence of plants on the living and non-living environment (climate, parent material, mineral soil composition). It provides explanation for the selective manner in which nature operates, which enables plant communities adapted to specific sites to form from the surviving tree, scrub and other plant species; it gives an overview of these communities and their changes over time. The article gives an account of a comprehensive historical development of phytocoenology in Central Europe and a description of certain issues in the contemporary phytocoenological study of forest ecosystems with special regard to Slovenia and Croatia.
Phytocoenology developed in the 19th century when botanists did not only study individual plants, but also how entire vegetation changes within a landscape. The focus of their attention became plant formations or plant communities in relation to their environment. In the southeastern European region, phytogeographical (geobotanical) or vegetation studies in the second part of the 19th and at the beginning of the 20th century were published by F. Krašan, G. Beck and L. Adamović, for example. An important milestone was the Botanical Congress in Brussels (1910), where the concept of association was defined. This resulted in a fast development of the discipline, but different methods were developed in different parts of the world, and attention was paid to different issues. The most widespread, also in Slovenia and Croatia, was the Central-European (Braun-Blanquet, Zürich-Montpellier) method. Among other things, the pioneers of phytocoenological research in Slovenia (G. Tomažič, M. Wraber, and V. Tregubov) and Croatia (I. Horvat, S. Horvatić) conducted also thorough research of forest communities. In this respect, Horvat´s Biljnosociološka iztraživanja šuma u Hrvatskoj (Horvat 1938) is a pioneer work. In Slovenia and Croatia, phytocoenology established itself in forestry practice only after the Second World War. Soon after the end of the war two Horvat´s books, Nauka o biljnim zajednicama (1949) and Šumske zajednice Jugoslavije (1950), were published. Professors Dušan Mlinšek and Milan Anić deserve a lot of credit for the promotion of phytocoenology in the forestry of Slovenia and Croatia because they emphasised the significance of the knowledge and consideration of sites in contemporary silviculture. The result of a very fruitful cooperation of phytocoenologists in the then Yugoslavia and more widely, within the Eastern Alpine and Dinaric Society for Vegetation Ecology in the 1970s and 1980s, was also a map of natural potential vegetation of Yugoslavia in the scale of 1:1.000.000 (B. Jovanović et al. 1986) and Prodromus phytocoenosum Jugoslaviae (Zupančič et al. 1986). The work of the time was incorporated also into the Map of Natural Vegetation of Europe in the scale of 1:2500000 (Bohn et al. 2000).
Development of fast and more advanced personal computers in the 1980 s, which paved a way for relatively simple massive utilisation of mathematical methods (above all multivariate statistics) in comparisons of phytocoenological relevés and their grouping by environmental factors, brought about a significant turnaround in vegetation research conducted according to the Central-European and other methods. One of the first widely used software of this kind was TWINSPAN (Hill 1979). Later on other program packages, such as MULVA (Wildi & Orloci 1996), SYN-TAX (Podani 2001), JUICE (Tichy 2002), CANOCO (Ter Braak & Šmilaure 2002), PC-ORD (McCune & Mefford 2006), etc. were applied as well. In this respect, a problematic issue in the Central-European method is the subjective selection of relevés and subjective evaluation of cover or abundance of species with ordinal values (e.g. r, +, 1, 2, 3, 4, 5). There have been discussions among experts on the correct procedures for numeric processing of ordinal input data. Some, e.g. Podani (2005), believe that only ordinal classification and non-metric ordination methods are suitable for such data. Others disagree. A similar problem exists with the statistical analysis of data acquired using non-random (subjective) sampling, such as are also our relevés. Experts published their pro et contra views on when and to what extent such analysis is appropriate in the journal Folia geobotanica (Herben & Chytrý 2007). Despite the above concerns it is still true that the Central-European method allows a relatively fast, simple and inexpensive way of acquiring useful data on vegetation and its connections with the environment. Databases of vegetation relevés (e.g. TURBOVEG - Hennekens & Schaminée 2001) already keep large amounts of historic, several decades and even half a century old relevés that were made with subjective plot selection. Disregarding these relevés on account of their statistically problematic (subjective and non-random) origin would mean discarding very valuable ecological data. Ecologists therefore use these data to their advantage, but with regard to their limitations. These data are used also in contemporary overviews of vegetation of large regions (e.g. in Willner & Grabherr 2007). Using and processing large quantities of relevés has changed the views of the basic unit of the syntaxonomic system - association - in many ways, and has affected the way we see the concept of character and differential species (comp. Willner 2006). When selecting diagnostic species authors apply different computing procedures. A large number of relevés enable a relatively objective calculation of fidelity of species to certain syntaxa and their diagnostic value (e.g. with phi-coefficient - Tichý & Chytrý 2006). As a rule, in formalized classification the number of syntaxonomic units of a vegetation formation (e.g. forest communities) within a certain region is reduced. The question remains, however, whether such reduction is founded on the actual site conditions and on the actual phytocoenoses.
Before the turn of the century there was a shift from the knowledge (study) of plant communities to the knowledge (study) of habitats. It is an acknowledgement of the Braun-Blanquet method that the most widely used habitat type classification in Europe (Devillers & J. Devillers-Teschuren 1996) is in many ways based on this method itself, as well as on its findings and its review of plant communities, arranged in a hierarchical system.
If we compare Braun-Blanquet´s Phytocoenology from 1964 and van der Maarel´s Vegetation ecology, which was published in 2005, we can observe a significant development and a broad array of different approaches to the research of plants, including forest vegetation. Nevertheless, the foundations of phytocoenological study of forest ecosystems in the 21st century may stay similar to what they have been so far. This means the knowledge of plants, i.e. botanical knowledge, remains essential. A forester who is professionally active in the forest should be familiar with the flora and vegetation of his district, so botany and dendrology in the new study programmes should be taught in the same extent as before, with a sufficient number of lessons left for practical and field work. Forest phytocoenology is their upgrading and its composite part is the knowledge of different methods of vegetation analysis. There are more methods apart from the Central-European method. Lately functional approach has gained momentum in Europe in discussions and research of vegetation, especially of that in disturbed habitats, and in the study of syndynamic processes (compare e.g. Grime 1974, 2001, Klotz et al. 2002). It would be very useful for the southeastern Alpine-Dinaric region with its variegated vegetation to prepare and unify the databases of our numerous relevés, to process them and critically review the correctness of names and justification of some of the syntaxa. This can only be done with consideration of the actual site conditions and the actual phytocoenoses in nature, which means we should not act merely as statisticians or mathematicians, who hardly know anything about the forest. Forest communities, associations treated as abstract units, should be not only floristically (which can be adequately provided with a mathematical processing), but also ecologically grounded, foresters (who are the users of our research) should be able to recognise their stands in the field, and our descriptions ought to provide help to foresters in concrete interventions into the forest.


Croatia, historical development, multivariate methods, phytocoenology (phytosociology), Slovenia

Hrčak ID:



Article data in other languages: croatian

Visits: 1.823 *