Formalized reproduction of an expert-based phytosociological classification

608

Kočí, M. et al.

Fig. 1. Scheme of the Cocktail classification of relevés into

three vegetation units, followed by a similarity-based assign-

ment of unclassified relevés to these units. Circles denote the

boundaries of the three vegetation units X, Y and Z. Small

letters x, y, z and o are symbols for relevés distributed in a

vegetation continuum, different letters denoting their assign-

ment to the vegetation units X, Y, Z and non-assignment,

respectively. Underlined letters indicate the relevés assigned

to the vegetation units by Cocktail definitions, non-underlined

letters are the relevés assigned by calculating their similarity

to the vegetation units. A. Hierarchical approach of the earlier

studies, where vegetation unit X is defined first and has the

highest priority, while Z has the lowest priority; B.  Non-

hierarchical approach used in this paper, with relevés that

occurred in the overlapping parts of the Cocktail definitions of

vegetation units being assigned using a similarity criterion.

priority and each of the next units included the defini-

tions of all the previously defined units in the negative

part of its definition (Fig. 1A). Such an approach, how-

ever, is unduly influenced by vegetation units from

which the classification starts. The new procedure pro-

posed in this paper solves the overlap issue by a similar-

ity-based assignment of the equivocally classified relevés

to one of the possible vegetation units (Fig. 1B). The

advantage of this procedure is that definitions of vegeta-

tion units are independent of the earlier defined units.

An important feature of the Cocktail method is that

some relevés which do not meet any of the association

definitions are not classified. Most of the unclassified

relevés are from vegetation stands that lack specialist

species, which could be used as diagnostic species of

associations. This feature is similar to the traditional

Braun-Blanquet approach, in which the researchers pref-

erably sample vegetation stands with specialized spe-

cies and often neglect stands composed mainly of

generalists. This makes the Cocktail method ideally

suited for classification of phytosociological data which

are biased towards vegetation stands with specialist

species. In traditional phytosociology, the stands with-

out specialist species are usually referred to as impover-

ished, atypical, initial, degraded, transitional, trunk or

basal communities. Kopecký & Hejný (1978) proposed

to assign these stands directly to higher syntaxa, such as

alliances, orders or classes. Although their approach has

been followed only in studies of synanthropic vegeta-

tion (Kopecký et al. 1995), it is applicable to any vegeta-

tion type. Using the Cocktail method, definitions of

higher syntaxa can be formed in a similar way as defini-

tions of associations, and relevés unassigned to the

associations can be assigned directly to the higher

syntaxa.

In vegetation mapping and other applications of

vegetation classification, however, it might be a disad-

vantage if some patches of a single stand were as-

signed to an association and others directly to higher

syntaxa. We propose a solution by calculating similar-

ity between associations and the relevés not belonging

to the associations according to Cocktail definitions.

Then, users of the classification will distinguish three

categories of relevés or vegetation stands (Fig. 1), in-

cluding those (1) belonging to the association, i.e. corre-

sponding to its Cocktail definition; (2) not belonging to

the association but related to it, i.e. not corresponding to

the Cocktail definition but being similar to the relevés

corresponding to this definition; and (3) not belonging

and not related to the association. If necessary, the

second category can be merged with the first for the

purposes of vegetation mapping. The distinction be-

tween the second and third category can be made in two

ways. The first option is defining a similarity threshold

that provides a criterion for relevés to be considered as

either related or unrelated to the association. The second

option, used in this paper, is calculation of similarities

between the relevé and each of predefined set of asso-

ciations, and subsequent assignment of the relevé to the

most similar association.

---

- Formalized reproduction of an expert-based phytosociological classification -

609

applications], pp. 151-162, Opulus Press, Uppsala, SE. (In

Dutch.)

Hill, M.O. 1979. TWINSPAN. A FORTRAN program for ar-

ranging multivariate data in an ordered two-way table by

classification of the individuals and attributes. Ecology &

Systematics, Cornell University, Ithaca, NY, US.

Hill, M.O. 1989. Computerized matching of relevés and asso-

ciation tables, with an application to the British National

Vegetation Classification. Vegetatio 83: 187-194.

Jandt, U. 1999. Kalkmagerrasen am Südharzrand und im

Kyffhäuser. Gliederung im überregionalen Kontext,

Verbreitung, Standortsverhältnisse und Flora. Diss. Bot.

322: 1-246.

Jandt, U. 2000. Application of the species-group method to the

data base of calcareous grasslands in Germany. In: White,

P.S., Mucina, L. & Lepš, J. (eds.) Vegetation science in

retrospect and perspective, pp. 115-119, Opulus Press,

Uppsala, SE.

Kočí, M. 2001. Subalpine tall-forb vegetation (Mulgedio-

Aconitetea) in the Czech Republic: syntaxonomical revi-

sion. Preslia 73: 289-331.

Kopecký, K. & Hejný, S. 1978. Die Anwendung einer

deduktiven Methode syntaxonomischer Klassifikation bei

der Bearbeitung der strassenbegleitenden Pflanzen-

gesellschaften Nordostböhmens. Vegetatio 36: 43-51.

Kopecký, K., Dostálek, J. & Frantík, T. 1995. The use of the

deductive method of syntaxonomic classification in the

system of vegetational units of the Braun-Blanquet ap-

proach. Vegetatio 17: 95-112.

Krestov, P.V. & Nakamura, Y. 2002. Phytosociological study

of the Picea jezoensis forests of the Far East. Folia Geobot.

37: 441-473.

Kubát, K., Hrouda, L., Chrtek jun., J., Kaplan, Z., Kirschner, J.

& Stěpánek, J. (eds.) 2002. Klíč ke květeně České republiky

[Key to the flora of the Czech Republic]. Academia, Praha,

CZ. (In Czech.)

Mucina, L. & van der Maarel, E. 1989. Twenty years of

numerical syntaxonomy. Vegetatio 81: 1-15.

Mucina, L., Grabherr, G., Ellmauer, T. & Wallnöfer, S. 1993.

Die Pflanzengesellschaften Österreichs. Teil I-III. G.

Fischer, Jena, DE.

Noble, I.R. 1987. The role of expert systems in vegetation

science. Vegetatio 69: 115-121.

Oberdorfer, E. (ed.) 1977-1992. Süddeutsche Pflanzen-

gesellschaften. Fischer, Jena, DE.

Ostermann, O.P. 1998. The need for management of nature

conservation sites designated under Natura 2000. J. Appl.

Ecol. 35: 968-973.

Pflume, S. 1999. Laubwaldgesellschaften im Harz. Gliederung,

Ökologie, Verbreitung. M. Galunder-Verlag, Wiehl, DE.

Pignatti, S., Oberdorfer, E., Schaminée, J.H.J. & Westhoff, V.

1995. On the concept of vegetation class in phytosociology.

J. Veg. Sci. 6: 143-152.

Schaminée, J.H.J. 1995–1999. De vegetatie van Nederland

[The vegetation of the Netherlands]. Opulus Press, Uppsala,

SE. (In Dutch.)

Sokal, R.R. & Rohlf, F.J. 1995. Biometry. 3rd ed. Freeman,

New York, NY, US.

Täuber, T. 2000. Zwergbinsen-Gesellschaften (Isoëto-Nano-

Acknowledgements. We thank Helge Bruelheide, Jason Holt,

Ute Jandt, Onno van Tongeren and two anonymous referees

for their stimulating ideas and comments on a previous version

of the manuscript. Maruška Rafajová managed our vegetation

database. This research was supported by the grants GAČR

206/99/1523 and MSM 143100010.

References

Bergmeier, E. 2002. Plant communities and habitat differen-

tiation in the Mediterranean coniferous woodlands of Mt.

Parnon (Greece). Folia Geobot. 37: 309-331.

Bruelheide, H. 1995. Die Grünlandgesellschaften des Harzes

und ihre Standortsbedingungen. Mit einem Beitrag zum

Gliederungsprinzip auf der Basis von statistisch ermittelten

Artengruppen. Diss. Bot. 244: 1-338.

Bruelheide, H. 1997. Using formal logic to classify vegeta-

tion. Folia Geobot. Phytotax. 32: 41-46.

Bruelheide, H. 2000. A new measure of fidelity and its appli-

cation to defining species groups. J. Veg. Sci. 11: 167-178.

Bruelheide, H. & Chytrý, M. 2000. Towards unification of the

national vegetation classifications: A comparison of two

methods for the analysis of large data sets. J. Veg. Sci. 11:

295-306.

Chytrý, M. & Rafajová, M. 2003. Czech National Phyto-

Sociological Database: basic statistics of the available

vegetation-plot data. Preslia 75: 1-15.

Chytrý, M., Tichý, L., Holt, J. & Botta-Dukát, Z. 2002. Deter-

mination of diagnostic species with statistical fidelity

measures. J. Veg. Sci. 13: 79-90.

Dierschke, H. (ed.) 1996. Synopsis der Pflanzengesellschaften

Deutschlands.  Floristisch-soziologische Arbeitsgemein-

schaft & Reinhold-Tüxen Gesellschaft, Göttingen, DE.

Dodd, M.E., Silvertown, J., McConvay, K., Potts, J. & Crawley,

M. 1994. Application of the British National Vegetation

Classification to the communities of the Park Grass Ex-

periment through time. Folia Geobot. Phytotax. 29: 321-

334.

Doing, H. 1969. Sociological species groups. Acta Bot. Neerl.

18: 398-400.

Ejrnaes, R., Aude, E., Nygaard, B. & Münier, B. 2002. Predic-

tion of habitat quality using ordination and neural net-

works. Ecol. Appl. 12: 1180-1187.

Ewald, J. 2001. Der Beitrag pflanzensoziologischer Daten-

banken zur vegetationsökologischen Forschung. Ber.

Reinhold-Tüxen-Ges. 13: 53-69.

Feoli, E. 1984. Some aspects of classification and ordination

of vegetation data in perspective. Stud. Geobot. 4: 7-21.

Hennekens, S.M. & Schaminée, J.H.J. 2001. TURBOVEG, a

comprehensive data base management system for vegeta-

tion data. J. Veg. Sci. 12: 589-591.

Hennekens, S.M., van der Maarel, E. & Stortelder, A.H.F.

1995. Numerieke methoden [Numerical methods]. In:

Schaminée, J.H.J., Stortelder, A.H.F. & Westhoff, V. (eds.)

De vegetatie van Nederland. Deel 1. Inleiding tot de

plantensociologie – grondslagen, methoden en toepas-

singen [The vegetation of the Netherlands. Vol. 1. Intro-

duction to phytosociology - fundamentals, methods and

‹

---

610

Kočí, M. et al.

juncetea) in Niedersachsen – Verbreitung, Gliederung,

Dynamik, Keimungsbedingungen der Arten und Schutz-

konzepte. Cuvillier, Göttingen, DE.

Tichý, L. 2002. JUICE, software for vegetation classification.

J. Veg. Sci. 13: 451-453.

Valachovič, M., Otahelová, H., Stanová, V. & Maglocký, S.

1995.  Rastlinné spoločenstvá Slovenska 1. Pionierska

vegetácia [Plant communities of Slovakia 1. Pioneer veg-

etation]. Veda, Bratislava, SK. (In Slovak.)

Westhoff, V. 1967. Problems and use of structure in the

classification of vegetation. The diagnostic evaluation of

structure in the Braun-Blanquet system. Acta Bot. Neerl.

15: 495-511.

Westhoff, V. & van der Maarel, E. 1978. The Braun-Blanquet

approach. In: Whittaker, R.H. (ed.) Classification of plant

communities, pp. 289-399. W. Junk, The Hague, NL.

Wildi, O. 1989. A new numerical solution to traditional phyto-

sociological tabular classification. Vegetatio 81: 95-106.

Willner, W. 2002. Syntaxonomische Revision der südmittel-

europäischen Buchenwälder. Phytocoenologia  32: 337-

453.

Received  15 February 2002;

Revision received  2 February 2003;

Accepted 13 March 2003.

Coordinating Editor: H. Bruelheide.

‹

App. 1. Brief description of the associations recognized in the expert-based classification (Kočí 2001).

1.  Sphagno compacti-Molinietum caeruleae  Wagnerová ex Berciková 1976 – species-poor chionophilous and hygrophilous

community dominated by Molinia caerulea, which occupies mostly leeward edges of cirques

2–3. Crepido-Calamagrostietum villosae (Zlatník 1925) Jeník 1961 and Sileno vulgaris-Calamagrostietum villosae Jeník et al. 1980

– closed grasslands dominated by Calamagrostis villosa, confined to leeward sites with dry, deep and nutrient-rich brown

alpine soils; the former association includes species-rich, the latter species-poor stands

4. Bistorto-Deschampsietum alpicolae (Zlatník 1928) Burešová 1976 – species-poor community of Deschampsia cespitosa and

Polygonum bistorta occupying shallow, wet depressions with long-lasting snow cover on the ridges and moderate slopes

above the timberline

5. Violo sudeticae-Deschampsietum cespitosae (Jeník et al. 1980) Kočí 2001 – Deschampsia cespitosa and Poa chaixii dominated

community, occupying depressions near springs and shaded places around the timberline, on soils with permanently

percolating ground water

6. Bupleuro-Calamagrostietum arundinaceae (Zlatník 1928) Jeník 1961 – species-rich community dominated by Calamagrostis

arundinacea, confined to steep, dry and sunny places, mostly located at the bases of slopes of cirques and on avalanche paths

7. Salici silesiacae-Betuletum carpaticae Rejmánek et al. 1971 – subalpine open shrubberies of Betula carpatica and Salix silesiaca

with tall forbs, occurring on steep slopes with sliding snow and infrequent avalanches

8. Pado-Sorbetum W. Matuszkiewicz et A. Matuszkiewicz 1975 – subalpine open shrubberies of Sorbus aucuparia subsp. glabrata

with herb layer dominated by tall forbs, confined to the bottoms of cirques and to moist ravines on shallow, stony and acidic

soils

9. Piceo-Salicetum silesiacae Rejmánek et al. 1971 – open shrubberies of Salix silesiaca and Picea abies with ferns and woodland

species in herb layer, influenced by snow accumulation and spring floods, occurring along submontane rivers

10. Ranunculo platanifolii-Adenostyletum alliariae (Krajina 1933) Dúbravcová et Hadač ex Kočí 2001 – species-rich community

dominanted by Adenostyles alliariae, confined to moderate slopes, in the surroundings of springs and streams, in shaded

places and wet depressions around the timberline

11. Salicetum lapponum Zlatník 1928 – subalpine low-willow shrubberies occupying shallow wet depressions, surroundings of

springs and mires, and the upper edges of cirques with permanently moist, shallow, often peaty and acidic soils

12. Trollio altissimi-Geranietum sylvatici Jeník et al. 1980 – species-rich tall-forb community confined to the surrounding of streams

and springs with moist soils, occurring at their upper edges of cirques outside the avalanche tracks

13. Laserpitio-Dactylidetum glomeratae Jeník et al. 1980 – species-rich tall-forb community of the bottoms of cirques, confined to

deep and moist soils, rich in nutrients supplied by avalanches, aeolic sedimentation, and percolating ground water

14.  Chaerophyllo-Cicerbitetum alpinae  (Kästner 1938) Sýkora et Hadač  1984 – community dominated by Petasites albus and

Cicerbita alpina, confined to shaded and wet surroundings of streams and springs and the bottoms of V-shaped valleys and

canyons in the montane and supramontane belts

15. Daphno mezerei-Dryopteridetum filicis-maris Sýkora et Stursa 1973 – species-rich community dominated by Dryopteris filix-

mas, occupying mostly dry and warm screes and scree cones with shallow soil at the bases of steep slopes in cirques, covered

by thick snow accumulations in winter

16. Adenostylo-Athyrietum alpestris (Zlatník 1928) Jeník 1961 – species-poor, chionophilous community dominated by Athyrium

distentifolium, occurring on deep soils on shaded wet places around the timberline, in places with a thick snow cover in winter

‹