Vegetace CeskÈ republiky» 1 Travinn· a ke¯ÌËkov· vegetace · ACADEMIA PRAHA 2007 Vegetation of the Czech Republic Travinn· a ke¯ÌËkov· vegetace 1. Grassland and Heathland

ACADEMIAPRAHA 2007

Vegetation of the Czech Republic

Travinn· a ke¯ÌËkov· vegetace

1. Grassland and Heathland Vegetation

Milan Chytr˝ (editor) a kolektiv

1»Vegetace CeskÈ republiky

KATALOGIZACE V KNIZE – NÁRODNÍ KNIHOVNA ČR

Vegetace České republiky. 1, Travinná a keříčkovávegetace = Vegetation of the Czech Republic. 1,Grassland and heathland vegetation / Milan Chytrý (editor)a kolektiv. – Vyd. 1. – Praha : Academia, 2007. – 528 s. ;barev. il.ISBN 978-80-200-1462-7

581.524/.526 * 581.526.45 * 581.526.4:582.093/.095 *(437.3)- vegetace – Česko- luční rostlinná společenstva – Česko- stepní rostlinná společenstva – Česko- křovinná společenstva – Česko- monografie

581 – Obecná botanika [2]

Kolektiv autorů:

EditorMilan Chytrý1

Autoři textů a dílčích analýz datMilan Chytrý1, Martin Kočí1, Kateřina Šumberová2, Jiří Sádlo3, František Krahulec3,Petra Hájková1, 2, Michal Hájek1, 2, Aleš Hoffmann3, Denisa Blažková3, Tomáš Kučera4,Jan Novák5, Marcela Řezníčková1, Tomáš Černý3, Handrij Härtel3 & Deana Simonová1

Editace a analýza dat, software a technická spolupráceLubomír Tichý1, Ilona Knollová1, Zdenka Otýpková1, Jiří Danihelka1, 2,Ondřej Hájek1, Klára Kubošová1, Katrin Karimová1 & Jiří Rozehnal1

1 Přírodovědecká fakulta Masarykovy univerzity, Brno2 Botanický ústav AV ČR, pracoviště Brno3 Botanický ústav AV ČR, pracoviště Průhonice4 Ústav systémové biologie a ekologie AV ČR, pracoviště České Budějovice5 Biologická fakulta Jihočeské univerzity, České Budějovice

Editor © Milan Chytrý, 2007

ISBN 978-80-200-1462-7

7

ObsahContents

Úvod Introduction (M. Chytrý) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Vymezení vegetačních jednotek a jejich interpretaceDelimination and interpretation of vegetation units (M. Chytrý) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Project Vegetation of the Czech Republic: Preface and summaryof methods (M. Chytrý) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

Diverzita vegetace České republiky, její příčiny a historický vývojDiversity of vegetation of the Czech Republic, its determinants and history (J. Sádlo) . . . . . . . . . 53

Alpínská vřesoviště Alpine heathlands (M. Kočí & M. Chytrý) . . . . . . . . . . . . . . . . . . . . . . . . . 65Třída AA. Loiseleurio-Vaccinietea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65Svaz AAA. Loiseleurio procumbentis-Vaccinion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

AAA01. Avenello flexuosae-Callunetum vulgaris . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67AAA02. Junco trifidi-Empetretum hermaphroditi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

Acidofilní alpínské trávníky Alpine grasslands on base-poor soils (M. Kočí) . . . . . . . . . . . 76Třída AB. Juncetea trifidi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76Svaz ABA. Juncion trifidi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

ABA01. Cetrario-Festucetum supinae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77Svaz ABB. Nardo strictae-Caricion bigelowii . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

ABB01. Carici bigelowii-Nardetum strictae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

Bazifilní alpínské trávníky Alpine grasslands on base-rich soils (M. Kočí) . . . . . . . . . . . . . 84Třída AC. Elyno-Seslerietea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84Svaz ACA. Agrostion alpinae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

ACA01. Saxifrago oppositifoliae-Festucetum versicoloris . . . . . . . . . . . . . . . . . . . . . . . . . . . 86ACA02. Saxifrago paniculatae-Agrostietum alpinae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

Subalpínská vysokobylinná a křovinná vegetaceSubalpine tall-forb and deciduous-shrub vegetation (M. Kočí) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91Třída AD. Mulgedio-Aconitetea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92Svaz ADA. Calamagrostion villosae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

ADA01. Sphagno compacti-Molinietum caeruleae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94ADA02. Crepido conyzifoliae-Calamagrostietum villosae . . . . . . . . . . . . . . . . . . . . . . . . . . . 96ADA03. Violo sudeticae-Deschampsietum cespitosae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

Svaz ADB. Calamagrostion arundinaceae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106ADB01. Bupleuro longifoliae-Calamagrostietum arundinaceae . . . . . . . . . . . . . . . . . . . . .106

8

Obsah

Svaz ADC. Salicion silesiacae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .109ADC01. Salici silesiacae-Betuletum carpaticae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110ADC02. Pado borealis-Sorbetum aucupariae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .112

Svaz ADD. Adenostylion alliariae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .115ADD01. Ranunculo platanifolii-Adenostyletum alliariae . . . . . . . . . . . . . . . . . . . . . . . . . . .116ADD02. Salicetum lapponum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118ADD03. Trollio altissimi-Geranietum sylvatici . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .120ADD04. Laserpitio archangelicae-Dactylidetum glomeratae . . . . . . . . . . . . . . . . . . . . . . . 122ADD05. Chaerophyllo hirsuti-Cicerbitetum alpinae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .124

Svaz ADE. Dryopterido filicis-maris-Athyrion distentifolii . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126ADE01. Daphno mezerei-Dryopteridetum filicis-maris . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127ADE02. Adenostylo alliariae-Athyrietum distentifolii . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .129

Vegetace jednoletých halofilních travinVegetation of annual graminoids in saline habitats (K. Šumberová) . . . . . . . . . . . . . . . . . . . . . . . . 132Třída TA. Crypsietea aculeatae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132Svaz TAA. Cypero-Spergularion salinae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .133

TAA01. Crypsietum aculeatae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133TAA02. Heleochloëtum schoenoidis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135

Vegetace jednoletých sukulentních halofytůVegetation of annual succulent halophytes (K. Šumberová) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143Třída TB. Thero-Salicornietea strictae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143Svaz TBA. Salicornion prostratae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .145

TBA01. Salicornietum prostratae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .145TBA02. Spergulario marginatae-Suaedetum prostratae . . . . . . . . . . . . . . . . . . . . . . . . . . .147

Slaniskové trávníky Saline grasslands (K. Šumberová, J. Novák & J. Sádlo) . . . . . . . . . . . .150Třída TC. Festuco-Puccinellietea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150Svaz TCA. Puccinellion limosae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152

TCA01. Puccinellietum limosae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .153Svaz TCB. Juncion gerardii . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .155

TCB01. Scorzonero parviflorae-Juncetum gerardii . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .157TCB02. Loto tenuis-Potentilletum anserinae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .159TCB03. Agrostio stoloniferae-Juncetum ranarii . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162

Louky a mezofilní pastviny Meadows and mesic pastures(P. Hájková, M. Hájek, D. Blažková, T. Kučera, M. Chytrý, M. Řezníčková,K. Šumberová, T. Černý, J. Novák & D. Simonová) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165Třída TD. Molinio-Arrhenatheretea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .166Svaz TDA. Arrhenatherion elatioris . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .168

TDA01. Pastinaco sativae-Arrhenatheretum elatioris . . . . . . . . . . . . . . . . . . . . . . . . . . . . .170TDA02. Ranunculo bulbosi-Arrhenatheretum elatioris . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173TDA03. Poo-Trisetetum flavescentis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175TDA04. Potentillo albae-Festucetum rubrae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .178

Svaz TDB. Polygono bistortae-Trisetion flavescentis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .188TDB01. Geranio sylvatici-Trisetetum flavescentis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .189TDB02. Melandrio rubri-Phleetum alpini . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191TDB03. Meo athamantici-Festucetum rubrae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .193

9

Obsah

Svaz TDC. Cynosurion cristati . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .195TDC01. Lolio perennis-Cynosuretum cristati . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .197TDC02. Anthoxantho odorati-Agrostietum tenuis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .200TDC03. Lolietum perennis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .203TDC04. Prunello vulgaris-Ranunculetum repentis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .205TDC05. Alchemillo hybridae-Poëtum supinae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .207

Svaz TDD. Molinion caeruleae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .209TDD01. Molinietum caeruleae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .210TDD02. Junco effusi-Molinietum caeruleae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .213

Svaz TDE. Deschampsion cespitosae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .220TDE01. Poo trivialis-Alopecuretum pratensis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .223TDE02. Holcetum lanati . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226TDE03. Lathyro palustris-Gratioletum officinalis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .229TDE04. Cnidio dubii-Deschampsietum cespitosae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .233TDE05. Scutellario hastifoliae-Veronicetum longifoliae . . . . . . . . . . . . . . . . . . . . . . . . . . . .236

Svaz TDF. Calthion palustris . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .238TDF01. Angelico sylvestris-Cirsietum oleracei . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .241TDF02. Cirsietum rivularis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .244TDF03. Angelico sylvestris-Cirsietum palustris . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .247TDF04. Crepido paludosae-Juncetum acutiflori . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .255TDF05. Polygono bistortae-Cirsietum heterophylli . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .257TDF06. Chaerophyllo hirsuti-Calthetum palustris . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .259TDF07. Scirpo sylvatici-Cirsietum cani . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262TDF08. Scirpetum sylvatici . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .264TDF09. Caricetum cespitosae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .267TDF10. Scirpo sylvatici-Caricetum brizoidis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .269TDF11. Junco inflexi-Menthetum longifoliae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .271TDF12. Filipendulo ulmariae-Geranietum palustris . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .274TDF13. Lysimachio vulgaris-Filipenduletum ulmariae . . . . . . . . . . . . . . . . . . . . . . . . . . . . .276TDF14. Chaerophyllo hirsuti-Filipenduletum ulmariae . . . . . . . . . . . . . . . . . . . . . . . . . . . .278

Smilkové trávníky a vřesovištěNardus grasslands and heathlands (F. Krahulec, M. Chytrý & H. Härtel) . . . . . . . . . . . . . . . . . . . .281Třída TE. Calluno-Ulicetea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .281Svaz TEA. Nardion strictae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .283

TEA01. Festuco supinae-Nardetum strictae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .284TEA02. Thesio alpini-Nardetum strictae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286

Svaz TEB. Nardo strictae-Agrostion tenuis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .295TEB01. Sileno vulgaris-Nardetum strictae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .295

Svaz TEC. Violion caninae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .298TEC01. Festuco capillatae-Nardetum strictae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .299TEC02. Campanulo rotundifoliae-Dianthetum deltoidis . . . . . . . . . . . . . . . . . . . . . . . . . . .301

Svaz TED. Nardo strictae-Juncion squarrosi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .304TED01. Juncetum squarrosi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .304

Svaz TEE. Euphorbio cyparissiae-Callunion vulgaris . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .307TEE01. Euphorbio cyparissiae-Callunetum vulgaris . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .308

Svaz TEF. Genisto pilosae-Vaccinion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .311TEF01. Vaccinio-Callunetum vulgaris . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .312TEF02. Calamagrostio arundinaceae-Vaccinietum myrtilli . . . . . . . . . . . . . . . . . . . . . . . . .314TEF03. Festuco supinae-Vaccinietum myrtilli . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .317

10

Obsah

Pionýrská vegetace písčin a mělkých půdPioneer vegetation of sandy and shallow soils (J. Sádlo, M. Chytrý & T. Černý) . . . . . . . . . . . . . .320Třída TF. Koelerio-Corynephoretea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .321Svaz TFA. Corynephorion canescentis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .323

TFA01. Corniculario aculeatae-Corynephoretum canescentis . . . . . . . . . . . . . . . . . . . . . .325TFA02. Festuco psammophilae-Koelerietum glaucae . . . . . . . . . . . . . . . . . . . . . . . . . . . . .328

Svaz TFB. Thero-Airion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .337TFB01. Airetum praecocis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .338TFB02. Vulpietum myuri . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .341

Svaz TFC. Armerion elongatae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343TFC01. Sileno otitae-Festucetum brevipilae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .345TFC02. Erysimo diffusi-Agrostietum capillaris . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .347

Svaz TFD. Hyperico perforati-Scleranthion perennis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 350TFD01. Polytricho piliferi-Scleranthetum perennis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .351TFD02. Jasiono montanae-Festucetum ovinae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .353

Svaz TFE. Arabidopsion thalianae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356TFE01. Festuco-Veronicetum dillenii . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .356

Svaz TFF. Alysso alyssoidis-Sedion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359TFF01. Cerastietum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .360TFF02. Alysso alyssoidis-Sedetum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363

Písečné stepi Sandy steppes (M. Chytrý) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .366Třída TG. Festucetea vaginatae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 366Svaz TGA. Festucion vaginatae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367

TGA01. Diantho serotini-Festucetum vaginatae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367

Suché trávníky Dry grasslands (M. Chytrý, A. Hoffmann & J. Novák) . . . . . . . . . . . . . . . . . . .371Třída TH. Festuco-Brometea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .372Svaz THA. Alysso-Festucion pallentis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 376

THA01. Festuco pallentis-Aurinietum saxatilis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .378THA02. Seselio ossei-Festucetum pallentis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .380THA03. Sedo albi-Allietum montani . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382THA04. Helichryso arenariae-Festucetum pallentis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .384

Svaz THB. Bromo pannonici-Festucion pallentis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .395THB01. Poo badensis-Festucetum pallentis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .396

Svaz THC. Diantho lumnitzeri-Seslerion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .398THC01. Carici humilis-Seslerietum caeruleae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .399THC02. Minuartio setaceae-Seslerietum caeruleae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .402THC03. Saxifrago paniculatae-Seslerietum caeruleae . . . . . . . . . . . . . . . . . . . . . . . . . . . . 404THC04. Asplenio cuneifolii-Seslerietum caeruleae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .406

Svaz THD. Festucion valesiacae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .409THD01. Festuco valesiacae-Stipetum capillatae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 411THD02. Erysimo crepidifolii-Festucetum valesiacae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .413THD03. Festuco rupicolae-Caricetum humilis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .416THD04. Koelerio macranthae-Stipetum joannis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 418THD05. Stipetum tirsae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .421THD06. Astragalo exscapi-Crambetum tatariae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423

Svaz THE. Cirsio-Brachypodion pinnati . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .425THE01. Scabioso ochroleucae-Brachypodietum pinnati . . . . . . . . . . . . . . . . . . . . . . . . . . .427THE02. Cirsio pannonici-Seslerietum caeruleae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 429THE03. Polygalo majoris-Brachypodietum pinnati . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .432

11

Obsah

THE04. Plantagini maritimae-Caricetum flaccae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .435Svaz THF. Bromion erecti . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .443

THF01. Carlino acaulis-Brometum erecti . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 444THF02. Brachypodio pinnati-Molinietum arundinaceae . . . . . . . . . . . . . . . . . . . . . . . . . . .447

Svaz THG. Koelerio-Phleion phleoidis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .449THG01. Potentillo heptaphyllae-Festucetum rupicolae . . . . . . . . . . . . . . . . . . . . . . . . . . . .451THG02. Avenulo pratensis-Festucetum valesiacae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .454THG03. Viscario vulgaris-Avenuletum pratensis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .456

Svaz THH. Geranion sanguinei . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .458THH01. Trifolio alpestris-Geranietum sanguinei . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .459THH02. Geranio sanguinei-Dictamnetum albae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .461THH03. Geranio sanguinei-Peucedanetum cervariae . . . . . . . . . . . . . . . . . . . . . . . . . . . . .463

Svaz THI. Trifolion medii . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .466THI01. Trifolio medii-Agrimonietum eupatoriae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .467THI02. Trifolio-Melampyretum nemorosi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .468

Literatura References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 471

Rejstřík vědeckých jmen rostlin a syntaxonůIndex of plant and syntaxon names . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .499

35

Project Vegetationof the Czech Republic:

Preface and summary of methodsMilan Chytrý

of the Czechoslovak Academy of Sciences inPrůhonice became the centre of phytosociologi-cal research on Czechoslovak vegetation. Theteam led by Rudolf Mikyška, which included dis-tinguished scientists such as Jaroslav Moravec,Robert Neuhäusl and Zdenka Neuhäuslová, stud-ied Czechoslovak vegetation within the frameworkof an ambitious project aimed at mapping re-constructed natural vegetation. The map waspublished as a book with explanatory text (Mikyš-ka et al. 1968) and map sheets at the scale of1 : 200 000. Besides vegetation mapping, whichcontributed a great deal to the knowledge on thediversity of natural forest vegetation, research intoother types of plant communities also progressedmarkedly in the 1960s. Jan Jeník, Charles Univer-sity, Prague, focussed on alpine and subalpinevegetation, while Slavomil Hejný and Karel Ko-pecký from the Institute of Botany in Průhonicestudied aquatic, wetland and synanthropic vege-tation types. University and academic centres inBrno also contributed to this kind of research: JiříVicherek studied various types of grassland andwetland vegetation, Emilie Balátová-Tuláčková in-vestigated meadows and Kamil Rybníček mires.The results obtained in this period appeared ina new overview of plant communities at the levelof phytosociological alliances and classes (Holubet al. 1967).

Since the 1960s phytosociological classifica-tion of vegetation has been routinely implementedas an approach to inventory in institutions en-gaged in environmental protection and botanicalresearch of selected areas. Further major contri-butions to the knowledge of plant communities invarious regions within the Czech Republic werepublished by the above-mentioned authors as wellas Denisa Blažková, František Grüll, Emil Hadač,Miroslava Husová, Vladimír Jehlík, Jiří Kolbek,Jarmila Kubíková, Antonín Pyšek, Jaroslav Rydlo,

The development and currentstate of phytosociologicalresearch in the Czech Republic

The classification of plant communities usingrelevés, i.e. the lists of plant species includingdata on their quantitative representation in smallplots, has a long tradition in Central Europe, dat-ing back to the first decades of the 20th century.The basic methodological principles of phyto-sociology were formulated by the Swiss plantecologist Josias Braun-Blanquet (1921, 1928).Braun-Blanquet’s methods spread quickly in anumber of European countries. The first pioneer-ing studies in the former Czechoslovakia werecarried out e.g. by Jaromír Klika, Vladimír Kraji-na, Rudolf Mikyška, Pavel Sillinger and AloisZlatník in the 1920s and 1930s. They focussedon classification and inventory of plant commu-nities and provided a good overview of majorvegetation types, summarized by Jaromír Klikain the annotated lists of plant communities ofCzechoslovakia and Central Europe (Klika in Kli-ka & Novák 1941: 53–71, Klika & Hadač 1944,Klika 1948, 1955). These lists were elaborated tothe level of phytosociological alliance and nei-ther included associations – the fundamentalunits of vegetation classification. However, theywere compiled at approximately the same timeas the foremost representatives of European phy-tosociology, Josias Braun-Blanquet and ReinholdTüxen, attempted to make an analogous list ofthe vegetation of Central Europe (Braun-Blanquet& Tüxen 1943). Czech vegetation scientists thuscontributed significantly to the formation of thebasic classification scheme of European vege-tation.

After World War II, the newly established Geo-botanical Laboratory and later Institute of Botany

36

Project Vegetation of the Czech Republic: Preface and summary of methods

Jaromír Sofron, Tomáš Sýkora, Miloslav Tomanand others. At the beginning of the 1980s theknowledge of plant communities of the Czech Re-public was sufficiently detailed to allow thecompilation of the first overview of vegetationunits at the level of association (Moravec et al.1983a; the second, updated edition was publishedin 1995). The second edition contains 665 asso-ciations, 150 alliances, and 44 classes. Althoughthis overview played a key role in the synthesis ofcurrent knowledge and remains the only com-plete survey of vegetation units in the CzechRepublic at the level of associations, it has a num-ber of shortcomings. The specification of manyvegetation units included in the overview is ratherambiguous and their characteristics are too briefto be sufficient for identification. Attempts weremade to eliminate these shortcomings by addingmore detailed descriptions to all associations inthe series entitled Vegetation Survey of the CzechRepublic, edited by Jaroslav Moravec. However,even this series did not provide the necessarybaseline data of a modern survey, i.e. tables withfloristic composition of vegetation units and dis-tribution maps. The volumes that have beenpublished so far (Moravec 1998, Moravec et al.2000, Husová et al. 2002, Neuhäuslová 2003) dealonly with forest vegetation.

Since the 1990s phytosociological classifi-cation has acquired increasing importance inEurope, particularly in the connection with theadoption of the Habitats Directive (92/43/EEC).The Habitats Directive implemented the principlethat the selection of protected areas shouldproceed on the basis of the representative distri-bution of endangered habitats. Vegetation is themost suitable component for the typification ofterrestrial habitats and phytosociology thus seemsto be a suitable method for habitat classification.Being based on the detailed analyses of speciescomposition of plant communities, phytosocio-logy is also crucial for biodiversity protection.Therefore, the phytosociological system wasadopted in the corrected form in the Europeanschemes of habitat classifications such as EUNIS(Davies & Moss 1997; http://eunis.finsiel.ro/eunis/habitats.jsp). The European systems of habitatclassification were adapted to the needs of theCzech Republic and interpreted in the HabitatCatalogue of the Czech Republic (Chytrý et al.2001a), which became the starting point for the

national mapping of habitats within the Natura2000 project from 2001 to 2004.

The “Vegetation of the CzechRepublic” project

The demand by environment protection agenciesfor consistent and well-documented systems ofvegetation classification led in the 1990s to theimplementation of modern national projects ofvegetation classification in some European coun-tries, e.g., Great Britain (Rodwell 1990–2000),Austria (Mucina et al. 1993b), the Netherlands(Schaminée et al. 1995–1999), Slovakia (Valacho-vič et al. 1995, Jarolímek et al. 1997, Valachovič2001), Germany (Dierschke 1996 et seq.) and theGerman federal state of Mecklenburg-Vorpom-mern (Berg et al. 2004). These projects have thefollowing features in common: thorough revisionof previously described vegetation units on thebasis of a critical reassessment of large relevédata sets; documentation of accepted asso-ciations using species composition tables; thedetailed revision of the nomenclature of vegeta-tion units; and the compilation of the distributionmaps of phytosociological associations within thearea in question. Research teams have sharedtheir experiences from these projects with vege-tation scientists from other European countriesat the annual meetings of the European Vegeta-tion Survey working group that have been heldevery year since 1992 (Mucina et al. 1993c, Rod-well et al. 1995). This working group has alsoproduced the European synopsis of vegetationunits at levels ranging from classes to alliances(Mucina 1997a, Rodwell et al. 2002).

Despite a long tradition of phytosociologicalresearch and a good level of documentation ofvegetation, the Czech Republic did not possessthe modern classification of vegetation. In 1995 adecision was made to start to work on a newmonograph entitled Vegetation of the Czech Re-public. The initial, partial goal was to generate theCzech National Phytosociological Database whichwould contain the representative sample of rele-vés from different habitats and regions of theCzech Republic in an easily accessible electronicformat (Chytrý & Rafajová 2003). Such relevésexisted but were scattered in a number of scien-tific books, articles, theses, unpublished research

37


reports, inventory surveys of protected areas, fieldbooks and other written materials maintained bydifferent botanists. Thanks to Professor John S.Rodwell from Lancaster University (UK) and Ste-phan M. Hennekens from Alterra – Green WorldResearch in Wageningen (The Netherlands), wewere able to use know-how from the British andDutch vegetation classification projects fromthe very beginning of our work on the database.In 1995–1997, John Rodwell arranged a seriesof courses to acquaint with the principles ofphytosociological database management andparticularly with the computer program TURBO-VEG (Hennekens 1995, Hennekens & Schaminée2001). The author of this program Stephan M.Hennekens kindly provided it to Czech users freeof charge. TURBOVEG was prepared for use inCentral Europe in cooperation with colleaguesfrom Austria (Ladislav Mucina, Harald Niklfeld,Walter Gutermann) and Slovakia (Milan Valacho-vič, Ivan Jarolímek) and in 1996 it was madeaccessible to all vegetation scientists in the CzechRepublic (Chytrý 1996). In February 1997 MasarykUniversity in cooperation with John Rodwell or-ganized a TURBOVEG training course in Brno forcolleagues and students from the Czech Repub-lic and Slovakia. Subsequently, a network of localTURBOVEG coordinators was established, whichcovered all major botanical institutions in theCzech Republic.

Financial support from the Czech ScienceFoundation has enabled the employment of adatabase administrator since 1999. This positionwas occupied by Marie Rafajová (1999–2003),Ilona Knollová (2003–2005) and Štěpánka Králová(since 2005). Professional database managementand contributions from a number of co-workersfrom Masaryk University, the Institute of Botanyof the Academy of Sciences of the Czech Repub-lic and other institutions resulted in the fastpopulation of the database (Chytrý & Rafajová2003). By the end of 2005, the database includeda total of 72,476 relevés, recorded on the territo-ry of the Czech Republic in the period from 1922to 2005. The database is thus probably third larg-est in the world following the Dutch and Frenchdatabases (Ewald 2001).

Besides the creation of the relevé database,the preparation of the monograph Vegetation ofthe Czech Republic also necessitated the devel-opment and testing of methods for vegetation

classification using large data sets. In the caseof data sets with tens of thousands of relevés,standard methods developed for classification ofsmaller data sets are not suitable or do not allowthe potential of these data to be fully exploited. Inaddition, there has been only very limited expe-rience with how various shortcomings in dataquality affect the analysis of large relevé datasets. It was therefore necessary to perform vari-ous methodological studies. One of the challengeswas to establish a method of selection of rele-vés from the database which would prevent thenegative impact on the resultant classificationcaused by the uneven distribution of relevés with-in the Czech Republic (Knollová et al. 2005). Forvegetation classification on the basis of a phyto-sociological database, we selected the Cocktailmethod (Bruelheide 1995, 2000). This method cre-ates explicit definitions of vegetation units whichallow an unambiguous assignment of every relevéto these units. It thus allows matching of newlyobtained relevés to the units of establishedclassification. The Cocktail method underwentcomprehensive testing and modifications and wasextended with a procedure that enabled the as-signment of relevés to vegetation units based onsimilarity (Kočí et al. 2003, Tichý 2005). Attentionwas also devoted to the testing and developmentof statistical methods to determine species fidel-ity to vegetation units (Chytrý et al. 2002, Tichý& Chytrý 2006), an important criterion in thedetermination of diagnostic species and the pre-sentation of vegetation classification in tables.Chytrý & Tichý (2003) calculated species fidelitiesto vegetation classes and alliances of the currentstandard vegetation classification of the CzechRepublic (Moravec et al. 1995), using data fromthe Czech National Phytosociological Database.Based on this analysis, they were able to evaluatethe quality of delimitation of vegetation units. Thiswork was used as a guideline for identification of(1) which vegetation units from the current classi-fication should be adopted in the new classificationsystem and (2) which should be eliminated ormodified. Since 1998 all methods of analysis ofphytosociological data used in the project havebeen included in the computer program JUICE(Tichý 2002), which has become a tool for the com-prehensive analysis of phytosociological data andis currently being used by a number of individualsand institutions in many countries worldwide.

38


The preparation of the first volume of themonograph Vegetation of the Czech Republic wassupported by the Czech Science Foundation(GAČR) grant no. 206/02/0957 (2002–2004). Theproject was managed at the former Departmentof Botany (now Department of Botany and Zoolo-gy), Masaryk University, Brno, with the Institute ofBotany of the Academy of Sciences of the CzechRepublic acting as a cooperating institution. Be-sides the authors of the treatments of individualsyntaxa, the project involved several colleagueswithout whose contributions this book would nev-er have been completed. Lubomír Tichý developedsoftware tools for the comprehensive analysis ofphytosociological data. Ilona Knollová managedthe phytosociological database and prepared thedata sets for analysis. Zdenka Otýpková andKatrin Karimová were involved in editing of thedatabase. Jiří Danihelka produced a database forautomatic conversion of species taxonomy andnomenclature used in the TURBOVEG programto those used in the Key to the Flora of the CzechRepublic and also undertook a detailed linguisticrevision of the Czech text. Ondřej Hájek preparedall the maps used in this book and took part inthe preparation of the predictive distribution mod-els. Klára Kubošová developed the statisticalmodels used for the prediction of potential distri-bution. Jiří Rozehnal was responsible for thedesign of graphs and provided the hardware andsoftware support. The final editing of the manu-script and publication of this book were supportedby the grants nos. GAČR 206/05/0020, GAČR206/06/0659 and MSM 0021622416.

The hierarchy of the systemof plant communities

The European phytosociological school recog-nizes four main hierarchical levels (ranks) ofphytosociological units (syntaxa), which differ bythe endings of their Latin names. The ranks, ar-ranged from the lowest to the highest, are asfollows: association (ending -etum), alliance (-ion),order (-etalia) and class (-etea). Apart from syn-taxa of these main ranks, there are also syntaxaof supplementary ranks, subassociation (-eto-sum), suballiance (-enion), suborder (-enalia) andsubclass (-enea). The nomenclature of these syn-taxa is governed by the International Code of

Phytosociological Nomenclature (ICPN; hereafterreferred to as the Code; Weber et al. 2000).

We believe that a simple hierarchy is con-venient for the practical use of a system ofvegetation units on the national level. Therefore,we only use four hierarchical ranks in the pre-sented system, namely class, alliance, associationand variant. Classes, alliances and associationsare frequently used in practice to identify variousvegetation types. We do not use orders as theyare not of great significance on the national level,and because classes usually contain a manage-able number of alliances, which do not need tobe arranged hierarchically by introducing yet an-other classification rank. Moreover, the applicationof orders has often not stabilized, since theirmeaningful definition would require the revisionof the respective vegetation classes throughouttheir distribution range; this has not yet been pos-sible with the current level of knowledge of thediversity of European vegetation. We also do notuse suballiances, suborders or subclasses. Untilnow suballiances have been used in the CzechRepublic for only a few alliances (Moravec et al.1995). However, we do consider it suitable to useclassification units below the rank of associationin order to express internal variability. Subassoci-ations could be suitable for this purpose, but theproblem is that for many associations, there areplenty of subassociations described that haverather limited, local validity, overlap each otherand are defined according to mutually incompat-ible criteria. Based on analysis of the variability ofrelevés within a given association we have at-tempted in the present work to define two to fourmajor subtypes inside the associations. Oftenthese subtypes could not be unambiguously iden-tified with the described subassociations, andthus named correctly according to the Code. Wetherefore used the variant as a classification unitat the hierarchical level below the association,because its nomenclature, unlike that of subas-sociation, is not governed by the Code. Thismeans that ad hoc names can be used irrespec-tive of the units described previously. We do notdistinguish variants for relatively homogeneousassociations with small internal variability.

For the purpose of the coding of vegetationunits in maps and databases, all vegetation unitsused are given unique codes, which simulta-neously reflect their rank in the classification

39


hierarchy, e.g. TBB03, or TBB03a, respectively.The meaning of these abbreviations is as follows:– The first letter indicates the formation group

according to the Habitat Catalogue of theCzech Republic (Chytrý et al. 2001a) and isselected to reflect the name of the formationgroup. For instance, the letter A denotes alpinevegetation while the letter T indicates vegeta-tion of secondary grasslands and heathlands;

– The second letter is presented in the alpha-betical order and indicates the class withinthe formation group;

– The third letter is also given in alphabeticalorder and indicates the alliance within theclass;

– The two-digit number indicates the associa-tion within the alliance;

– The small letter provided in alphabetical orderindicates the variant within the association.

The concept of associationsand other vegetation units

The variability of species composition in plantcommunities is usually continuous: clearly definedvegetation units are rather rare in nature. How-ever, vegetation classification is important forpractical purposes, e.g. for habitat inventory andmapping. Classification can be performed usingseveral alternative approaches, of which none canbe declared the best. The classification in themonograph Vegetation of the Czech Republiclargely adopted traditional vegetation units usedin the last edition of the list of plant communitiesof the Czech Republic (Moravec et al. 1995), inthe vegetation overviews of neighbouring coun-tries (Mucina et al. 1993b, Schaminée 1995–1999,Pott 1995, Valachovič et al. 1995 et seq., Matusz-kiewicz 2001, Schubert et al. 2001, Borhidi 2003,Berg et al. 2004) and in the pan-European over-views (Mucina 1997a, Rodwell et al. 2002). Wedid not aim at the formation of a new classifi-cation but at a critical review of the currentclassification, to which Czech and Central Euro-pean users are accustomed. The main goals ofthis revision, based on analysis of large relevédata sets, were as follows: (1) to eliminateoverlaps in the definitions of the currently distin-guished vegetation units, (2) to exclude from thesyntaxonomic system units with poor floristic dif-

ferentiation, which are difficult to recognize bothin the field and in the databases, and (3) to adaptthe definitions of the vegetation units of the CzechRepublic to concepts accepted in neighbouringcountries, provided they are not in conflict withthe variability of the vegetation occurring on theterritory of the Czech Republic.

The large relevé data sets can be classifiedusing methods of unsupervised or supervisedclassification (Ejrnæs et al. 2004). Unsupervisedclassification algorithms seek the main gradientsin species composition, more significant dis-continuities and relatively homogeneous relevégroups within the relevé data sets. The use ofthese algorithms yields the division of the relevésinto groups which is dependent only on the in-formation contained in the data set. The mostcommonly used methods of unsupervised classi-fication include TWINSPAN (Hill 1979) and clusteranalysis (Legendre & Legendre 1998, Podani2000). Different variants of these algorithms aswell as different data transformations provideslightly different results. However, the majority ofthem reflect the variability inside the data set quitewell. The main disadvantage of unsupervised clas-sification is that the result is always unique forthe given set of relevés. If the data set is partiallyaltered, e.g. by the addition of newly obtainedrelevés, the classification may change consider-ably and some of the relevés that have beenassigned to a particular group during the classifi-cation of the original set will be assigned toa different group during the classification of thealtered set. Unsupervised classifications thus donot ensure the stability of vegetation classifica-tion of large areas. The project Vegetation of theCzech Republic employed unsupervised classi-fications in pilot studies to identify the maingradients within individual vegetation types on theterritory of the Czech Republic and neighbouringcountries (e.g. Havlová et al. 2004, Botta-Dukátet al. 2005). However, the final version of theclassification was generated using a supervisedclassification method.

Supervised classifications use external pre-defined criteria of what the individual vegetationtypes should look like. These criteria are inde-pendent of the data set being classified. Thevegetation classification at the level of associ-ations in the project Vegetation of the CzechRepublic was performed using the supervised

40


classification method Cocktail (Bruelheide 1995,2000). This method imitates the traditional phy-tosociological classification approach using thesociological groups of species. It was tested withvarious data sets and slightly modified in com-parison with its original version (Bruelheide &Chytrý 2000, Kočí et al. 2003, Lososová 2004).Based on large sets of relevés, Cocktail quanti-fies species co-occurrence rates and generatessociological groups from species with a strongtendency to occur together in relevés. When so-ciological groups are to be generated, the initialspecies of the groups can be selected subjectivelyso as to characterize the traditional vegetationunits well. However, the assignment of otherspecies to a particular group is subjected to astatistical check to indicate whether one speciesmatches the group better than others. Sociologi-cal groups are named after one of the group’sspecies. The next phase delimits the vegetationunits by means of formal definitions with the log-ical operators AND, OR or NOT. These definitionsdetermine which sociological groups must bepresent or absent in a relevé in order this relevécan be assigned to the particular vegetation unit.We produced Cocktail definitions of associationsbut not of the vegetation units of other ranks. Thetesting of the Cocktail method revealed that alarge number of traditional associations could notbe defined purely by floristic composition, with-out considering dominance of some species(Kočí et al. 2003). As a result, the formal defini-tions include the dominance of selected individualspecies, in addition to the presence/absence ofsociological species groups. For example, theCocktail definition of the association Angelicosylvestris-Cirsietum oleracei has the followingform:

Group Caltha palustris AND Group Cirsiumoleraceum NOT Group Cirsium rivulare NOTCarex cespitosa cover > 25 % NOT Filipendu-la ulmaria cover > 25 %.

This means that a relevé is assigned to the asso-ciation if it contains both the Caltha palustrissociological group and the Cirsium oleraceumgroup and, at the same time, it does not containthe Cirsium rivulare group nor the species Carexcespitosa with the cover higher than 25% nor thespecies Filipendula ulmaria with the cover higher

than 25%. The sociological group is consideredto be represented in the relevé if the relevé con-tains at least half of all species of the group. Theoverview of sociological species groups used inthe first volume of Vegetation of the Czech Re-public is provided in Table 1 (page 22). The formaldefinitions of associations were elaborated insuch a way that the groups of relevés specifiedby them matched as much as possible the tradi-tionally distinguished associations. In the courseof the elaboration of formal definitions attemptwas made to formally define all associations in-cluded in the latest list of vegetation units ofthe Czech Republic (Moravec et al. 1995) andsome associations not yet reported from theCzech Republic but recognized in neighbouringcountries. It was shown that many associationsmentioned by Moravec et al. (1995) overlap in theirdelimitations while others cannot be defined atthe national scale at all due to their poor floristicdifferentiation. As a result, the number of associ-ations recognized in the monograph Vegetation ofthe Czech Republic is lower than in the list byMoravec et al. (1995). However, the acceptedassociations are clearly defined and well recog-nizable. Thus, the Cocktail method prevents theinflation of vegetation units (Pignatti 1968), i.e.the description of an increasing number of asso-ciations which are usually of limited local validityor have poor differentiation as compared with thepreviously described associations.

It should be emphasized that associationsdefined by the Cocktail method are defined sub-jectively, i.e. like the associations in the traditionalphytosociological classification. However, theyhave one major advantage compared with tra-ditional classification: they are defined usingunambiguous criteria which allow the consistentassignment of any relevé to the particular associ-ation. They can therefore serve as a suitable basisfor computer expert systems, i.e. programs thatwill compare every submitted relevé with the for-mal definitions of associations and assign it tothe particular association.

An important property of the Cocktail methodis that some relevés, particularly those consistingof mainly generalist species, are not assigned toany association and thus remain unclassified. Upto 50–70% of relevés usually remained unclassi-fied in our tests. This feature reflects traditionalphytosociological experience that the majority of

41


vegetation stands occurring in the field cannotbe assigned to associations. However, in somepractical applications of phytosociological clas-sification, e.g. in vegetation mapping, it is notdesirable that some stands are not assigned toclassification units. We have therefore developeda two-step classification. In the first step, therelevés are assigned to associations accordingto formal definitions. In the second step, therelevés that have not been assigned to any asso-ciation by formal definitions are compared withthe species composition of groups of relevés thathave already been assigned to individual associ-ations and subsequently assigned to the mostmatching association (Kočí et al. 2003, Tichý2005).

The adopted concept of alliances and class-es relies particularly on the statistical analysis ofthe quality of the delimitation of alliances andclasses described by Moravec et al. (1995) thatwas performed by Chytrý & Tichý (2003) on thebasis of the data from the Czech National Phy-tosociological Database. The analysis evaluatednot only the sharpness of each unit, i.e. the num-ber and the quality of its diagnostic species, butalso the uniqueness of the unit, i.e. the degree ofoverlap of the particular vegetation unit with othervegetation units. As an auxiliary criterion, themajor alliances and classes accepted in the veg-etation overviews of neighbouring countries andinternational overviews were also taken into ac-count.

The technical procedureof defining associationsand other vegetation units

The set of 53,097 relevés from the Czech Republicavailable in the Czech National PhytosociologicalDatabase on 1 July 2002 was used to generatesociological groups of species and formal defini-tions using the Cocktail method. This set waselaborated to contain as many different vegeta-tion types as possible. It thus includes not onlygrassland and low-shrub vegetation, but alsorelevés of anthropogenic, aquatic, wetland, chas-mophytic, shrub and forest vegetation. First, atotal of 636 relevés recorded using plots of un-usually large or small size (Chytrý & Otýpková

2003) were eliminated from the data set. This ap-plied to any plots < 50 m2 or > 1000 m2 for forests,< 10 m2 or > 100 m2 for shrub vegetation and< 4 m2 or > 100 m2 for herbaceous vegetation.This was followed by the elimination of 12,740relevés that had not been assigned to syntaxa atleast at the level of the class according to thestandard national list (Moravec et al. 1995) or thatlacked geographical localization with an accura-cy of at least one geographical minute. All otherdata analyses were performed in the JUICE pro-gram (Tichý 2002), unless stated otherwise.

The geographical distribution of the remain-ing relevés was relatively uneven: some territorieswere surveyed sufficiently, with quite a large num-ber of relevés available, while fewer or no relevésat all were available from other territories (Chyt-rý & Rafajová 2003). We therefore performeda stratified selection of relevés to ensure that novegetation type was represented by a high num-ber of relevés from a small area and that theformation of sociological groups was not affect-ed by locally specific coincidences of speciesoccurrence (Knollová et al. 2005). Stratificationwas performed in a geographical grid with cellssized 1.25 minutes of longitude × 0.75 minutes oflatitude, i.e. approximately 1.5 × 1.4 km. If two ormore relevés assigned by their authors to thesame association fell in the same grid cell, onlyone of them was selected. The selection preferredrelevés with a record of the moss layer and morerecent relevés. If there were still some relevés ofthe same record date left, one of them was select-ed at random. This stratified selection generateda data set containing 21,794 relevés, which wasused to generate sociological species groups, testthe Cocktail definitions of associations and per-form other analyses.

All records of juvenile trees and shrubs in theherb layer were deleted from the stratified dataset because some authors recorded them whileothers did not. The multiple records of the identi-cal species in the tree and shrub layers werecombined. Similarly, the records of low shrubs orhigh herbs recorded in the herb layer by someauthors and in the shrub layer by others (e.g.Calluna vulgaris, Cotoneaster spp., Daphne me-zereum, Prunus fruticosa, Reynoutria spp., Rosagallica, Rubus spp. and Sambucus ebulus) werealso combined in one layer. The same approachwas applied to lianas recorded in the tree, shrub

42


and herb layers (e.g. Clematis vitalba, Hedera he-lix, Humulus lupulus and Solanum dulcamara). Asa result, all the species were represented onlyonce in the final data.

Bryophytes, foliose and fruticose lichens andmacroscopic algae were maintained in the dataset although they were not recorded in all relevés.This fact could lead in some cases to underesti-mating the significance of these plants in theformation of species groups. Therefore the spe-cies groups were formed with particular emphasison vascular plants since these are recorded in allrelevés.

The taxonomic concepts and nomenclatureof species and subspecies were standardizedaccording to standard works for vascular plants(Kubát et al. 2002), bryophytes (Kučera & Váňa2003) and lichens (Vězda & Liška 1999). Thenames of vascular plants that do not occur inthe Czech Republic mostly follow Ehrendorfer(1973), provided there is no contradiction in thetaxonomic concept accepted in this publicationand in Kubát et al. (2002). Narrowly defined spe-cies or subspecies were unified into a broaderconcept in all cases where most relevés con-tained determinations of broadly defined speciesor when the data on narrowly defined specieswere likely to contain errors. In cases when theKey to the Flora of the Czech Republic (Kubát etal. 2002) specifies species aggregates, theseaggregates (agg.) were used for broadly definedspecies. In other cases, the broader species con-cepts were defined particularly for the purposeof the monograph Vegetation of the Czech Re-public and marked with the name of a particularspecies and abbreviation s. lat. (sensu lato).These species are listed in the Czech version ofthe text (page 25). Broader species conceptswere used in the analysed data set. If statisticalanalysis of this data set determined some broad-ly conceived species as diagnostic, constant ordominant of an association, and it is known thatsome narrowly conceived species grows in thisassociation, the species list contains the broad-er species concept, followed by the narrowerconcept in brackets.

The resultant set of 21,794 relevés was usedto generate sociological species groups with theCocktail method (Bruelheide 1995, 2000), whichincluded a few modifications described by Kočíet al. (2003). The degree of co-occurrence was

calculated for each species using the phi coeffi-cient of association (Sokal & Rohlf 1995, Chytrýet al. 2002). When the phi coefficient between twospecies was high, indicating a strong interspecif-ic association that exceeded the association ofeach of these two species to any other species,these two species were used as the basis of asociological group. The next step tested the as-sociation between the common occurrence ofthese two species and the occurrence of the otherspecies. Based on the strength of this associa-tion, third species was added to group. It was aspecies with strong (usually the strongest) asso-ciation with the relevés containing both speciespreviously included in the sociological group. Thespecies with the strongest association was notincluded in the sociological group in the caseswhen it had already been included in anothergroup or when its frequency (i.e. the number ofoccurrences in the data set) differed by an orderof magnitude from those of the species alreadyincluded in the sociological group. In such cases,species with the second or third strongest asso-ciation was included in the group. This was donebecause incorporation of either too rare or toocommon species would lead to the formation ofheterogeneous sociological groups. In the nextsteps, the association between relevés contain-ing at least half of the species included in thesociological group on the one hand and the oc-currence of other individual species on the otherwas calculated. On the basis of this calculation,another species with a strong association wasadded to the group and the process continued inthe same way. The generation of sociologicalgroups usually stopped when they contained 3–5species, because larger groups usually appearedtoo heterogeneous.

Sociological groups and dominances of se-lected species were used to generate logicaldefinitions of associations, in which the require-ments regarding the occurrence of groups orspecies dominances were linked via the logicaloperators AND, OR, and NOT (Bruelheide 1997).The definitions of associations were generatedstepwise so that the group of relevés assigned toa particular association using its logical defini-tion overlapped the group of relevés includedin this association by the relevé authors as muchas possible. These two groups of relevés werecompared after every modification of the logical

43


definition using the phi coefficient. Finally, the def-inition was selected that was as close to thetraditional subjective delimitation of associationas possible. However, at the same time, the re-quirement that relatively simple definitions aremaintained, without too many criteria, was ap-plied. If the definitions assigned some relevés offorest vegetation to some grassland or heathlandassociations, these relevés were automaticallyexcluded.

In the definitions of associations marginaloverlap was allowed, i.e. some relevés could beassigned to more than one association at thesame time. These relevés were compared withthe groups of relevés that had already beenassigned to each of these associations unam-biguously. The comparison was based on theFrequency-Positive Fidelity Index (FPFI; Tichý2005), which took account of the similarity ofspecies composition between the relevés to beassigned and the respective group of relevés,positively weighting the species with the diag-nostic value for the particular group of relevés.Every relevé subjected to comparison was thenassigned to the group of relevés (association) thatit best matched according to the FPFI index.

Variants, i.e. lower vegetation units inside theassociations, were defined using an unsupervisedclassification method – cluster analysis performedseparately for each group of relevés assigned tothe individual associations. The cluster analysiswas calculated using the PC-ORD 4 program(McCune & Mefford 1999), with the chord distanceas a measure of dissimilarity and the beta-flexi-ble linkage method with the coefficient $ = –0.25.The clusters obtained were interpreted subjective-ly with respect to their ecological interpretation.Generally, two or three (occasionally four) clustersat the highest hierarchical level were interpretedas variants. If the clusters distinguished in thisway showed only indistinctive floristic differentia-tion or did not have an unambiguous ecologicalinterpretation, no variants in the particular asso-ciation were differentiated.

Associations were grouped into alliances andclasses on the basis of the subjective evaluationof their mutual similarity, following the CentralEuropean phytosociological tradition. In somecases, this evaluation was supported by clusteranalysis or ordination of several related associa-tions.

Nomenclature of plantcommunitiesThe nomenclature of plant communities in con-tinental Europe usually adheres to the rulesimplemented in the International Code of Phytoso-ciological Nomenclature (Weber et al. 2000). Alsoin monograph Vegetation of the Czech Republic,the Code was accepted as the nomenclature au-thority. The individual nomenclature solutionspartly follows those adopted in previous nomen-clature revisions (particularly Mucina et al. 1993b,Rennwald 2000, Berg et al. 2004), but predomi-nantly they are based on the independent revisionof the Czech and international syntaxonomic lit-erature. Many vegetation units have been givennames different from those used in the latest listof plant communities of the Czech Republic(Moravec et al. 1995) because this overview con-tains not only correct names but sometimes alsoinvalid names or names that have never been de-scribed.

All names used in the monograph Vegetationof the Czech Republic were checked in the publi-cations containing their original description sothat we could ensure that all the accepted namesare valid. However, it is possible that a different,older and valid name will be found for some ofthe syntaxa in the vast body of phytosociologicalliterature in the future, and this name will have tobe accepted as the correct name.

Apart from the adopted name, which is themost likely correct name, the text also providesfrequent synonyms such as legitimate youngernames of the same syntaxon and illegitimatenames. The list of synonyms is not exhaustive butit is limited particularly to the synonyms usedby Moravec et al. (1995) and in the vegetationmonographs of neighbouring countries such asGermany (Oberdorfer 1993a, b, Pott 1995, Dier-schke 1996 et seq., Rennwald 2000, Schubert etal. 2001), Austria (Mucina et al. 1993b), Poland(Matuszkiewicz 2001), Slovakia (Valachovič et al.1995 et seq., Stanová & Valachovič 2002) andHungary (Borhidi 2003). All synonyms older thanthe respective accepted name include a refer-ence to the article or paragraph of the Code(indicated as §) according to which the name hasto be rejected as being ineffective, invalid or ille-gitimate. This is not included in case of the namesyounger than the accepted one because such

44


names are not usable on the basis of the priorityprinciple. The most frequent reasons of rejectinga name are listed below:

§ 1 – The name has not been published in aprinted publication, it is therefore ren-dered ineffective (nomen ineditum)

§ 2b – The name has not been published witha sufficient original diagnosis (nomennudum). Sufficient original diagnosismeans: (1) in the case of associations,at least one relevé or constancy table ifpublished before 1978 or at least onerelevé if published after 1978; (2) in thecase of the units of a higher rank, a bib-liographically unambiguous reference tothe valid name of the syntaxon of theclosest subordinate principal rank andsince 1980 also with the list of diagnos-tic species

§ 3a – The name has been merely cited as asynonym by its author

§ 3b – The name has been suggested as pro-visional by its author

§ 3c – The syntaxonomic rank of the vegeta-tion unit has not been indicated

§ 3d – The rank of the syntaxon did not corre-spond to the rank of the Code or it is anassociation of the Uppsala School pub-lished before 1936

§ 3e – The syntaxonomic rank did not corre-spond to the form of the name

§ 3f – The name-giving taxa were not indicat-ed in the original diagnosis

§ 3g – The name was published after 1978 andit is not clear from which taxon name(s)it has been formed

§ 5 – The name was published after 1978 with-out indication of the nomenclature type

§ 29 – The syntaxon has been renamed be-cause another taxon characterizes itbetter

§ 31 – The name is a younger homonym, i.e. itis spelled like a previously and validlypublished name

§ 33 – The name is one of the homonyms ofequal age but another of these hom-onyms was adopted by other earlierauthors

§ 34a – The name contains an epithet in thenominative case that indicates a geo-

graphical, ecological or morphologicalproperty, e.g. Fagetum sudeticum orVaccinietum myrtilli subalpinum

§ 34c – The name was formed from more thantwo taxon names

§ 36 – Due to earlier misinterpretations, thename was often used in a false sensethat excludes its type (nomen ambigu-um); it may therefore be proposed forrejection as nomen ambiguum rejicien-dum propositum

§ 37 – The type relevé of the association is soincomplete or complex that it cannot beassigned to any one of the currently dis-tinguished associations (nomen dubium)

§ 43 – The taxon providing the name of thesyntaxon was determined erroneously

Besides the above-mentioned reasons, synony-my may also include the names that are oftenquoted in the literature and attributed to a partic-ular author who neither created nor used thename. Such cases occur surprisingly often andare called phantoms according to Mucina (in Mu-cina et al. 1993a: 19–28).

Another synonymy problem concerns pseu-donyms, i.e. the names of syntaxa used with theoriginal author citation or with reference to it butmisinterpreted by later authors. If these namesare used more frequently, we also place them insynonymy, referring to the misinterpreting author,preceded by the word sensu, and followed by thename of the author of the original description (af-ter the word non). For example, the author citation‘sensu Šmarda 1961 non Tüxen 1937’ means thatŠmarda used Tüxen’s name for a syntaxon otherthan that originally described by Tüxen. If moreauthors used a certain name in a manner differ-ent from that of the original description, theabbreviation auct. non is used instead of the nameof the misinterpreting author.

There are many cases when the correct formof the accepted name differs from that given inthe original diagnosis. As a result, every accept-ed name in the monograph Vegetation of theCzech Republic is accompanied by its originalwording attached after the abbreviation ‘Orig.’,including the original wording of the author cita-tion if this was indicated in the original diagnosis.The accepted names of associations and alliancesthat contained only the genus name(s) in the orig-

45


inal diagnosis were supplemented with speciesepithets in accordance with Recommendation10C of the Code. One exception applies to theassociations and alliances whose original diag-nosis contains more than one species of thegenus used in the name, and it is not thereforeclear from which species name the name of thesyntaxon was formed. In the cases that only oneof these species is indicated in the list of diag-nostic species or has much higher constancy orcover as compared with the other species, theformer species is considered as name-giving andincluded in the name of the syntaxon. In othercases, when it is not clear which is the name-giving species, only the genus name is used andthe list of species that can potentially provide thename is placed in brackets after the original word-ing of the name. The names of classes composedof two taxon names are usually left without spe-cies epithets in accordance with establishedtradition while the names of classes formed fromthe name of a single species are supplementedwith the species epithet.

For practical reasons in some cases, we alsoused the modified form of names which is sub-ject to approval by the Nomenclature Commissionof the International Association for VegetationScience. The form used is considered here as theproposal to modify the name. This concerns nomi-na inversa and nomina mutata. According toArticle 42 of the Code, nomina inversa are thenames of syntaxa in which, as compared with theoriginal diagnosis, the order of the names of taxawas changed so that the dominant taxon or thetaxon of the higher layer is in the second place.According to Article 45, nomina mutata replacesyntaxon names which were originally formedfrom the names of taxa not used in the recenttaxonomic and floristic literature, with syntaxonnames that include the names of taxa that are inaccordance with the contemporary taxonomic lit-erature. The names of taxa that were not acceptedin common taxonomic and nomenclature sourc-es used in the Czech Republic over the last 30years (Ehrendorfer 1973, Smejkal 1981, Neuhäus-lová & Kolbek 1982, Dostál 1982, 1989, Hejný etal. 1988 et seq., Kubát et al. 2002) were generallyreplaced by names of taxa accepted in the Key tothe Flora of the Czech Republic (Kubát et al. 2002).In order to facilitate work with synonyms, we pro-vide the conversion of the old taxon name to the

name from the Key in brackets after the originalform of the syntaxon name. We also introducethis conversion in cases when the name of thesyntaxon maintains a different taxon name thanthat used in the Key.

Phytosociological tables andthe determination of diagnostic,constant and dominant speciesof vegetation units

Species composition of associations defined bythe Cocktail method was compared in synoptictables, which included groups of similar asso-ciations. Each table contains the percentagefrequency of the occurrence of species in relevésassigned to individual associations. Tables do notcontain all available relevés that can be assignedto the particular association; instead, they onlyshow the relevés of the stratified set of 21,794relevés of all vegetation types of the Czech Re-public (see above) that were assigned to individualassociations with the help of Cocktail definitions.The use of relevés from the stratified data setlimited potential distortion of the data due to localoversampling of some sites. If some associationshad less than 10 relevés assigned to them, addi-tional relevés were added such that the number10 was attained, provided they were available andcomplied with the Cocktail definition.

In this stratified data set fidelity of each spe-cies to each association, i.e. the occurrenceconcentration of species in relevés of the particu-lar association, was calculated. Fidelity expressesthe diagnostic value of the species for a particu-lar association. Species with high fidelity can beconsidered diagnostic, i.e. character species ordifferential species. Fidelity was determined withthe phi coefficient, which was used as a measureof the statistical association between the occur-rence of species and the relevés assigned to theparticular phytosociological association. Since thevalue of the phi coefficient depends on the ratioof the number of relevés belonging to the particu-lar association to the total number of relevés, andeach association is represented by a differentnumber of relevés (Chytrý et al. 2002), the relativenumber of relevés of each association was virtuallyequalized to 1% of the total number of all relevés

46


in the stratified data set (Tichý & Chytrý 2006). Forcomparison with the target association, relevésof all vegetation types were retained in this strat-ified data set. As a result, diagnostic speciesdetermined in this manner have general validity incomparison with all other vegetation types of theCzech Republic. Species with a phi coefficientabove 0.25 were considered diagnostic for aparticular association while species with a phicoefficient above 0.50 were termed highly dia-gnostic. These thresholds were determinedsubjectively in order to obtain practical numbersof diagnostic species, i.e. not too many or toofew. After virtual equalization of the number ofrelevés in an association, the phi coefficient mayreach a high value even in cases where fidelity ofa particular species to a particular association isnot statistically significant. This occurs in caseswhen the particular association before equaliza-tion is represented by a small number of relevés.Therefore, in addition to the phi coefficient for eachspecies and association, the statistical signi-ficance of the fidelity prior to equalization wascalculated using the Fisher’s exact test (Chytrý etal. 2002). Based on this calculation, species whoseoccurrence concentration in relevés of the partic-ular association did not differ from random at alevel of significance of P < 0.001 were not includ-ed in a group of diagnostic species although theyshowed a high phi coefficient value. The fidelity ofbryophytes and lichens, which were not recordedin all relevés, was calculated only on the basis ofthe subset of relevés in which these plants wererecorded. Diagnostic species are marked withgreen in synoptic tables while highly diagnosticspecies are marked in a dark green colour. Thesespecies are also introduced in the lists of diag-nostic species in the textual descriptions ofassociations; highly diagnostic species are print-ed in bold.

Diagnostic species determined in the above-mentioned manner were also used to assess thequality of the definition of individual associationsusing the Cocktail method. Associations that hadno diagnostic species, i.e. those difficult to dis-tinguish floristically, were not accepted in theproposed system of vegetation units.

In addition to diagnostic species, species fre-quently occurring in vegetation stands (constantspecies) and species with high cover (dominantspecies) are also important for sufficient charac-

terization of phytosociological associations. Thesespecies were determined using the same dataset with which diagnostic species were deter-mined. Constant or highly constant species werethose with a frequency over 40% or 80%, re-spectively. Dominant species and highly dominantspecies were those that occurred with a covervalue exceeding 25% at least in 5% and 10% ofrelevés, respectively. However, in cases of asso-ciations with few relevés only, species occurringas dominants in a single relevé were not includedin the list of dominant species, even though thissingle dominant occurrence corresponded tomore than 5 or 10% of relevés. In the case ofbryophytes and lichens, constant and dominantspecies were only determined on the basis ofrelevés in which these plants had been recorded.

Diagnostic and constant species of the alli-ances and classes were determined in the sameway as those of the associations, based on therelevés assigned to all the subordinate associa-tions. As these groups of species are only basedon the data from the Czech Republic, they have alocal validity for the national territory. If vegeta-tion diversity across the entire geographic rangeof particular classes and alliances was taken intoaccount, the lists of diagnostic and constant spe-cies would probably be modified to some extent.In alliances with a single association recognizedin the Czech Republic, we consider their diag-nostic and constant species to be identical withthose of the association. The same principle ap-plies for classes with a single alliance. We did notdetermine dominant species of classes and alli-ances, because different associations assignedto them often have different dominant species.

Highly constant and highly dominant speciesare printed in bold in the text. Synoptic tablescontain a list of diagnostic species in the firstpart, followed by a list of species with frequencyof at least 10% in all relevés of a particular tableor at least 20% in one or more associations ofthe table. Less frequent and non-diagnostic spe-cies were omitted due to space limitation.

Graphic calibrationof associations

Ellenberg indicator values (Ellenberg et al. 1992),altitudinal range and the cover of the herb layer

47


for each association were illustrated in box-and-whiskers plots, which provide an overview of thehabitat requirements and the physiognomy of in-dividual associations. These plots were formedon the basis of relevés from the stratified selecti-on, i.e. from the relevés used for making synoptictables. They illustrate the median (i.e. the horizon-tal line in the middle of the box), lower and upperquartiles, i.e. the interval accomodating 50% ofthe observed values (box), and 5% and 95% per-centiles, i.e. the interval containing 90% of theobserved values (whiskers). The background ofeach graph contains the median (the colouredhorizontal line) and the inter-quartile range (thecolour strip) for all associations of the grasslandand heathland vegetation of the Czech Republic,allowing a comparison of the values between in-dividual graphs that use different scales on thevertical axis. Whether the respective variable inthe individual associations has higher, lower ornearly equal values as the other types of grasslandand heathland vegetation can be derived fromthe comparison of the positions of the boxes andcolour strips.

Ellenberg indicator values in the ordinal scaleexpress the relationship of plant species to light,temperature, continentality, humidity, soil reactionand nutrients. This scale contains twelve degreesin the case of humidity and nine degrees for othervariables. Although the applicability and interpre-tations of Ellenberg indicator values have beenrepeatedly challenged in the recent literature(Schaffers & Sýkora 2000, Wamelink et al. 2002),their main advantage is that they allow to com-pare large numbers of relevés with regard toenvironmental factors that cannot usually bedetermined on the basis of short-term measure-ments. The major disadvantage is the ordinalcharacter of Ellenberg indicator values, which lim-its the usability of basic arithmetic operations.However, the values in a data set with a largernumber of species tend to behave like continuousvariables and calculations of arithmetic meansfrom the values of species for a relevé are oftenused as a rough estimate of site conditions (terBraak & Barendregt 1986, Ertsen et al. 1998,Schaffers & Sýkora 2000). Ellenberg values of allrepresented vascular plants were used to calcu-late the unweighted arithmetic mean for eachrelevé of the stratified data set that was assignedby the Cocktail definitions to individual associa-

tions. Species which were lacking or not assigneda particular indicator value in the Ellenberg tableswere omitted. Thus we obtained indicator valuesfor each relevé and illustrated their distribution inthe box-and-whiskers plots.

Altitudes were taken directly from the accom-panying relevé data. If there was no indication ofaltitude, this was derived from the digital hypso-metric map in the geographical informationsystem ArcGIS 8.3 (www.esri.com).

Data on percent cover of the herb layer werealso taken from the relevés. Relevés that did notcontain this information were not used in thegraphical presentation.

Distribution mapsof associations*

The distribution of individual associations wasmapped in a geographical grid with cells of 5 min-utes of geographical longitude × 3 minutes oflatitude, i.e. approximately 6 × 5.5 km. The gridwas derived from the standard grid of the CentralEuropean mapping of flora and fauna, with thebasic cells divided into quadrants.

The source data used to make maps includ-ed all relevés of non-forest vegetation that werecontained in the Czech National Phytosociologi-cal Database by 15 December 2005 and localizedusing geographical coordinates with an accura-cy greater than 1 geographical minute. The totalnumber of relevés used was 51,940. The relevésfrom this data set were compared with the for-mal definitions of associations formed by theCocktail method. They were compared not onlywith the definitions of associations of grasslandand heathland vegetation but also with defini-tions of associations of other treeless vegetationtypes, such as mires, reed beds, tall-sedge veg-etation and chasmophytic vegetation, which willbe described in the next volumes of the mono-graph Vegetation of the Czech Republic. Thisparallel comparison enabled to identify all over-laps in the association delimitations. If any relevéwas assigned to more than one association, itsfinal assignment was decided on the basis ofthe similarity calculation using the FPFI index(Tichý 2005) and the relevé was assigned to the

*Elaborated by M. Chytrý, K. Kubošová & O. Hájek.

48


association whose species composition it bestmatched.

The distribution maps prepared on the basisof available relevés used different symbols forthose sites where only old relevés were available(recorded up to 1975) and sites with relevés re-corded after this date. The maps provide a reliableoverview of distribution in the case of rare asso-ciations, for which the majority of sites are wellknown and phytosociologically documented (e.g.associations of alpine vegetation). They are alsoreasonably reliable for associations that are abun-dant but represent vegetation types which havebeen popular among phytosociologists and ex-tensively sampled in the Czech Republic (e.g. drygrasslands). However, in the case of some wide-spread associations (e.g. meadows), the maps ofexisting relevé localities provide an incompletepicture of the real distribution. We believe thatpublication of these incomplete maps may stimu-late further research aimed at filling the gaps inthe presently known distribution. In future thesemaps will be regularly updated with the use ofnew relevés obtained for the Czech National Phy-tosociological Database and published online.

To provide further information about the dis-tribution of the associations, the maps of relevélocalities of some of the associations were sup-plemented with the estimate of their potentialdistribution. The estimate was based on the sta-tistical predictive model which quantified therelationship between the occurrence probabilityof a particular association and the explanatoryenvironmental variables that were available in theform of digital maps for the territory of the CzechRepublic. The models used the following explan-atory variables: altitude, soil acidity, averagetemperature (annual, January and June) and an-nual precipitation. The values of these variableswere obtained for each relevé by the overlayingof the respective digital maps with geographicalcoordinates of the relevés. The overlay was per-formed in the geographical information systemArcGIS 8.3 (www.esri.com). In order to minimizethe effect of the local oversampling of certainareas, the stratified set of 21,794 relevés (seeabove) was used for modelling. From this data set,the relevés complying with the Cocktail definitionof each association were selected for modelling.

Since the dependent variable, i.e. the pres-ence or absence of the association at the

particular site, is binary, predictive modelling usedthe generalized linear model (GLM) with binomialdistribution and the logistic linking function logit.In cases when the binary dependent variable didnot have a binomial distribution, the quasi distri-bution was used (McCullagh & Searle 2001).Regression coefficients were estimated using themaximum likelihood method and their significancewas tested using the likelihood ratio test. The se-lection of variables proceeded from the full modelwith all explanatory variables. The variables shownto be significant and to contribute to the increasedpredictive ability of the model were selected step-wise. In this process, each association obtainedan equation with the estimated regression coeffi-cients; this equation was used to predict theprobability of occurrence at the sites with no dataavailable. The probability of the occurrence of aparticular association in a range of <0, 1> wasobtained using the inverse logistic transformationof the linear predictor values calculated from re-gression equations.

The generalized index R2N (generalized

R-square) was used (Cox & Snell 1989) to verifythe predictive abilities of the models. If the R2

N

value was higher than 0.8, the model was consid-ered highly predictive. Equations with an R2

N value

of 0.7 to 0.8 were also accepted, showing slightlylower but still good predictive ability. The predic-tive abilities of models were low particularly incases when available relevés did not provide arepresentative coverage of the range of ecologicalfactors linked to the occurrence of the associa-tion in question. Associations with less than 50available relevés were not modeled at all; similar-ly, some models based on a larger number ofrelevés but providing meaningless predictionsaccording to the expert judgement were alsorejected. For some associations, two to three al-ternative models were developed, of which themodel that seemed to best reflect the biologicalreality was subjectively selected.

Probabilities of occurrence, calculated fromindividual models, were plotted on grid mapsprovided they were higher than a subjectively se-lected threshold. However, they were only plottedat sites where grasslands or heathlands currentlyoccur according to the digital map of CORINEland cover. In such a way the observed distribu-tion was supplemented with the estimate of thepotential distribution. The symbols used in the

49


maps are as follows: sites with relevés record-ed after 1975; sites with no relevés recordedafter 1975 but with relevés recorded earlier; ● siteswith no relevés but with a high probability of oc-currence of the association according to thepredictive model.

On the practical applicationof the present phytosociologicalsystem

Vegetation can be classified in many differentways and the system presented here is just oneof them. Its main advantage is that it is supportedby the analysis of phytosociological data andprovides unambiguous criteria for inclusion ofparticular vegetation stands or relevés in associ-ations. It does not aspire to classify every existingstand of vegetation but defines the cores of as-sociations, which are usually characterized by theoccurrence of ecologically specialized species.This reflects the common experience that phy-tosociological systems work well with relativelyhomogeneous stands that contain specific com-binations of species with a narrow ecologicalrange, but at the same time leave a large pro-portion of vegetation stands existing on thelandscapes unclassified. Still it is possible toquantify the similarity of any vegetation stand tothe cores of the associations and assign it to themost similar association, if necessary.

The identification of associations included inthe proposed classification can be performedusing the computer expert system available atwww.sci.muni.cz/botany/vegsci/vegetace.php.This expert system runs in the environment of theJUICE program and uses relevés exported fromthe TURBOVEG program as input data. In orderto ensure the correct function of the expert sys-tem, relevés intended for automatic assignmentto associations should be exclusively from non-forest vegetation. Formal definitions of grasslandand heathland associations prepared with theCocktail method and contained in the expert sys-tem were formed with an assumption that theywill not be used for classification of forest vege-tation (if they were, they could assign some forestrelevés to non-forest associations, e.g. some drypine forest relevés from acidic soils to heathland

associations). Sometimes different relevés from asingle relatively homogeneous vegetation standare assigned to different associations by the ex-pert system. Such stands should be interpretedas transitional between these associations. If theexpert system assigns some of the relevés from asingle vegetation stand to a particular associationwhile others remain unassigned to any associa-tion, it means that the stand consists of patcheswith typical and less typical species compositionwith respect to the association core. Relevés thatremained unassigned to any association can becompared by the expert system using the FPFIindex with the total species composition of indi-vidual associations and subsequently assignedto the association that they best match. Thiskind of assignment can be interpreted as follows:although the relevé does not belong to the coreof the particular association and is not typicalof it, it is close or similar to it. In addition to therequirement that the relevé should match the par-ticular association better than any other, it is alsoappropriate to determine the particular thresholdvalue of similarity that the relevé must exceed inorder to be assigned to the particular associa-tion. A large number of stands exist that are mainlycomposed of species with a broad ecologicalrange or which contain unusual species combi-nations whose assignment to any associationwould be in conflict with phytosociological tradi-tion. The determination of the threshhold value issubjective and depends on the user, who mustdecide how large deviations from the typicalspecies composition he or she is willing to ac-cept while assigning the relevé to a particularassociation.

With respect to the practical use of vegetationclassification, one should realize that applicabilityof any classification is scale-dependent. The pro-posed classification was optimized for the territoryof the Czech Republic. It is therefore possible thatsome associations that have been distinguished init will not be clearly recognizable in broader Cen-tral European or European vegetation classificationsystems. However, it is also evident that in the caseof strictly local vegetation description, e.g. in asmall nature reserve, it may be more suitable todefine ad hoc local vegetation units. Such unitsmay be difficult to transpose into other territoriesor to larger scales but will provide a better de-scription of local vegetation variability. Even so, it

50


may be desirable to compare such local vegeta-tion units with the national classification, e.g. bymeans of the assignment of relevés to associa-tions by the expert system, and thus place localdiversity patterns into a broader context.

Acknowledgments

This project would never have been possible with-out the hundreds of enthusiastic Czech botanistswho have enjoyed and continue to enjoy crawlingalong on their knees within the restricted areas oftheir relevés, searching for further plant speciesamong blades of grass and tiny leaves in standsof vegetation. The enormous amount of field workperformed over the last 80 years by these enthu-siasts has made the Czech Republic one of thebest explored countries in the world in terms ofvegetation diversity.

The methodology of the project was estab-lished thanks to the support and inspiration offriends from the European Vegetation Surveyworking group. John S. Rodwell and Julian Dring(Lancaster, UK) shared their experience in themanagement of phytosociological databases withus at the beginning of the project. Stephan M.Hennekens (Wageningen, NL) provided us withthe TURBOVEG program and was always willingto improve and modify this program accordingto our requirements, which were usually verydemanding. Harald Niklfeld, Walter Gutermann (Vi-enna, AT), Ladislav Mucina (Stellenbosch, ZA),Milan Valachovič and Ivan Jarolímek (Bratislava,

SK) cooperated with us on the development of astandard list of Central European plant speciesfor the TURBOVEG database. Helge Bruelheide(Halle, DE) inspired us with his ideas to formalizethe methods of traditional phytosociology andwas always willing to discuss with us on the de-velopment and application of these new methods.Jason Holt (Brno, now Hinsdale, Montana, USA)and Zoltán Botta-Dukát (Vácrátót, HU) contribut-ed to the development of the concept of statisticalmeasurement of species fidelity. Jean-Paul Theu-rillat (Champex, CH) and Valentin Golub (Togliatti,RU) gave us advice on the syntaxonomic no-menclature and diversity of saline vegetation,respectively. Ms. Iva Adamová, the librarian at theformer Department of Botany and now of theDepartment of Botany and Zoology, MasarykUniversity, helped us obtain difficult-to-access lit-erature. Toby Spribille (Göttingen, DE) kindly madea thorough linguistic revision of the English textsin this book. Photographs of some vegetation ty-pes were provided, besides the authors of thetreatments of individual syntaxa, by Vít Grulich,Viera Horáková, Záboj Hrázský, Jitka Kopáčová,Zdenka Otýpková, Jan Roleček, Jan Vaněk, Jiří Vi-cherek and Alena Vydrová. Many valuablecomments on the previous version of the text wereprovided by the reviewer of this book, LadislavMucina (Stellenbosch). Last but not least, we wantto thank all of the colleagues who supplied thenational phytosociological database with relevésand cooperated in database compilation. Thesecontributors are listed in the Czech version of thistext (page 17).

Czech-English glossaryof basic keywords

alpínský, -á, -é alpineasociace associationBílé Karpaty White Carpathiansbiotop, -y habitat, -sbohatý, -á, -é richbor, -y pine forest, -sbučina, -y beech forest, -sbylina, -y herb, -sbylinný, -á, -é herbaceousČechy BohemiaČeská republika Czech Republic

Českomoravská Bohemian-Moravianvrchovina Uplands

český, -á, -é 1. Czech; 2. Bohemianchladný, -á, -é coolchudý, -á, -é poordiagnostický, -á, -é diagnosticdominantní dominant, dominatingdoubrava, -y oak forest, -sdřevina, -y woody plant, -sdruh, -y speciesdruhová bohatost species richness

51


druhové složení species compositiondubohabřina, -y oak-hornbeam forest, -sdynamika dynamicsflyš flyschfytocenologické relevés

snímkyfytocenologický relevé

snímekhadec, hadce serpentine, -shluboký, -á, -é deephojný, -á, -é commonhora, -y mountain, -shorský, -á, -é montanehospodářský význam economic importancejednoletá rostlina annual plantjehličnatý, -á, -é coniferousjižní southernkarpatský, -á, -é CarpathianKarpaty Carpathianskeř, -e shrub, -s (subst.)keřový, -á, -é shrub (adj.)keříček, keříčky low shrub, -s (subst.)keříčkový, -á, -é low-shrub (adj.)klasifikace classificationkonstantní constantkontinentální continentalkras karstKrkonoše Giant Mountainskřovina, -y scrub, shrubbery (subst.)křovinný, -á, -é shrub, shrubby (adj.)kyselý, -á, -é acid, acidicLabe Elbeles), -y forest, -s, woodland,

-s (subst.)lesní lem, -y forest fringe (saum)lesní forest, woodland (adj.)lišejník, -y lichen, -slouka, -y meadow, -služní riverine, floodplain (adj.)Maďarsko Hungarymalý, -á, -é smallmech, -y moss, -esmechorost, -y bryophyte, -smělký, -á, -é shallowměsto, -a town, -smírný, -á, -é gentle, moderateMorava Moraviamoravský Moraviannadmořská výška, -y altitude, -snarušovaný, -á, -é disturbedNěmecko Germany

nížina, -y lowland, -snízký, -á, -é lowoceanický, -á, -é oceanicohrožení endangermentolšina, -y alder forest, -sopadavý, -á, -é deciduousopuštěný, -á, -é abandonedorná půda arable landpahorkatina, -y hilly (colline) landscape, -spanonský, -á, -é Pannonianpaseka, -y forest clearing, -spastvina, -y pasture, -spatro, -a layer, -spísčina, -y sand area, -spísečný, -á, -é sand, sandy (adj.)písek, písky sand, -splevel, -e weed, -spočet druhů number of speciespodhorský, -á, -é submontanepokryvnost, -i cover, -spole field, -sPolsko Polandporost, -y stand, -spotok, -y brook, -sPraha Pragueprameniště water spring, -spřevážně mainly, mostlypřirozený, -á, -é naturalprůměrný, -á, -é averagepůda, -y soil, -sřád, -y order, -srákosina, -y reed bed, -sRakousko Austriarašeliniště mire, -sřeka, -y river, -srostlina, -y plant, -srostlinná společenstva plant communitiesrostlinné společenstvo plant communityrozšíření distributionrula, -y gneiss, -esrybník, -y fishpond, -ssečený, -á, -é mownsešlapávaný, -á, -é trampledseverní northernširokolistý, -á, -é broad-leavedskála, -y rock, -s (subst.)skalní rock, rocky (adj.)skupina, -y group, -s

(abbrev. skup.)slanisko, -a salt marsh, -esslatina, -y fen, -s

52


sloupec, sloupce columns, -sSlovensko Slovakiasmrčina, -y spruce forest, -sspolečenstvo, -a community, -iessrážky precipitationstanoviště habitat, -sstrmý, -á, -é steepstojatá voda, -y standing waterstrom, -y tree, -s (subst.)stromový, -á, -é tree (adj.)střední centralstruktura, -y structure, -ssubalpínský, -á, -é subalpinesuchý, -á, -é dryŠumava Bohemian Forestsuť, sutě scree, -ssvah, -y slope, -ssvaz, -y alliance, -stabulka, -y table, -stekoucí voda, -y running waterteplomilný, -á, -é thermophilousteplota, -y temperatureteplý, -á, -é warmtráva, -y grass, -estravina, -y graminoid, -stravinný, -á, -é grassland (adj.)trávník, -y grassland, -s (subst.)

třída, -y class, -esúdolí valley, -súzemí territory, -iesúzkolistý, -á, -é narrow-leavedvápenec, vápence limestone, -svápnitý, -á, -é calcareousvarianta, -y variant, -svegetace vegetationvelký, -á, -é large, bigvesnice village, -svlhkost moisturevlhký, -á, -é wetvoda, -y water, water bodiesvodní aquaticvrchoviště bog, -svřesoviště heathland, -svýchodní easternvýchoz, -y outcrop, -svysokobylinný, -á, -é tall-forbvysoký, -á, -é tallvytrvalá rostlina perennial plantvzácný, -á, -é rarezamokřený, -á, -é water-loggedzápadní westernzaplavovaný, -á floodedživina, -y nutrient, -sžula, -y granite, -s

Documents

Vegetace CeskÈ republiky» 1 Travinn· a ke¯ÌËkov· vegetace · ACADEMIA PRAHA 2007 Vegetation of the Czech Republic Travinn· a ke¯ÌËkov· vegetace 1. Grassland and Heathland