Variabilita luk svazu Molinion v České a Slovenské republice

MASARYKOVA UNIVERZITA

Prodovdeck fakulta

DISERTAN PRCE Marcela EZNKOV Brno, 2006

MASARYKOVA UNIVERZITA Prodovdeck fakulta

Marcela EZNKOV

VARIABILITA LUK SVAZU MOLINION V ESK A SLOVENSK REPUBLICE

Disertan prce

kolitel: Doc. RNDr. Milan Chytr, Ph.D. Brno, 2006

Bibliografick identifikace Jmno a pjmen autora: Marcela eznkov

Nzev disertan prce: Variabilita luk svazu Molinion v esk a Slovensk republice

Nzev disertan prce anglicky: Variability of the Molinion meadows in the Czech and Slovak

Republics

Studijn program: Biologie

Studijn obor (smr), kombinace obor: Botanika

kolitel: doc. RNDr. Milan Chytr, Ph.D.

Rok obhajoby: 2006

Klov slova v etin: esk republika, fytocenologie, klasifikace, lun vegetace, ordinace,

Slovensk republika

Klov slova v anglitin: Czech Republic, phytosociology, classification, meadow vegetation,

ordination, Slovak Republic

Marcela eznkov, Masarykova univerzita, 2006

Podkovn Na tomto mst bych rda podkovala pedevm mmu koliteli doc. RNDr. Milanu Chytrmu, Ph.D. za veden disertan prce, za jeho trplivost a ochotu pomoci po celou dobu jejho zpracovvn. Rovn moc dkuji manelm Hjkovm, Honzovi Rolekovi a doc. RNDr. Vtu Grulichovi, CSc. za podntn pipomnky a za poskytnut daj o lokalitch bezkolencovch luk. Mj dk pat i vem ostatnm lidem, kte mi pi zpracovn disertan prce jakkoli pomohli. Publikace v disertan prci byly pipraveny s podporou tchto projekt: GA R 206/02/0957, GA R 206/05/0020, GD R 524/05/H536, APVT-51-015804, MSM 143100010, MSM 0021622416.

Abstrakt Disertan prce je zamena zejmna na studium variability bezkolencovch luk svazu Molinion na zem esk republiky a Slovenska. Zabv se vak rovn variabilitou cel tdy Molinio-Arrhenatheretea a variabilitou druhovho sloen vlhkch aluvilnch luk podl klimatickho gradientu kontinentality ve stedn Evrop. Pomoc detrendovan korespondenn analzy bylo zjitno, e nejvznamnjm faktorem, kter ovlivuje variabilitu druhovho sloen lun vegetace tdy Molinio-Arrhenatheretea v esk republice, je vlhkost, druhm nejdleitjm faktorem jsou iviny, s nimi koreluje pH pdy. Shlukovou analzou bylo v tto vegetaci rozlieno deset hlavnch vegetanch typ. Prvn skupina sdruovala ninn aluviln louky svazu Deschampsion, skupiny 2, 3, 5, 9 a 10 rzn typy vlhkch pchovch luk svazu Calthion, skupina 4 louky svazu Molinion, skupina 6 zahrnovala horsk pchov louky, kter pedstavuj pechodn typ vegetace mezi svazy Calthion a Polygono-Trisetion, skupina 7 ninn ovskov louky svazu Arrhenatherion a skupina 8 ovskov louky vych poloh na ivinami chudch pdch, horsk trojttov louky svazu Polygono-Trisetion a mezofiln pastviny svazu Cynosurion. Analzou beta-diverzity bylo zjitno, e v lun vegetaci na vlhkch stanovitch jsou vt rozdly v druhovm sloen mezi jednotlivmi typy vegetace, co odpovdalo jednak vsledkm shlukov analzy, kde vlhk pchov louky byly rozdleny ve srovnn s mezofilnmi loukami do vce skupin, a jednak lenn v tradin fytocenologii, kde je ve svazu Calthion rozliovn nejvt poet asociac.

Vlhk ninn louky (vyskytujc se do 350 m n. m.) z esk republiky, vchodnho Rakouska, Slovenska, Maarska a severnho Chorvatska byly pouitm nov metody na uren optimlnho potu shluk rozlenny shlukovou analzou nejprve na ti shluky, z nich prvn zahrnoval pchov louky svazu Calthion, druh bezkolencov louky svazu Molinion a tet kontinentln zaplavovan louky, azen do svaz Agrostion albae, Alopecurion pratensis, Cnidion venosi, Deschampsion cespitosae a Veronico longifoliae-Lysimachion vulgaris. V dalm kroku byl soubor rozdlen na devt shluk, z nich tyi zahrnovaly kontinentln zaplavovan louky, ti pchov louky a dva shluky odpovdaly bezkolencovm loukm. Protoe dal dlen kontinentlnch zaplavovanch luk neodrelo hranice mezi svazy tak, jak jsou uvdny v tradin fytocenologick literatue, bylo navreno slouit tyto svazy do jednoho a pojmenovat je nejstarm platnm jmnem, tj. Deschampsion cespitosae Horvati 1930.

Protoe bezkolencov louky svazu Molinion byly v obou tchto studich jasn odlieny od ostatn lun vegetace, lze je povaovat za dobe vyhrann vegetan typ. V esk republice byly tyto louky v dosavadn literatue lenny do deseti fytocenologickch asociac, na Slovensku do dvancti. Ne vechny tyto asociace vak byly jasn popsny, mnoh z nich byly rozlieny pouze na zklad vy pokryvnosti nebo ptomnosti jednoho nebo dvou druh. Z tohoto dvodu bylo druhov sloen i diagnostick druhy ady asociac velmi podobn a pro uivatele fytocenologick klasifikace bylo asto obtn piadit k nim fytocenologick snmky, nebo dan snmek svm druhovm sloenm zpravidla odpovdal nkolika z nich. Po proveden syntaxonomick revize byly na zem esk republiky a Slovenska rozlieny pouze dv asociace. Asociace Molinietum caeruleae Koch 1926, kter se vyskytuje na bzemi bohatch stanovitch a je charakterizovna druhy Betonica officinalis, Galium boreale nebo Serratula tinctoria, a asociace Junco effusi-Molinietum caeruleae Txen 1954, vzan na kysel, oligotrofn pdy s velmi nzkm obsahem vpnku a typick vskytem druh smilkovch trvnk. Podle Ellenbergovch indikanch hodnot, kter byly pouity pro interpretaci hlavnch gradient, koreloval hlavn gradient tchto luk pozitivn s pdn reakc, kontinentalitou, teplotou a ivinami. Klasifikace pro eskou republiku byla dlna metodou Cocktail, pro Slovensko shlukovou analzou.

Dissertation Abstract The main topic of the Ph.D. thesis is variability of meadows of the Molinion caeruleae alliance in the Czech and Slovak Republics. Variability of the Molinio-Arrhenatheretea class in the Czech Republic and variability of lowland wet meadows along a climatic continentality gradient in Central Europe is also studied. Detrended correspondence analysis of meadows of the Molinio-Arrhenatheretea class from the Czech Republic revealed moisture as the main gradient, and nutrients, correlated with soil base status, as a secondary gradient controlling species composition. Ten clusters were distinguished in this vegetation using cluster analysis. Lowland alluvial meadows of the Deschampsion alliance were included in the first cluster, intermittently wet meadows of the Molinion alliance in cluster 4, montane Cirsium meadows, which represent transitional type of vegetation between the Calthion and the Polygono-Trisetion alliances, in cluster 6, nutrient-rich types of Arrhenatheretum meadows from lower altitudes in cluster 7 and nutrient-poor types of Arrhenatheretum meadows from higher altitudes, montane meadows of the Polygono-Trisetion alliance and mesic pastures of the Cynosurion alliance in cluster 8. Wet meadows of the Calthion alliance were divided among several clusters, which corresponded to the main associations recognized in traditional phytosociological literature. This patterns suggests that wet meadows have a higher beta-diversity than mesic meadows. This hypothesis was tested by calculating mean pair-wise Srensen dissimilarity of meadow relevs for partitions of the moisture gradient, and confirmed that beta-diversity of meadows increases with increasing soil moisture. In traditional phytosociological literature, this fact is reflected by higher numbers of associations distinguished within wet meadows than in mesic meadows.

A statistical analysis of variation in species composition of Central European lowland wet meadows from the Czech Republic, eastern Austria, Slovakia, Hungary and northeastern Croatia was performed, using a data set of 387 geographically stratified vegetation relevs sampled at altitudes < 350 m. Principal coordinates analysis was used to identify and partial out the noise component in the variation in this data set. The relevs were classified by cluster analysis. A new method for identifying the optimal number of clusters was developed, based on species fidelity to particular clusters. This method suggested the optimum level of classification with three clusters and secondary optimum levels with five and nine clusters. Classification based on three clusters separated the traditional phytosociological alliances of Calthion and Molinion, both with a suboceanic phytogeographical affinity, and a group of flooded meadows of large river alluvia, with a continental affinity. The latter group included the traditional alliances of Agrostion albae, Alopecurion pratensis, Cnidion venosi,Deschampsion cespitosae and Veronico longifoliae-Lysimachion vulgaris. However, the internal heterogeneity of this group did not reflect putative boundaries between these alliances as proposed in the phytosociological literature. Therefore it was suggested to unite these alliances in a single alliance Deschampsion cespitosae Horvati 1930 (the oldest valid name).

As meadows of the Molinion alliance were distinguished in both of these studies, it is possible to consider them as a clearly defined vegetation type. Ten phytosociological associations were reported in the previous phytosociological literature for this alliance from the Czech Republic and twelve associations from Slovakia. However, most of these associations were not clearly described. They were often distinguished only on the basis of higher cover or presence of one or two species. Therefore species composition and diagnostic species of these associations were often very similar. It was difficult to assign phytosociological relevs to them because one relev may have corresponded with its species composition to several associations. After a syntaxonomical revision only two associations

were distinguished in the Molinion alliance in the Czech Republic and also in Slovakia Molinietum caeruleae Koch 1926 and Junco effusi-Molinietum caeruleae Txen 1954. The former occurs on more base-rich soils and is characterized by species such as Betonica officinalis, Galium boreale or Serratula tinctoria, while the latter is found in more acidic habitats and is typical by the occurrence of species of Nardus grasslands. Ellenberg indicator values were used for interpretation of main gradients of these meadows. In both countries the main gradient positively correlates with soil base status, continentality, temperature and nutrients.

Obsah 1. vod...10 2. Souasn stav poznn vegetace bezkolencovch luk12

2.1. Pvod 12 2.2. Pdy...13 2.3. Kolsn hladiny podzemn vody...13 2.4. Obhospodaovn...15 2.5. Souasn stav a rozen...15 2.6. Syntaxonomick lenn16 2.7. Charakteristika druh typickch pro stdav vlhk stanovit.17

3. Cle disertan prce18 4. Struktura prce19

4.1. Vymezen podlu spoluautor20

5. Hlavn vsledky prce21 6. Literatura23 7. Publikovan nebo k tisku pipraven texty27

7.1. Havlov M., Chytr M. & Tich L. (2004): Diversity of hay meadows

in the Czech Republic: major types and environmental gradients. Phytocoenologia 34: 551567..28

7.2. Botta-Dukt Z., Chytr M., Hjkov P. & Havlov M. (2005):

Vegetation of lowland wet meadows along a climatic continentality gradient in Central Europe. Preslia 77: 8911146

7.3. Havlov M. (2006): Syntaxonomical revision of the Molinion meadows

in the Czech Republic. Preslia 78: 87101...70 7.4. eznkov M. (2007): Svaz TDD. Molinion caeruleae. In: Chytr

M. (ed.): Vegetace esk republiky 1. Travinn a kekov vegetace. Academia, Praha (pijato k tisku) ...86

7.5. eznkov M. (2007): Variability of the Molinion meadows in

Slovakia. Biologia (pijato k tisku) 101 8. Publikace..120

1. vod

Louky tdy Molinio-Arrhenatheretea pedstavuj v cel stedn Evrop ohroen typ vegetace. Zemdlsky vyuvan lun porosty jsou kvli zskn co nejvtch vnos intenzivn hnojeny a koseny nkolikrt ron. Na druh stran znan st lunch porost je ponechna ladem a postupn zarstaj (Ellenberg 1996, Linusson et al. 1998, Dupr & Diekmann 2001, Jensen et al. 2001).

Vtina fytocenologickch klasifikac lun vegetace tdy Molinio-Arrhenatheretea ve stedn Evrop je zaloena na expertnm sudku vdeckch pracovnk (Oberdorfer 1993, Ellmauer & Mucina 1993, Ellmauer 1994, Dierschke 1995, 1997, Blakov & Baltov in Moravec et al. 1995, Zuidhoff et al. 1996, Kuera & umberov 2001). Stejnm zpsobem byly odhadnuty i hlavn gradienty prosted, kter nejvznamnji ovlivuj variabilitu druhovho sloen (Ellenberg 1996). Jednm z cl disertan prce proto bylo ovit, zda stejn vegetan typy a stejn hlavn gradienty prosted budou rozlieny i s pouitm numerickch metod. Rovn jsme chtli otestovat, zda beta-diverzita vlhkch luk je vy ne beta-diverzita luk mezofilnch, co by umonilo vysvtlit, pro jsou vlhk louky lenny v tradin fytocenologii do vce asociac.

Kontinentln ninn vlhk louky jsou do jist mry svm druhovm sloenm podobn bezkolencovm loukm svazu Molinion a je dleit mezi nimi a vegetac svazu Molinion vymezit hranici. Kontinentln ninn vlhk louky jsou vak v rznch zemch stedn Evropy azeny do rznch svaz Agrostion albae, Alopecurion pratensis, Cnidion venosi, Deschampsion cespitosae a Veronico longifoliae-Lysimachion vulgaris (Passarge 1964, pnikov 1975, Oberdorfer 1993, Ellmauer & Mucina 1993, Dierschke 1995, Kuera & umberov 2001, Stanov & Valachovi 2002, Borhidi 2003, Burkart et al. 2004) a nen zcela jasn, jak se tyto svazy od sebe li. Druhov sloen snmk azench do tchto svaz je toti asto velmi podobn, stejn jako stanovit, na kterch se vyskytuj. Otzkou je, zda je opodstatnn vechny tyto svazy rozliovat a zda by nebylo vhodnj rozliovat pouze jeden svaz s nkolika asociacemi. Clem tedy bylo pouitm rozshl fytocenologick databze snmk ninnch vlhkch luk (z lokalit do 350 m n. m.) z esk republiky, vchodnho Rakouska, Slovenska, Maarska a severnho Chorvatska vyhodnotit variabilitu v druhovm sloen tchto luk, numerickmi metodami identifikovat hlavn vegetan typy a porovnat je s vegetanmi typy rozliovanmi v tradin fytocenologii.

Bezkolencov louky svazu Molinion pat v rmci tdy Molinio-Arrhenatheretea k jednomu z nejohroenjch typ vegetace. Jedn se toti o louky, kter byly v minulosti extenzivn obhospodaovny a poskytovaly pouze nzk vnosy (Kopeck 1960, Kovcs 1962, Zahradnkov-Roetzk 1965). Z dvodu intenzifikace zemdlstv a stle vt poteby zemdlsk pdy v prbhu 19. a 20. stolet byla vtina z nich postupn odvodnna a rozorna, ppadn pemnna na kulturn louky. Jet z padestch a edestch let minulho stolet je tato vegetace dokladovna z ady lokalit v esk republice (Klika 1947, Vlek 1954, 1956, Kopeck 1960) i na Slovensku (Klika 1958, marda 1951, Zahradnkov-Roetzk 1965, pnikov 1971, Boskov 1970, 1975). Bhem druh poloviny dvactho stolet vak byly i tyto lokality vtinou znieny a do souasn doby se zachovaly pouze fragmenty, nap. v nkterch chrnnch zemch (NPR Polabsk ernava, NPR Hrabanovsk ernava, PR Abrod), roztrouen na lesnch loukch, okrajch rybnk a zejmna ve vojenskch prostorech (nap. VVP Ddice nebo V Boletice).

Ve svazu Molinion bylo na zem esk republiky a Slovenska rozlieno celkem trnct asociac (Blakov & Baltov-Tulkov in Moravec et al. 1995, Baltov-Tulkov 1996, 2001). Vtinou vak byly tyto asociace popisovny na zklad vt pokryvnosti nebo ptomnosti jednoho nebo dvou druh, a tak druhov sloen a rovn i diagnostick druhy jednotlivch asociac byly velmi podobn. Dalm clem disertan prce proto bylo provst

10

syntaxonomickou revizi, na zklad kter by byly s pouitm numerickch metod vylenny jasn definovan asociace s vlastnmi diagnostickmi druhy.

11

2. Souasn stav poznn vegetace bezkolencovch luk 2.1. Pvod

Vtina bezkolencovch luk pravdpodobn vznikla a v souvislosti s innost lovka na mstech pvodnch lesnch porost. Jak lze v souasn dob na ad lokalit pozorovat, bez managementu tyto louky postupn zarstaj a v konen fzi je nahrad lesn spoleenstva. Na druhou stranu se pedpokld, e tyto louky vznikaly i bez psoben lovka, a to jako jedno ze sukcesnch stadi pi zazemovn pirozench vodnch ndr (Krippel 1965, Pacltov & Huben 1994). Tento zpsob vzniku bezkolencovch luk byl zejm mon i v dob ped vraznjm psobenm lovka na vegetaci. Dokzat existenci bezkolencovch luk v tchto obdobch je vak obtn, nebo na zklad pylovch analz nen mon vtinu druh charakteristickch pro svaz Molinion identifikovat.

Krippel (1965), kter se zabval postglacilnm vvojem vegetace na Zhorsk nin, se domnv, e bezkolencov louky mohly existovat u v obdob mladho dryasu. Pedpokld, e zde mohly vznikat jednak pi zazemovn mrtvch ramen eky Moravy a jednak v mezidunovch sneninch, kde na povrch vystupovaly prameny nebo podzemn voda a vytvely mezi psenmi pesypy mal jezrka. Tyto vodn ndre byly nejprve osidlovny okehkovitmi rostlinami a vodnmi makrofyty, kter postupn nahradily spoleenstva rkosin nebo vysokch ostic. Dalm hromadnm rkosov a osticov slatiny se postupn sniovala hladina vody a spoleenstva rkosin a vysokch ostic nahradily na stanovitch bohatch vpnkem kalcitrofn nebo mezotrofn slatinn raelinit, na stanovitch s nzkm obsahem vpnku pak pechodov raelinit nebo vrchovit. Pokud se i nadle sniovala hladina podzemn vody nebo dochzelo k jejmu vraznjmu kolsn v prbhu vegetan sezony, druhy charakteristick pro vegetaci raelini postupn ustupovaly a zaaly se uplatovat druhy typick pro stdav vlhk stanovit, jako je Molinia caerulea agg., Succisa pratensis nebo Selinum carvifolia. Na stanovitch bohatch vpnkem tak mohla vzniknout vegetace bazifilnch bezkolencovch luk, na mstech chudch vpnkem vegetace acidofilnch bezkolencovch luk.

Podobnm zpsobem pravdpodobn vznikaly bezkolencov louky i na jinch mstech. Pkladem mohou bt polabsk ernavy, u nich se pedpokld, e vznikly postupnm zazemovnm prohlubn (jezer), kter se vytvoily koncem pleistocnu a potkem holocnu, kdy se v polabsk nin ukldaly vt psky a vytvely hrze nebo naopak prohlubn zaplovan vodou (Pacltov & Huben 1994).

Spoleenstva bezkolencovch luk mohla na takovch mstech existovat adu let. V tto vegetaci se vak ji mohly uplatovat i semenky ke a strom a je pravdpodobn, e za uritou dobu byly bezkolencov louky nahrazeny lesnmi spoleenstvy. K jejich zachovn by bylo nutn alespo obasn seen, kter by zamezilo en ke a strom. Bez existence managementu se bezkolencov louky pravdpodobn udrovaly pouze maloplon, nap. na such okrajch raelini. Na zamezen nletu devin se mohla podlet i pastva zve.

Na ad mst vak souvis vznik bezkolencovch luk a s innost lovka, kter postupn likvidoval pirozen lesn porosty. Bazifiln bezkolencov louky tak nejastji vznikaly na mstech pvodnch slatinnch olin, dubohabin nebo teplomilnch doubrav. Kysel bezkolencov louky se vyvjely na stanovitch raelinnch bor nebo kyselch doubrav, ppadn buin. Odstrannm slatinnch olin vznikala zptn spoleenstva vysokch ostic. Protoe vak lovk poteboval stle vce zemdlsk pdy, snail se zamoken stanovit odvodnit. Melioracemi a regulacemi vodnch tok dochzelo ke sniovn hladiny podzemn vody a spoleenstva vysokch ostic byla postupn nahrazena kalcitrofnmi nebo mezotrofnmi slatinnmi raeliniti, pi dalm poklesu hladiny podzemn vody, respektive pi jejm vraznjm kolsn v prbhu roku, pak bazifilnmi

12

bezkolencovmi loukami. Tmto zpsobem vznikly bezkolencov louky na mnoha mstech, nap. v Polab (Klika 1947, Vlek 1954, Kopeck 1960), v Podunajsk rovin (Zahradnkov-Roetzk 1965) nebo na Zho (marda 1951, Klika 1958, Boskov 1970, 1975). 2.2. Pdy

Bezkolencov louky se mohou vyskytovat jak na organogennch, tak na minerlnch pdch. Co se tk organogennch pd, tak bazifiln bezkolencov louky se nachzej na neutrlnch a bazickch raelinitnch pdch bohatch vpnkem. Geologickm podkladem jsou nejastji kdov horniny, slny, psit vpence, vpenit pskovce a bidlice. Acidofiln bezkolencov louky jsou vzny na kysel raelinitn pdy s velmi nzkm obsahem vpnku, kter se vyvj na horninch chudch pdotvornmi minerly (Vlek 1954).

Z minerlnch pd je nejastjm typem pseudoglej. V dsledku vraznho oglejen se jedn o pdu se znan nepznivmi fyziklnmi vlastnostmi. Zhutnl spodiny zpsobuj siln seznn pevlhovn povrchovch horizont, jeho dsledkem je pedevm citeln nedostatek vzduchu v pd (Tomek 1995).

Bezkolencov louky, kter se vyskytuj na aluvich ek, se nachzej na nivnch pdch, acidofiln bezkolencov louky jsou rovn rozeny na kyselch hndch pdch. 2.3. Kolsn hladiny podzemn vody

Pro vegetaci bezkolencovch luk je charakteristick vrazn kolsn hladiny podzemn vody bhem vegetan sezony. Prbh tohoto kolsn sledovalo v esk a Slovensk republice nkolik autor (Kopeck 1960, Zahradnkov-Roetzk 1965, Baltov-Tulkov 1966, 1968, Blakov 1973).

Kopeck (1960) mil bhem roku 1957 kolsn hladiny podzemn vody v bezkolencovch loukch asociace Molinietum caeruleae na nkterch lokalitch v Polab. Jak vyplv z jeho vsledk, na zatku vegetan sezony (bezen a duben) byla hladina podzemn vody pomrn vysoko (1025 cm pod povrchem), pot postupn klesala a do pelomu ervence a srpna (do hloubky 5070 cm), kdy zejm z dvodu silnch de prudce vystoupila k povrchu. V prbhu srpna opt pomrn prudce klesala a nejne byla v z a listopadu (vce ne 70 cm pod povrchem), v prosinci zaala postupn stoupat.

Podrobn men kolsn hladiny podzemn vody v bezkolencovch loukch asociace Molinietum caeruleae provdla v roce 1959 na itnm ostrov rovn Zahradnkov-Roetzk (1965). Na tto lokalit je vka hladiny podzemn vody ovlivnna vkou hladiny Dunaje, kter dosahuje maxima v ervenci a srpnu. V tomto obdob byla tud hladina podzemn vody nejve i v bezkolencovch loukch (u Molinietum caeruleae deschampsietosum caespitosae dosahovala a k povrchu, u M. c. caricetosum paniceae do 35 cm a u M. c. brometosum erecti kolsala v rozmez 5561 cm pod povrchem pdy). Klesat zaala a v podzimnch mscch a nejne byla v listopadu a prosinci u M. c. deschampsietosum caespitosae 3235 cm, u M. c. caricetosum paniceae 7080 cm a u M. c. brometosum erecti 105110 cm pod povrchem.

Baltov-Tulkov (1966, 1968) mila kolsn podzemn vody v porostech asociace Serratulo-Festucetum commutatae na jin Morav a v porostech asociac Molinietum caeruleae a Silatum pratensis na Zho. V loukch piazench k asociaci Serratulo-Festucetum commutatae, kter se nachzely na vyvench mstech v aluviu eky Moravy, dosahovala hladina podzemn vody maxima v obdob jarnch zplav, pot s menmi vkyvy postupn klesala a nejne byla v pozdnm lt a na podzim (ne ne 1 m pod povrchem). Podobn prbh kolsn podzemn vody byl pozorovn i v porostech asociac Molinietum

13

caeruleae a Silatum pratensis. Nejve byla hladina podzemn vody na jae, kdy dosahovala a povrchu, nejne opt v pozdnm lt a na podzim (vce jak 60 cm pod povrchem).

Blakov (1973) mila v letech 196465 kolsn hladiny podzemn vody v porostech asociac Junco-Molinietum caeruleae a Sanguisorbo-Festucetum pratensis na nkolika lokalitch v jinch echch. U obou tchto asociac dosahovala hladina podzemn vody maxima opt na zatku vegetan sezony, v obdob od kvtna do listopadu vrazn kolsala.

Z uvedench studi vyplv, e na stanovitch bezkolencovch luk dochz k vraznmu kolsn hladiny podzemn vody v prbhu roku. Na zatku vegetan sezony dosahuje obvykle hladina podzemn vody nejble k povrchu, pot postupn kles s obasnmi vkyvy, zpsobenmi silnjmi deti v dob men, a nejne je v pozdnm lt a na podzim. Pokud by nedochzelo k vraznm vkyvm a hladina spodn vody by kolsala pouze mrn (tj. asi do 3040 cm pod povrchem), byly by na stejnch stanovitch jin typy vegetace. Na vpnkem bohatch slatinnch pdch by byla vyvinuta vegetace svazu Caricion davallianae, na kyselch a vpnkem chudch pdch raelinn louky nebo pechodov raelinit. Vznik bezkolencovch luk v dsledku poklesu nebo vraznjho kolsn hladiny podzemn vody z tchto typ vegetace je dokladovn z ady lokalit (Klika 1947, Vlek 1954, 1956, Kopeck 1960, Boskov 1970, 1975). Naopak dlouhodob pokles hladiny podzemn vody vede k zniku bezkolencovch luk. Kysel bezkolencov louky nahrad vegetace smilkovch trvnk, bazick vegetace suchch trvnk nebo v ppad, e byla podzemn voda zdrojem vpnku, rovn vegetace smilkovch trvnk. Rovn tento typ sukcese byl pozorovn na nkolika lokalitch (Vlek 1954, 1956, Kopeck 1960, Kovcs 1962, Boskov 1970, Blakov 1973).

Kolsn hladiny podzemn vody je vak pouze jednou z podmnek existence bezkolencovch luk na danm stanoviti. Dalmi nezbytnmi podmnkami je vhodn typ substrtu a zpsob obhospodaovn.

Krom bezkolencovch luk dochz k vraznmu kolsn hladiny podzemn vody i v jinch typech lun vegetace. Pkladem mohou bt aluviln louky svazu Deschampsion, kter jsou na jae pravideln zaplavovny a v lt naopak pda znan vysych a podzemn voda kles hluboko pod povrch (Baltov-Tulkov 1966, Vicherek et al. 2000). Hlavn rozdl proti vegetaci bezkolencovch luk je ten, e jarn zplavy s sebou pinej mnostv ivin, a tm jsou tyto louky kadoron pirozen hnojeny. Z tohoto dvodu se mohou v tto vegetaci uplatnit druhy nronj na iviny, kter v bezkolencovch loukch bu pln chybj nebo se vyskytuj jen s malou pokryvnost. Rovn v loukch svazu Deschampsion chybj druhy tdy Scheuchzerio-Caricetea fuscae, vzan na oligotrofnj stanovit. Dal podstatn rozdl je ten, e jsou tyto louky koseny 23krt ron, piem prvn se bv v kvtnu a ervnu.

K vraznjmu kolsn hladiny podzemn vody me dochzet i v nkterch spoleenstvech pchovch luk svazu Calthion. Ta se vak na rozdl od bezkolencovch luk vyskytuj na ivinami bohatch stanovitch nebo jsou pravideln pihnojovny (Baltov-Tulkov 1968, Baltov-Tulkov et al. 1977).

Vedle absolutn vky hladiny podzemn vody a intenzity jejho kolsn v prbhu roku je dleitm ekologickm faktorem rovn vzlnavost. Psit pdy maj kapilrn zdvih mnohem ni ne pdy jlovit. To ovlivuje nejen vlhkostn reim svrchnch vrstev pdy, ale i jejich zsoben ivinami, zejmna vpnkem (Kopeck 1960).

Vrazn kolsn pdn vlhkosti v prbhu vegetan sezony zpsobuje znan rozdly v obsahu vzduchu v pd. Na jae, kdy hladina podzemn vody dosahuje asto a k povrchu, je obsah vzduchu v pd minimln, v pozdnm lt a na podzim, kdy je hladina podzemn vody nejne, je naopak obsah vzduchu v pd vysok. Pevn vtina pdnch pr je nekapilrnch. To zpsobuje, e tyto pdy maj velmi mal mnostv dynamicky pohybliv,

14

fyziologicky inn vody a povrch pdy trp v letnch mscch silnm nedostatkem vody (Vlek 1956).

2.4. Obhospodaovn

Zpsob obhospodaovn je dalm faktorem, kter je nezbytn pro zachovn vegetace bezkolencovch luk. Pro tyto louky je charakteristick obhospodaovn extensivn, tzn. se seou, ppadn pepsaj pouze jednou za rok, nkdy i jednou za vce let. Velmi dleit je tak to, e se nehnoj nebo pouze nepravideln pi obasn pastv. Tento zpsob obhospodaovn vak umouje pouze nzk vnosy, a je proto v dnen dob pro zemdlsk vrobce nezajmav. To je tak jeden z hlavnch dvod, pro jsou bezkolencov louky v souasn dob tak vzcn.

V ppad, e se zanou bezkolencov louky intenzivn obhospodaovat, zmn se postupn v jin vegetan typy. Pravidelnm hnojenm se su typy bazifilnch i acidofilnch bezkolencovch luk pemn v mezofiln louky svazu Arrhenatherion, vlh typy v louky svazu Calthion (Klika 1947, Kopeck 1960, Blakov & Kuera in Kolbek et al. 1999). Nevhodn je rovn ast kosen, nebo vtina druh charakteristickch pro tuto vegetaci, jako nap. Molinia caerulea agg., Serratula tinctoria nebo Succisa pratensis, kvete a v ervenci nebo srpnu. Tak intenzivn pastva vede vlivem nadmrnho psunu ivin k postupnmu zniku tchto luk a k nahrazen eutrofnjmi vegetanmi typy, jako jsou louky svazu Arrhenatherion nebo Cynosurion.

Dalm nezbytnm pedpokladem existence vegetace bezkolencovch luk je zachovn vodnho reimu. Pi jeho dlouhodobm naruen dochz rovn k jejich postupnmu zniku a k nahrazen jinmi vegetanmi typy (viz pedchoz kapitola).

Do souasn doby zstaly zachovny pouze fragmenty bezkolencovch luk, kter jsou nejvce ohroeny zarstnm. Vtina tchto luk vznikla a v dsledku lidsk innosti na mstech pvodnch les a byla udrovna prv obasnm seenm nebo pastvou. V dnen dob se vak velk st tchto luk ji neobhospodauje, vjimkou jsou zpravidla pouze porosty v chrnnch zemch. pln absence managementu se projevuje postupnm zarstnm. V oputnch porostech tak asto pevldne njak vrazn dominanta, nejastji Molinia caerulea agg., a zrove kles druhov bohatost. Tak se zanou uplatovat semenky ke a strom, kter postupn tuto vegetaci zcela perostou. 2.5. Souasn stav a rozen

Bezkolencov louky byly v minulosti pomrn hojn rozeny v cel stedn Evrop (Klika 1947, Wagner 1950, Vlek 1954, 1956, Kovcs 1962, Zahradnkov-Roetzk 1965, pnikov 1971, Boskov 1975). V dsledku rostouc poteby zemdlsk pdy zejmna v prbhu 19. a 20. stolet vak byla vtina z nich postupn odvodnna a rozorna. Jet z 50. let minulho stolet jsou bezkolencov louky dokladovny z ady lokalit v esk republice i na Slovensku. Bazifiln bezkolencov louky byly rozeny zejmna v ninch, jako nap. v Polab (Klika 1947, Vlek 1954, Kopeck 1960), na Zho (marda 1951, Klika 1958, Boskov 1970, 1975), v Podunajsk rovin (Zahradnkov-Roetzk 1965) nebo v Koick kotlin (pnikov 1971). Acidofiln bezkolencov louky se vyskytovaly zejmna ve vych nadmoskch vkch a v oblastech s geologicky chudm podlom, nap. na Tebosku nebo eskomoravsk vrchovin (Vlek 1956). Na Slovensku byla tato vegetace pravdpodobn mn ast, je uvdna nap. z nkterch lokalit na Zho (Boskov 1970, 1975), kde se vyskytovala na psitch pdch. V prbhu druh poloviny minulho stolet vak byly i tyto lokality z velk sti odvodnny a pemnny na ornou pdu nebo kulturn louky. Do souasn doby se zachovaly pouze mal fragmenty, nap. v rznch prodnch rezervacch

15

(nap. NPR Polabsk ernava, NPR Hrabanovsk ernava, PR Abrod). Mimo chrnn zem se bezkolencov louky nachz hojnji jet ve vojenskch prostorech (nap. VVP Ddice, V Boletice), kde zstaly zachovny v dsledku extensivnho hospodaen, tzn. obasnho kosen a nehnojen. 2.6. Syntaxonomick lenn

V bezkolencovch loukch svazu Molinion bylo do souasn doby rozlieno v esk republice dvanct asociac (Blakov & Baltov-Tulkov in Moravec et al. 1995, Baltov-Tulkov 2001), na Slovensku trnct (Baltov-Tulkov 1996).

Jedn se o nsledujc asociace: Molinietum caeruleae Koch 1926 do tto asociace byly azeny porosty na bzemi

bohatch stanovitch, obvykle s vy pokryvnost Molinia caerulea agg. (Klika 1947, marda 1951, Vlek 1954, Klika 1958, Kopeck 1960, Zahradnkov-Roetzk 1965, Vicherek 1967, Boskov 1970, 1975, pnikov 1971, Baltov-Tulkov 1994).

Selino-Molinietum Kuhn 1937 bezkolencov louky, ve kterch se s vt pokryvnost uplatovaly Selinum carvifolia a Molinia caerulea agg. (Baltov-Tulkov 1968, 1993a, 1994)

Junco-Molinietum caeruleae Preising 1951 zpravidla druhov chud porosty na kyselch a oligotrofnch pdch s dominujc Molinia caerulea agg. (Blakov 1973, Baltov-Tulkov 1983, 1993a, 1994, 1997, Ruikov 1985, Baltov-Tulkov & Hberov 1996, Baltov-Tulkov & Kontriov 1999)

Silatum pratensis Knapp 1954 louky s dominanc druhu Silaum silaus (Baltov-Tulkov 1968, 1994, Baltov-Tulkov & Hjek 1998, Vicherek et al. 2000, Baltov-Tulkov 2001)

Diantho superbi-Molinietum caeruleae Passarge 1955 porosty s dominanc Molinia caerulea agg. a ptomnost druhu Dianthus superbus (Baltov-Tulkov 1996)

Scorzonero-Molinietum Oberdorfer & Krause 1955 bezkolencov louky s vy pokryvnost Scorzonera humilis a Molinia caerulea agg. (Baltov-Tulkov 1983, 1991, Baltov-Tulkov & Kontriov 1999)

Sanguisorbo-Festucetum commutatae Baltov-Tulkov 1959 louky s vy pokryvnost Sanguisorba officinalis a Festuca rubra agg. (Baltov-Tulkov & Zapletal 1959, Baltov-Tulkov 1965, 1974, 1977, 1991, 1993a, 1994, Duchoslav 1997, Baltov-Tulkov & Kontriov 1999)

Serratulo-Festucetum commutatae Baltov-Tulkov 1966 porosty s dominanc Festuca rubra agg. nebo Festuca rupicola a ptomnost Serratula tinctoria, obvykle bez Molinia caerulea agg. (Baltov-Tulkov 1966, 1968, 1975, 1994, Kov 1981, Baltov-Tulkov & Hjek 1998, Baltov-Tulkov & Kontriov 1999)

Silao-Molinietum caeruleae Baltov-Tulkov 1968 bezkolencov louky s dominujc Molinia caerulea agg. a ptomnost druhu Silaum silaus (Baltov-Tulkov 1968)

Gladiolo palustris-Molinietum caeruleae Baltov-Tulkov 1968 porosty s dominanc Molinia caerulea agg. a ptomnost Gladiolus palustris (Baltov-Tulkov 1968)

Gentiano pneumonanthis-Molinietum litoralis Ilijani 1968 bezkolencov louky s vy pokryvnost druhu Molinia litoralis (Baltov-Tulkov 1993b, 2000,

16

Baltov-Tulkov & Hjek 1998, Baltov-Tulkov & Kontriov 1999, Vicherek et al. 2000)

Sanguisorbo-Festucetum pratensis Blakov 1973 louky s vy pokryvnost trav nronjch na iviny (Blakov 1973, Kolbek et al. 1999)

Carici davallianae-Molinietum caeruleae pnikov 1978 porosty bezkolencovch luk, ve kterch jsou zastoupeny druhy tdy Scheuchzerio-Caricetea fuscae (pnikov 1978)

Galio borealis-Cirsietum cani Baltov-Tulkov 2001 porosty s vy pokryvnost druhu Cirsium canum a ptomnost Galium boreale (Baltov-Tulkov 2001)

Z uvedenho pehledu asociac vyplv, e vtina z nich byla rozliena na zklad vt

pokryvnosti nebo ptomnosti jednoho nebo dvou druh. Porovnnm asociac mezi sebou vak zjistme, e celkov druhov sloen nkterch z nich je velmi podobn a rovn, e pro adu z nich jsou uvdny stejn diagnostick druhy. Nen proto zcela jasn, kter z tchto asociac m smysl rozliovat a kter pedstavuj pouze synonyma ji dve popsanch asociac. Z tohoto dvodu je proto nezbytn provst syntaxonomickou revizi, na zklad kter budou vylenny pouze asociace, kter budou jasn definovny, zejmna v tom smyslu, e budou mt vlastn diagnostick druhy. 2.7. Charakteristika druh typickch pro stdav vlhk stanovit

Pro bezkolencov louky jsou charakteristick druhy stdav vlhkch stanovi, jako je nap. Betonica officinalis, Dianthus superbus, Galium boreale, Gentiana pneumonanthe, Iris sibirica, Molinia caerulea agg., Scorzonera humilis, Selinum carvifolia, Serratula tinctoria nebo Succisa pratensis. Jedn se o druhy, kter jsou pizpsobeny na vrazn zmny vlhkosti pdy bhem roku.

Vtina tchto druh (krom Scorzonera humilis) zasahuje svm rozenm daleko na vchod, do oblast s kontinentlnm klimatem. Kontinentln klima je relativn such (ron hrn srek je vtinou 300600 mm), s velkmi teplotnmi rozdly v prbhu roku. Zimy jsou extrmn chladn, teploty klesaj i pod 40 C, lta jsou naopak vrazn hork. Dostatek vlhy je pouze v jarnch mscch, zatmco v lt je pda extrmn vyschl. V tchto oblastech jsou ve uveden druhy vzny na vegetaci lesostep (tda Brachypodio-Betuletea).

V naich podmnkch rostou na stanovitch, kde z njakch pin dochz k podobnmu stdn vlhkosti pdy bhem roku. Druhy jako Betonica officinalis, Galium boreale, Gentiana pneumonanthe, Iris sibirica, Selinum carvifolia a Serratula tinctoria se krom bezkolencovch luk vyskytuj rovn v ninnch aluvilnch loukch svazu Deschampsion, kter jsou na jae pravideln zaplavovny, zatmco v lt pda znan vysych. Betonica officinalis, Galium boreale, Molinia caerulea agg. a Serratula tinctoria se rovn vyskytuj v subkontinentlnch suchch trvncch svazu Cirsio-Brachypodion pinnati, kter jsou rozeny na tkch jlovitch pdch, ktermi tko prosakuje voda. Proto jsou tyto pdy na jae zamokeny a v lt siln vysychaj. Krom lunch spoleenstev se druhy Betonica officinalis, Dianthus superbus, Galium boreale, Molinia caerulea agg., Selinum carvifolia a Serratula tinctoria nachz tak v rozvolnnch teplomilnch doubravch asociace Potentillo albae-Quercetum (svaz Quercion petraeae), kter je podobn jako such trvnky svazu Cirsio-Brachypodion pinnati vzna na tk jlovit nebo slinit pdy, a v subkontinentlnch teplomilnch doubravch svazu Aceri tatarici-Quercion, kter se vyskytuj na spraovch ploinch.

17

3. Cle disertan prce 1. Provst analzu luk tdy Molinio-Arrhenatheretea z esk republiky s clem zjistit hlavn

faktory prosted, kter nejvraznji ovlivuj variabilitu jejich druhovho sloen. Pouitm numerickch metod stanovit hlavn vegetan typy a porovnat je s vegetanmi typy rozliovanmi v tradin fytocenologii. Otestovat, zda je beta-diverzita vlhkch luk vy ne beta-diverzita luk mezofilnch, co by mohlo bt dvodem toho, e jsou vlhk louky v tradin fytocenologii lenny do vce asociac ne louky mezofiln. Ovit, zda je svaz Molinion jako fytocenologick jednotka dobe floristicky vymezen vi jinm svazm.

2. Vyhodnotit variabilitu druhovho sloen vlhkch ninnch luk podl gradientu

klimatick kontinentality ve stedn Evrop. Numerickmi metodami identifikovat hlavn vegetan typy tchto luk. Provst syntaxonomickou interpretaci rozliench vegetanch typ. Zjistit, zda jsou ninn bezkolencov louky svazu Molinion dobe floristicky vyhranny vi ostatnm vlhkm ninnm loukm.

3. Provst syntaxonomickou revizi asociac svazu Molinion uvdnch pro eskou republiku

a stanovit jasn pravidla pro piazen snmk k rozlienm asociacm. Prostednictvm numerickch metod najt faktory prosted, kter nejvraznji ovlivuj variabilitu druhovho sloen v bezkolencovch loukch. Stanovit diagnostick druhy rozliench asociac, charakterizovat jejich strukturu a druhov sloen, stanovit, na kterch se vyskytuj, popsat dynamiku, management a rozen v Evrop. Numerickmi metodami vyhodnotit variabilitu uvnit asociac. Zpracovat vegetaci svazu Molinion pro monografii Vegetace esk republiky.

4. Pouitm numerickch metod zjistit hlavn vegetan typy bezkolencovch luk svazu

Molinion na Slovensku a porovnat je s vegetanmi jednotkami rozliovanmi v tradin fytocenologii. Najt faktory prosted, kter nejvraznji ovlivuj variabilitu druhovho sloen v tchto loukch. Urit diagnostick druhy vylennch vegetanch jednotek, popsat jejich strukturu, druhov sloen a stanovit, na kterch se vyskytuj.

18

4. Struktura prce Prce se skld ze t publikovanch lnk a dvou rukopis pijatch k tisku:

1) Havlov M., Chytr M. & Tich L. (2004): Diversity of hay meadows in the Czech

Republic: major types and environmental gradients. Phytocoenologia 34: 551567. 2) Botta-Dukt Z., Chytr M., Hjkov P. & Havlov M. (2005): Vegetation of lowland

wet meadows along a climatic continentality gradient in Central Europe. Preslia 77: 89111.

3) Havlov M. (2006): Syntaxonomical revision of the Molinion meadows in the Czech

Republic. Preslia 78: 87101. 4) eznkov M. (2007): Svaz TDD. Molinion caeruleae. In: Chytr M. (ed.): Vegetace

esk republiky 1. Travinn a kekov vegetace. Academia, Praha (pijato k tisku). 5) eznkov M. (2007): Variability of the Molinion meadows in Slovakia. Biologia

(pijato k tisku).

19

4.1. Vymezen podlu spoluautor Milan Chytr Spoluautor lnk 1 a 2. Podlel se na vytven hypotz, na metodice a textech lnk. Lubomr Tich Je spoluautorem lnku 1. Naprogramoval funkci v programu Juice, pomoc kter byla spotna beta-diverzita. Zoltn Botta-Dukt Hlavn autor lnku 2. Navrhoval hypotzy, metodiku a zpracoval charakteristiku luk svazu Deschampsion. Petra Hjkov Podlela se na zpracovn lnku 2. Napsala charakteristiku pchovch luk svazu Calthion.

20

5. Hlavn vsledky prce 1. Pouitm detrendovan korespondenn analzy a Ellenbergovch indikanch hodnot

bylo zjitno, e nejvznamnjm faktorem, kter ovlivuje variabilitu druhovho sloen lun vegetace v esk republice, je vlhkost, druhm nejdleitjm faktorem jsou iviny, s nimi koreluje pH pdy. Pomoc shlukov analzy bylo rozlieno 10 hlavnch vegetanch typ. Skupina 1 sdruovala ninn aluviln louky svazu Deschampsion, skupiny 2, 3, 5, 9 a 10 rzn typy vlhkch pchovch luk svazu Calthion, skupina 4 louky svazu Molinion, skupina 6 zahrnovala horsk pchov louky, kter pedstavuj pechodn typ vegetace mezi svazy Calthion a Polygono-Trisetion, skupina 7 ninn ovskov louky svazu Arrhenatherion a skupina 8 ovskov louky vych poloh na ivinami chudch pdch, horsk trojttov louky svazu Polygono-Trisetion a mezofiln pastviny svazu Cynosurion. Skutenost, e vlhk pchov louky byly rozdleny ve srovnn s mezofilnmi loukami do vce skupin, odpovd jednak lenn v tradin fytocenologii, kde je ve svazu Calthion rozliovn nejvt poet asociac, a jednak i vsledkm analzy beta-diverzity, kter ukzala, e v lun vegetaci na vlhkch stanovitch jsou vt rozdly v druhovm sloen mezi jednotlivmi typy vegetace. Z vsledk tto prce rovn vyplv, e bezkolencov louky svazu Molinion pedstavuj v rmci tdy Molinio-Arrhenatheretea zeteln vyhrann typ vegetace. viz lnek 1

2. V tto prci bylo analyzovno 387 fytocenologickch snmk vlhkch ninnch luk z

nadmoskch vek do 350 m z esk republiky, vchodnho Rakouska, Slovenska, Maarska a severnho Chorvatska za elem sjednocen jejich klasifikace, kter se v jednotlivch zemch liila. Pomoc analzy hlavnch koordint (PCoA) byla odstranna nesystematick sloka variability v datech a cel soubor byl klasifikovn shlukovou analzou. Pouitm nov vyvinut metody pro uren optimlnho potu shluk byl soubor nejprve rozdlen na ti shluky. Prvn zahrnoval pchov louky svazu Calthion, druh bezkolencov louky svazu Molinion a tet kontinentln zaplavovan louky, azen do svaz Agrostion albae, Alopecurion pratensis, Cnidion venosi, Deschampsion cespitosae a Veronico longifoliae-Lysimachion vulgaris. V dalm kroku byl soubor rozdlen na devt shluk, z nich tyi zahrnovaly kontinentln zaplavovan louky, ti pchov louky a dva shluky odpovdaly bezkolencovm loukm. Protoe dal dlen kontinentlnch zaplavovanch luk neodrelo hranice mezi svazy tak, jak jsou uvdny v tradin fytocenologick literatue, bylo navreno slouit tyto svazy do jednoho a pojmenovat je nejstarm platnm jmnem, tj. Deschampsion cespitosae Horvati 1930. Bezkolencov louky svazu Molinion byly v prvnm kroku jasn odlieny od ostatnch ninnch luk. Dalm dlenm byly tyto louky rozlenny na dva shluky, kter vak mly velmi podobn druhov sloen, hlavn rozdl byl ten, e jeden shluk obsahoval druhov chud snmky, zatmco druh shluk druhov bohat snmky. Rozdlen bezkolencovch luk na acidofiln a bazifiln nebylo mon v tomto ppad oekvat, protoe acidofiln bezkolencov louky jsou ve stedn Evrop rozeny pevn ve vych nadmoskch vkch, a tud byly v analyzovanm souboru zastoupeny jen nkolika snmky. viz lnek 2

3. Z pvodn deseti asociac svazu Molinion uvdnch pro eskou republiku byly metodou

Cocktail rozlieny pouze dv Molinietum caeruleae Koch 1926 a Junco effusi-Molinietum caeruleae Txen 1954. Ostatn asociace je mon povaovat za synonyma tchto dvou asociac. Molinietum caeruleae se vyskytuje na bazitjch pdch a je charakterizovno druhy Betonica officinalis, Galium boreale a Serratula tinctoria, zatmco Junco effusi-Molinietum caeruleae je rozeno na pdch kyselejch a je typick

21

vskytem druh smilkovch trvnk. V asociaci Molinietum caeruleae byly shlukovou analzou rozlieny ti varianty varianta Carex hostiana, pro kterou je charakteristick vskyt druh svazu Caricion davallianae a kter se v minulosti hojnji vyskytovala zejmna v Polab, varianta Bromus erectus, kter zahrnuje nejsu porosty asociace Molinietum caeruleae a je charakterizovna vskytem druh svaz Bromion erecti, Cirsio-Brachypodion pinnati a Arrhenatherion elatioris, a varianta Scorzonera humilis, kter se v porovnn s pedchozmi dvma variantami vyskytuje na kyselejch pdch a pedstavuje pechodn typ k asociaci Junco effusi-Molinietum caeruleae. V tto asociaci byly rozlieny dv varianty varianta Valeriana dioica, kter je rozena na vlhch stanovitch a je typick vskytem druh svaz Caricion fuscae a Calthion, a varianta Leucanthemum vulgare, charakteristick druhy smilkovch trvnk a mezofilnch luk. Podle Ellenbergovch indikanch hodnot, kter byly pouity pro interpretaci hlavnch gradient, koreloval hlavn gradient v druhovm slouen luk svazu Molinion pozitivn s pdn reakc, kontinentalitou, teplotou a ivinami a negativn s nadmoskou vkou. viz lnky 3 a 4

4. Na Slovensku bylo pvodn uvdno dvanct asociac svazu Molinion. Po proveden

syntaxonomick revize vak byly podobn jako v esk republice rozlieny pouze asociace Molinietum caeruleae Koch 1926 a Junco effusi-Molinietum caeruleae Txen 1954. Shlukovou analzou bylo identifikovno devt hlavnch vegetanch typ bezkolencovch luk. Osm z nich je mon piadit k Molinietum caeruleae, pouze jeden shluk odpovd svm druhovm sloenm asociaci Junco effusi-Molinietum caeruleae. Krom variant rozliench v esk republice (varianta Carex hostiana odpovd subasociaci caricetosum hostianae Koch 1926, varianta Bromus erectus subasociaci caricetosum tomentosae Koch 1926) byly na Slovensku identifikovny subasociace caricetosum paniceae Koch 1926, rozen na vlhkch mezotrofnch stanovitch, subasociace potentilletosum albae Boskov 1970, kter se vyskytuje na such oligotrofnch psitch pdch a je charakteristick vskytem druh suchch a smilkovch trvnk, a subasociace deschampsietosum cespitosae Zahradnkov-Roetzk 1965 vyskytujc se na vlhkch, ivinami bohatch stanovitch a zahrnujc druhy tdy Phragmito-Magnocaricetea a svazu Deschampsion. Podobn jako v esk republice koreloval hlavn gradient s pdn reakc, kontinentalitou, teplotou a ivinami. viz lnek 5

22

6. Literatura Baltov-Tulkov E. (1965): Die Sumpf- und Wiesenpflanzengesellschaften der

Mineralbden sdlich des Zbeh bei Hlun. Vegetatio 13: 151. Baltov-Tulkov E. (1966): Synkologische Charakteristik der sdmhrischen

berschwemmungswiesen. Rozpr. s. Akad. Vd, ser. math.-natur., 76: 141. Baltov-Tulkov E. (1968): Grundwasserganglinien und Wiesengesellschaften.

(Vergleichende Studie der Wiesen aus Sdmhren und der Sdwestslowakei). Prod. Pr. st. s. Akad. Vd 2: 137.

Baltov-Tulkov E. (1974): Fytocenologick charakteristika lunho komplexu u Dolnch ivotic (Opavsko). as. Slez. Muz., ser. A, 23: 5769.

Baltov-Tulkov E. (1975): Das Succiso-Festucetum commutatae und das Serratulo-Festucetum commutatae als Beispiel der Konvergenz in der Pflanzensoziologie. In: Dierschke H. (ed.), Vegetation und Substrat, pp. 117130, Ber. Int. Symp. IV Rinteln 1969, J. Cramer, Vaduz.

Baltov-Tulkov E. (1977): Zur Kenntnis der Nass- und Feuchtwiesen im Graben Hornomoravsk val. Preslia 49: 135160.

Baltov-Tulkov E. (1983): Feuchtwiesen des Landschaftsschutzgebietes umava (Bhmerwald). Folia Mus. Rer. Natur. Bohem. Occid., Bot. 1819: 182.

Baltov-Tulkov E. (1991): Feuchtwiesen des Brdy-Berglandes und seiner Randgebiete (Mittelbhmen). Folia Geobot. Phytotax. 26: 179.

Baltov-Tulkov E. (1993a): Nass- und Feuchtwiesen des sdwestbhmischen Region Blatensko. Folia Mus. Rer. Natur. Bohem. Occid., Bot. 3738: 137.

Baltov-Tulkov E. (1993b): Das Gentiano pneumonanthis-Molinietum litoralis Ilijani 1968 in Sd-Mhren und der Slowakei. Tuexenia 13: 193201.

Baltov-Tulkov E. (1994): Alopecurion- und Molinion-Gesellschaften NW-Bhmens (Phytoznologische und kologische Charakteristik). Acta Sci. Nat. Brno 28: 155.

Baltov-Tulkov E. (1996): Diversitt der Nass- und Feuchtwiesen der Ordnung Molinietalia in der Tschechischen und der Slowakischen Republik. Acta Bot. Gallica 143: 299307.

Baltov-Tulkov E. (1997): Feuchtwiesen- und Hochstaudengesellschaften des Landschaftsschutzgebietes Luick hory und der angrenzenden Randgebiete (Nordbhmen). Verh. Zool.-Bot. Ges. sterr. 134: 233304.

Baltov-Tulkov E. (2000): Vergleich der aus Mhren und der Slowakei beschriebenen Subassoziationen der Gentiano pneumonanthis-Molinietum litoralis. Acta Bot. Croat. 59: 157165.

Baltov-Tulkov E. (2001): Silaetum pratensis und Galio borealis-Cirsietum cani in Ost- und Zentral-Bhmen. Verh. Zool.-Bot. Ges. sterr. 138: 157176.

Baltov-Tulkov E. & Hberov I. (1996): Feuchtwiesen des Landschaftsschutzgebietes Slovensk Kras (SO-Slowakei). Tuexenia 16: 227250.

Baltov-Tulkov E. & Hjek M. (1998): Feuchtwiesengesellschaften des sdlichen Teiles des Landschaftsschutzgebietes Bl Karpaty (Sdost-Mhren). Verh. Zool.-Bot. Ges. sterr. 135: 140.

Baltov-Tulkov E. & Kontriov O. (1999): Quell-, Wiesen- und Hochstauden-Gesellschaften der Ordnung Molinietalia im Landschaftsschutzgebiet und Biosphrenreservat Poana (Zentralslowakei). Tuexenia 19: 351392.

Baltov-Tulkov E. & Zapletal A. (1959): Druh pspvek k typologii luk Slezska (Louky katastru obce Zbeh u Hluna). Prod. as. Slez. 20: 435470.

23

Baltov-Tulkov E., Zelen V. & Tesaov M. (1977): Synkologische Charakteristik einiger wichtiger Wiesentypen des Naturschutzgebietes rsk vrchy. Rozpr. eskoslov. Akad. Vd, ada Mat. Pr. Vd, 87: 1115.

Blakov D. (1973): Pflanzensoziologische Studie ber die Wiesen der sdbhmischen Becken. Stud. SAV 1973/10: 1170.

Borhidi A. (2003): Magyarorszg nvnytrsulsai. Akadmiai Kiad, Budapest. Boskov E. (1970): Kvetena a rastlinn spoloenstv ttnej prrodnej rezervcie Abrod na

Zhor. Pr. td. s. Ochr. Prr. 2: 183. Boskov E. (1975): Rastlinn spoloenstv slatinovch lk na Zhorskej nine. s. Ochr.

Prr. 15: 173273. Burkart M., Dierschke H., Hlzel N., Nowak B. & Fartmann T. (2004): Synopsis der

Pflanzengesellschaften Deutschlands. Heft 9. Molinio-Arrhenatheretea (E1). Kulturgrasland und verwandte Vegetationstypen. Teil 2: Molinietalia. Futter- und Streuwiesen feucht-nasser Standorte und Klassenbersicht Molinio-Arrhenatheretea. Floristisch-soziol. Arbeitsgem., Gttingen.

Dierschke H. (1995): Syntaxonomical survey of Molinio-Arrhenatheretea in Central Europe. Coll. Phytosoc. 23: 387399.

Dierschke H. (1997): Molinio-Arrhenatheretea (E 1). Kulturgrasland und verwandte Vegetationstypen. Teil 1: Arrhenatheretalia. Wiesen und Weiden frischer Standorte. Synopsis der Pflanzengesellschaften Deutschlands. Heft 3. Gttingen.

Duchoslav M. (1997): The present state of meadow vegetation (Molinio-Arrhenatheretea) in the Morava river floodplain (Hornomoravsk val area). Zpr. es. Bot. Spole. 15: 131176.

Dupr C. & Diekmann M. (2001): Differences in species richness and life-history traits between grazed and abandoned grasslands in southern Sweden. Ecography 24: 275286.

Ellenberg H. (1996): Vegetation Mitteleuropas mit den Alpen in kologischer, dynamischer und historischer Sicht. Ed. 5. Verlag Eugen Ulmer, Stuttgart.

Ellmauer T. (1994): Syntaxonomie der Frischwiesen (Molinio-Arrhenatheretea p.p.) in sterreich. Tuexenia 14: 151168.

Ellmauer T. & Mucina L. (1993): Molinio-Arrhenatheretea. In: Mucina L., Grabherr G. & Ellmauer T. (eds.): Die Pflanzengesellschaften sterreichs. Teil I. Anthropogene Vegetation, pp. 297401. Gustav Fischer Verlag, Jena.

Jensen C., Vorren K. D., Eilertsen S. M. & Samuelsen R. (2001): Successionary stages of formerly cultivated grassland in northern Norway, abandoned for 10, 20 and 35 years. Nord.J. Bot. 21: 305320.

Klika J. (1947): Rostlinosociologick jednotky slatin a lunch porost v Polab. Vstn. Krl. esk Spolen. Nauk, T. Mat.-Pr. 1945/1: 131.

Klika J. (1958): K fytocenologii raelinnch a slatinnch spoleenstev na Zhorsk nin. Biol. Pr. 4: 134.

Koch W. (1926): Die Vegetationseinheiten der Linthebene unter Bercksichtigung der Verhltnisse in der Nordostschweiz. Systematisch-kritische Studie. Jahrb. St. Gallischen Naturwiss. Ges. 61/2: 1144.

Kolbek J., Blakov D., Bzov E., Loek V., Rybnkov E., Rybnek K. & Rydlo J. (1999): Vegetace Chrnn krajinn oblasti Kivokltsko. 1. Vvoj krajiny a vegetace, vodn, poben a lun spoleenstva. Agentura ochrany prody a krajiny R & Botanick stav AV R, Praha.

Kopeck K. (1960): Fytocenologick studie slatinnch luk v severovchodnch echch. Rozpr. s. Akad. Vd, ser. math.-natur. 70/4: 164.

Kovcs M. (1962): Die Moorwiesen Ungarns. Akadmiai Kiad, Budapest.

24

Kov P. (1981): The grassland communities of the southeastern basin of the Labe river. 1. Syntaxonomy. Folia Geobot. Phytotax. 16: 143.

Krippel E. (1965): Postglacilny vvoj lesov Zhorskej niny (Historicko-geobotanick tdia). Biol. Pr. 11: 199.

Kuera T. & umberov K. (2001): Louky a pastviny. In: Chytr M., Kuera T. & Ko M. (eds.), Katalog biotop esk republiky, pp. 109123, Agentura ochrany prody a krajiny R, Praha.

Linusson A. C., Berlin G. A. I. & Olsson E. G. A. (1998): Reduced community diversity in semi-natural meadows in southern Sweden, 19651990. Pl. Ecol. 136: 7794.

Moravec J., Baltov-Tulkov E., Blakov D., Hada E., Hejn S., Husk ., Jenk J., Kolbek J., Krahulec F., Krop Z., Neuhusl R., Rybnek K., ehoek V. & Vicherek J. (1995): Rostlinn spoleenstva esk republiky a jejich ohroen. Ed. 2. Severoes. Pr., suppl. 1995: 1206.

Oberdorfer E. (ed.) (1993): Sddeutsche Pflanzengesellschaften. Teil 3. Wirtschaftswiesen und Unkrautgesellschaften. Ed. 3. Gustav Fischer Verlag, Jena.

Pacltov B. & Huben E. (1994): To the history of forest formations of the central Labe-River region and to paleoecological conditions at Hrabanov fen peat (Hrabanovsk ernava). Palynological study. In: Rikov E. & Zeman A. (eds.), Holocene flood plain of the Labe-River. Contemporary state of research in the Czech Republic, pp. 6676, Prague.

Passarge H. (1964): Pflanzengesellschaften des nordostdeutschen Flachlandes I. Gustav Fischer Verlag, Jena.

Ruikov H. (1985): Bezkolencov lky (asociacia Junco-Molinietum Preising 1951) na Holom vrchu. In: Zbornk XXI. TOPu, pp. 513, Povadlo.

Stanov V. & Valachovi M. (eds.) (2002): Katalg biotopov Slovenska. Daphne, Intitt aplikovanej ekolgie, Bratislava.

marda J. (1951): Rostlinn spoleenstva slovenskho Zho. as. Mor. Mus. 36: 3868. pnikov A. (1971): Fytocenologick tdia lk juhozpadnej asti Koickej kotliny. Biol.

Pr. 17: 1103. pnikov A. (1975): Die Pflanzengesellschaften mit Alopecurus pratensis in der Slowakei.

Biolgia 30: 523534. pnikov A. (1978): Gesellschaften mit Molinia caerulea in der Slowakei. Biolgia 33:

291305. Tomek M. (1995): Atlas pd esk republiky. esk geologick stav, Praha. Txen R. (1954): Pflanzengesellschaften und Grundwasser-Ganglinien. Angew.

Pflanzensoz. (Stolzenau/Weser) 8: 6498. Vlek B. (1954): Pdy porost Molinia coerulea (W. Koch) v echch a jejich vztah k pdm

ostatnch raelinnch porost. I. Molinietum coeruleae na pdch alkalickch. Preslia 26: 385414.

Vlek B. (1956): Pdy porost Molinia coerulea (W. Koch) v echch a jejich vztah k pdm ostatnch raelinnch porost. II. Molinietum coeruleae na pdch s kyselou pdn reakc. Preslia 28: 169192.

Vicherek J. (1967): Pspvek k poznn vegetace slatinnch luk s Carex davalliana Sm. u Vranovic na jin Morav. Preslia 39: 9596.

Vicherek J., Antonn V., Danihelka J., Grulich V., Gruna B., Hradlek Z., ehoek V., umberov K., Vampola P. & Vgner A. (2000): Flra a vegetace na soutoku Moravy a Dyje. Masarykova univerzita, Brno.

Wagner H. (1950): Das Molinieum caeruleae (Pfeifengraswiesen) im Wiener Becken. Vegetatio 2: 128165.

Zahradnkov-Roetzk K. (1965): Geobotanick charakteristika slatinnch lk a pasienkov (Molinion Koch 1926) na itnom ostrove. Biol. Pr. 11: 145.

25

Zuidhoff A. C., Schamine J. H. J. & van t Veer R. (1996): Molinio-Arrhenatheretea. In: Schamine J. H. J, Stortelder A. H. F. & Weeda E. J. (eds.), De vegetatie van Nederland. Deel 3. Plantengemeenschappen van graslanden, zomen en droge heiden, pp. 163226, Opulus Press, Uppsala, Leiden.

26

7. Publikovan nebo k tisku pipraven texty

27

7.1. Havlov M., Chytr M. & Tich L. (2004): Diversity of hay meadows in the Czech Republic: major types and environmental gradients. Phytocoenologia 34: 551567.

28

Phytocoenologia 34 (4) 551567 BerlinStuttgart, December 8, 2004

Diversity of hay meadows in the Czech Republic:major types and environmental gradients

by Marcela Havlova, Milan Chytry and Lubomr Tichy, Brno

with 3 figures and 4 tables

Abstract. A stratified data set of 3102 releves of meadows and mesic pastures of theCzech Republic was analysed by detrended correspondence analysis and cluster analysis.Major gradients and clusters were interpreted using Ellenberg indicator values. The majorgradient in species composition was associated with soil moisture and the second mostimportant gradient with available nutrients. Clusters proposed by numerical classificationreproduced some of the traditional phytosociological alliances, namely Arrhenather-ion, Molinion and Polygono-Triset ion, while some other alliances were less clearlydifferentiated (e.g. Alopecurion, Cnidion and Cynosurion). Wet meadows of theCalthion alliance were divided among several clusters, which corresponded to the mainassociations recognized in traditional phytosociological literature. This patterns suggeststhat wet meadows have a higher beta-diversity than mesic meadows. We tested this hy-pothesis by calculating mean pair-wise Srensen dissimilarity for bootstrap subsamplesof meadow releves for partitions of the moisture gradient, and confirmed that beta-diver-sity of meadows increases with increasing soil moisture. In traditional phytosociologicalliterature, this fact is reflected by higher numbers of associations distinguished withinwet meadows than in mesic meadows.

Keywords: beta-diversity, classification, Ellenberg indicator values, grassland vegetation,ordination, phytosociology.

IntroductionHay meadows are the most widespread type of semi-natural vegetation inCentral Europe. Due to socio-economic changes in agriculture, which tookplace in the second half of the 20th century, areas of species-rich meadowshave been increasingly reduced by abandonment of some meadow tractsand introduction of intensive management systems with massive applica-tion of artificial fertilizers in other tracts (Ellenberg 1996, Linusson etal. 1998, Dupre & Diekmann 2001, Jensen et al. 2001). Hay meadow eco-systems have therefore appeared in the focus of nature conservation author-ities and many applied projects have been initiated with the aim of creatingmeadow inventories or re-establishing traditional management (Prach1996, Dzwonko & Loster 1998, Joyce & Wade 1999, Seffer & Stanova1999, Krahulec et al. 2001, Sedlakova & Fiala 2001, Vecrin et al. 2002,Losvik & Austad 2002).

Nature conservation survey projects are in need of a robust classificationof meadow vegetation. Phytosociological classification is perfectly suited

DOI: 10.1127/0340-269X/2004/0034-0551 0340-269X/04/0034-0551 $ 4.25 2004 Gebrder Borntraeger, D-14129 Berlin D-70176 Stuttgart

29

552 M. Havlova et al.

for this purpose, as it is based on floristic composition and thus directlylinked to biodiversity. In Central Europe, there is a long tradition of phyto-sociological study of meadows (Dierschke 1995), which has resulted in ageneral agreement as to the major types of meadow vegetation (Ober-dorfer 1993, Ellmauer & Mucina 1993, Ellmauer 1994, Dierschke1995, 1997, Blazkova & Balatova in Moravec et al. 1995, Zuidhoffet al. 1996, Kucera & Sumberova 2001). The main environmental gradientresponsible for variation in species composition of Central European mead-ows is moisture, as recognized in ecograms of Ellenberg (1996). Thisgradient is used for the primary division of meadows in the classificationstudies. Secondary gradients include altitude, nutrient availability, soil pHand water fluctuations. At the level of phytosociological alliances, the majortypes of meadow vegetation include mesic meadows of low altitudes (Arr-henatherion), manured pastures of low altitudes (Cynosurion), mesicmeadows of montane belt (Polygono-Triset ion), mesic meadows ofsubalpine belt (Poion alpinae), manured wet meadows (Calthion) andunmanured wet meadows (Molinion). Less agreement has been achievedupon justification of separate alliances for tall-forb vegetation replacing wetmeadows after abandonment (Fil ipendulion, Veronico longifol iae-Lysimachion vulgar is ) and for wet meadows of lowland river flood-plains (Alopecurion, Cnidion). Even less agreement is found at the as-sociation level, where delimitations of individual syntaxa often greatly varyamong different authors. It is striking that wet meadows, namely those ofthe Calthion alliance, are usually divided into more associations than me-sic meadows in phytosociological surveys (Balatova-Tulackova 1984).

Most of the phytosociological classifications of Central Europeanmeadow vegetation produced so far have been largely based on expertknowledge, and so was the estimation of main environmental gradients(Ellenberg 1996). Some classification exercises involved manual editing ofsynoptic tables taken from individual studies and some others did not in-volve data analysis at all. It is therefore important to test whether the majortypes and gradients recognized in the expert-based phytosociologicalstudies can also be recognized by numerical analysis of large data sets ofvegetation releves (Bruelheide & Chytry 2000, Studer-Ehrensberger2000). In this study, we use a large data set of meadow vegetation relevesfrom the Czech Republic and analyse it with respect to the following objec-tives: (1) to reveal the major environmental gradients responsible for varia-tion in floristic composition of meadow vegetation; (2) to establish the ma-jor vegetation types resulting from numerical classification and to comparethem with the major types recognized in the traditional expert-based classi-fication; (3) to test whether beta-diversity of wet meadows is higher thanof mesic meadows, a pattern which would justify finer differentiation ofwet meadows at the association level, as accepted in phytosociological tra-dition.

30

553Diversity of hay meadows in the Czech Republic

Materials and methods

The basic source of the data were releves of the Molinio-Arrhenathere-tea class from the Czech National Phytosociological Database (Chytry &Rafajova 2003). Releves for the current analysis were selected accordingto their assignment to this class by the original authors. Releves of extremesize (i.e. < 4 m2 or > 100 m2), releves without recorded bryophytes andreleves lacking sufficiently accurate indication of locality were deleted.With respect to geographical coverage of the national territory, some areasappeared to be oversampled while there were gaps in some other areas.Possible negative effects of spatial autocorrelation resulting from thispattern were therefore reduced, although not entirely eliminated, by per-forming a geographically stratified selection of releves from the database.Only one releve of each association (according to original authors assign-ment) per grid square of 1.25 longitudinal 0.75 latitudinal minute (ca.1.5 1.4 km) was selected at random. This selection yielded 3102 releveswhich were used for the analysis. Species cover values recorded on ordinalscales (mostly Braun-Blanquet or Domin) were replaced by percentagesand square-root transformed.

Major gradients in species composition of meadow vegetation were ana-lysed by ordination of this data set, using detrended correspondence analy-sis (DCA) from the CANOCO 4.5 package (ter Braak & Smilauer2002). For ecological interpretation of the ordination axes, average Ellen-berg indicator values (Ellenberg et al. 1992) for releves were plotted ontoDCA ordination diagram as supplementary environmental data.

Classification of the data set was performed by cluster analysis in theprogram PC-ORD 4 (McCune & Mefford 1999), using relative Euclidean(chord) distance as a resemblance measure and flexible beta group linkagemethod with parameter = 0.3. Two classifications were done. The firstclassification used all 3102 releves as input data. In this classification, tenclusters at the highest level of classification hierarchy were accepted, be-cause this number roughly corresponds to the number of alliances tradi-tionally recognized in phytosociological literature. The second classifica-tion was done with 900 releves, including 300 randomly selected relevesfrom each of the three most common alliances, Calthion, Arrhenather-ion and Molinion. Assignment of releves to the alliances followed theexpert opinion of the original authors of these releves. This second classifi-cation was done in order to evaluate validity of the first classification, be-cause the larger data set (3102 releves) contained unequal numbers of rele-ves from different habitats. For example, wet meadows, assigned by theiroriginal authors to the Calthion alliance, were represented by 53 % ofreleves in that data set, and this fact could result in a disproportionatelymore detailed division of the Calthion meadows in the cluster analysisdendrograms of the first classification.

Diagnostic species for the clusters were determined a posteriori, by calcu-lating the fidelity of each species to each cluster, using the phi coefficientof association (Sokal & Rohlf 1995, Chytry et al. 2002) in the program

31


JUICE 6.1 (Tichy 2002). In these calculations, each cluster was comparedwith the rest of the releves in the data set, which were taken as a singleundivided group. In this way, partitioning of the rest of the data set did notinfluence fidelity of species to the target cluster. The threshold value fora species to be considered as diagnostic was set to 0.20. The results of theclassification were summarized in a synoptic table, in which both percen-tage species frequencies (constancies) and values (fidelities) were shown,and diagnostic species were ranked by decreasing fidelity, i.e. by decreasing value (Chytry et al. 2002).

Syntaxonomical interpretation of each cluster in terms of the standardnational vegetation classification of the Czech Republic (Moravec et al.1995) was made, using the list of diagnostic species for alliances of thisclassification, as produced by Chytry & Tichy (2003) on the basis of sta-tistical calculations applied to a large data set extracted from the CzechNational Phytosociological Database. This list represents the collectiveideas of Czech vegetation scientists about delimitation of alliances and pro-vides statistically reliable sets of diagnostic species for each alliance. Diag-nostic species of each cluster were compared with diagnostic species fromthe national list, which enabled interpretation of the clusters in terms ofphytosociological alliances. As different alliances contained different num-bers of diagnostic species and also our clusters included different numbersof diagnostic species, we standardized this comparison by calculatingSrensen similarity index between each group of diagnostic species for aphytosociological alliance and each group of diagnostic species for one ofthe clusters identified in the current analysis:

S = 2a/(2a + b + c),

where a is the number of shared (diagnostic) species, b and c are numbersof species present in one of the two groups of diagnostic species but absentin the other. In this paper, values of the Srensen coefficient were multipliedby 100, thus the range is from 0 to 100.

For further interpretation of the clusters, average Ellenberg indicatorvalues for releves of each cluster were subjected to principal componentsanalysis (PCA from CANOCO 4.5; ter Braak & Smilauer 2002) to showthe ecological relationships among these clusters.

In order to compare beta-diversity between wet and mesic meadows, wedivided the releves into groups according to average Ellenberg moisturevalue. The groups were defined by Ellenberg value intervals 3.04.9, 5.05.9, 6.06.9, 7.07.9, and 8.09.4. The extreme values 3.0 and 9.4 were thelowest and the highest values found in the data set. Releves with values3.04.0 or 9.09.4 were few and therefore they were merged with adjacentcategories. For each of these releve groups, we calculated beta-diversity asthe mean Srensen dissimilarity between all pairs of releves (100 S, whereS is Srensen similarity; Magurran 1988, Koleff et al. 2003), using theJUICE 6.1 program (Tichy 2002). Confidence intervals for beta-diversitywere obtained from 100 bootstrap samples (Efron 1979) taken from relevesof each interval.

32


Nomenclature follows Kubat et al. (2002) for vascular plants, Frey etal. (1995) for bryophytes and Moravec et al. (1995) for syntaxa.

Results

Fig. 1 shows species scatter plot of detrended correspondence analysisbased on individual releves (eigenvalues of the first two axes are 0.514 and0.314). Ellenberg indicator values, plotted a posteriori onto ordination dia-gram, show that the major variation in species composition of the meadowscorresponds to two major gradients, the moisture gradient and the gradientof soil nutrient availability, the latter combined with the soil reaction gra-dient.

Clusters distinguished by the classification of the data set of 3102 relevesare shown in Table 1, along with their diagnostic species. Using the exter-nally defined list of diagnostic species (Chytry & Tichy 2003), there is aclear interpretation for clusters 4, 5, 6 and 7, which in turn represent vegeta-tion of the alliances Molinion, Calthion, Polygono-Triset ion andArrhenatherion (Table 2). Cluster 8 is transitional between the alliancesArrhenatherion, Polygono-Triset ion and Cynosurion. Clusters 2,3, 9 and 10 possess a less clear interpretation in terms of diagnostic species,but all of them are most closely related to the Calthion alliance. They aremainly defined by dominants, including Cirsium rivulare in cluster 2, Cir-sium oleraceum and Carex cespitosa in cluster 3, Scirpus sylvaticus in cluster9 and Filipendula ulmaria in cluster 10. Cluster 1 is characterized by speciesof lowland alluvial meadows, but is poorly characterized in terms of bothdiagnostic and dominant species.

Individual clusters strikingly differ in their relationships to major eco-logical factors. The most important factors, as revealed by principal compo-nents analysis of the mean Ellenberg indicator values and clusters (eigen-values of the first two axes are 0.615 and 0.326), are identical with thoseidentified in DCA ordination of individual releves, i.e. moisture, soil reac-tion and nutrients (Fig. 2). Clusters 10, 9, 2 and 3, i.e. monodominant grass-lands of the Calthion alliance, occupy the wettest sites, while clusters 7and 8, related to the Arrhenatherion alliance are confined to the driestsites. The gradients of soil reaction and nutrients are mutually correlated,with clusters 1, 2, 3, 7 and 10 associated with high values and clusters 4, 5and 6 with low values.

The second cluster analysis with equal numbers of releves originally as-signed to Calthion, Arrhenatherion and Molinion resulted in tenclusters, of which five corresponded to Calthion, three to Arrhenathe-r ion and two to Molinion (results are not shown here).

Beta-diversity varied along the moisture gradient (Fig. 3). Mean Srensendissimilarity was comparatively low for releves with an average Ellenbergmoisture value lower than 6.0, i.e. for releves from mesic sites, and increasedtowards wetter sites.

33

Fig. 1. Detrended correspondence analysis (DCA) ordination diagram of meadow dataset. Species list: Achillea millefolium s. lat., A. ptarmica, Aegopodium podagraria, Agrostiscanina, A. capillaris, A. stolonifera, Ajuga reptans, Alchemilla vulgaris s. lat., Alopecuruspratensis, Anemone nemorosa, Angelica sylvestris, Anthoxanthum odoratum s. lat., An-thriscus sylvestris, Arrhenatherum elatius, Aulacomnium palustre, Avenula pubescens, Bel-lis perennis, Betonica officinalis, Bistorta major, Brachythecium rutabulum, Briza media,Calliergonella cuspidata, Caltha palustris, Campanula patula, C. rotundifolia s. lat., Car-daminopsis halleri, Carex acuta, C. brizoides, C. cespitosa, C. echinata, C. hirta, C. nigra,C. ovalis, C. pallescens, C. panicea, C. rostrata, Centaurea jacea, Cerastium holosteoidesssp. triviale, Chaerophyllum hirsutum, Cirriphyllum piliferum, Cirsium canum, C. hetero-phyllum, C. oleraceum, C. palustre, C. rivulare, Climacium dendroides, Colchicum au-tumnale, Crepis biennis, C. mollis, C. paludosa, Cynosurus cristatus, Dactylis glomerata,Dactylorhiza majalis, Daucus carota ssp. carota, Deschampsia cespitosa, Elytrigia repens,Epilobium palustre, Equisetum arvense, E. fluviatile, E. palustre, E. sylvaticum, Eriopho-rum angustifolium, Festuca ovina, F. pratensis, F. rubra s. lat., Filipendula ulmaria, Galiumboreale ssp. boreale, G. mollugo, G. palustre s. lat., G. uliginosum, G. verum s. lat., Gera-nium palustre, G. pratense, G. sylvaticum, Geum rivale, Glechoma hederacea s. lat., Hera-cleum sphondylium, Holcus mollis, Hypericum maculatum, Juncus acutiflorus, J. articula-tus, J. conglomeratus, J. effusus, J. filiformis, Knautia arvensis s. lat., Leontodon hispidus,Leucanthemum vulgare s. lat., Lotus corniculatus, L. uliginosus, Luzula campestris s. lat.,Lychnis flos-cuculi, Lysimachia nummularia, L. vulgaris, Lythrum salicaria, Mentha ar-vensis, M. longifolia, Molinia caerulea s. lat., Myosotis palustris s. lat., Nardus stricta,Phleum pratense s. lat., Pimpinella major, P. saxifraga, Plagiomnium affine s. lat., Plantagolanceolata, P. media, Poa palustris, P. pratensis s. lat., P. trivialis, Potentilla erecta, Primulaelatior, Prunella vulgaris, Ranunculus acris, R. auricomus s. lat., R. repens, Rhinanthusminor, Rhytidiadelphus squarrosus, Rumex acetosa, Scirpus sylvaticus, Selinum carvifolia,Stellaria graminea, Succisa pratensis, Symphytum officinale s. lat., Taraxacum sect. Rude-ralia, Thymus pulegioides, Trifolium dubium, T. hybridum, T. pratense, T. repens, Trisetumflavescens, Trollius altissimus, Urtica dioica, Valeriana dioica, Veronica chamaedrys s. lat.,Vicia cracca, V. sepium, Viola palustris. 34


Table 1. Synoptic table produced by cluster analysis, with the corresponding dendrogram.Values are percentage frequencies in the left-hand part of the table and fidelities ( valuesmultiplied by 1000) in the right-hand part. Diagnostic species for the clusters (defined asthose with > 0.20) are shaded and ranked by decreasing values, i.e. decreasing fideli-ties to each cluster. Negative values are not shown.

!

" #$ $ % $ & ' ( ! ) ) * *

*

& ! !

35


Table 1. (cont.)

!

" # $" # % &

" " #! !! "

"!

" %" ' ! ' & (

"" '" $

!!

"

36


Table 1. (cont.)

Cluster number 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10Number of relevs 475 268 130 435 410 85 360 384 265 290 475 268 130 435 410 85 360 384 265 290Cardaminopsis halleri 2 1 2 1 2 22 1 21 . 1 . . . . . 151 . 306 . .Thymus pulegioides 1 . . 6 1 . 18 25 . . . . . . . . 178 289 . .Carlina acaulis . 1 . 2 . . 6 16 . . . . . . . . 73 280 . .Euphrasia rostkoviana 1 . 1 7 1 . 6 19 . . . . . 63 . . 27 280 . .Silene dioica 2 1 . 1 1 8 . 13 . 1 . . . . . 67 . 263 . .Hypochaeris radicata 1 . . 3 1 . 5 15 . . . . . . . . 38 257 . .Leontodon autumnalis 7 3 1 9 1 1 5 23 1 . 13 . . 39 . . . 257 . .Potentilla aurea . 1 . 1 . 1 . 8 . . . . . . . . . 252 . .Veronica officinalis 1 1 1 6 2 . 3 16 . . . . . 54 . . . 239 . .Alchemilla vulgaris s. lat. 33 46 40 56 45 65 37 75 23 19 . 14 . 103 11 71 . 237 . .Polygala vulgaris 1 1 . 9 1 . 5 17 . . . . . 97 . . 12 228 . .Dianthus deltoides 1 . . 2 . . 5 12 . . . . . . . . 67 228 . .Festuca rubra s. lat. 43 57 63 81 85 89 69 92 29 16 . . . 155 177 93 51 228 . .Hieracium lachenalii . . . . . . 1 6 . . . . . . . . . 224 . .Thlaspi caerulescens 1 . . . . . 6 9 . . . . . . . . 100 204 . .Ononis spinosa 1 1 1 1 . . 1 8 . . . . . . . . . 200 . .Scirpus sylvaticus 23 56 67 18 49 20 1 1 100 62 . 133 139 . 113 . . . 414 183Epilobium obscurum 2 2 11 . 9 . . . 19 6 . . 67 . 83 . . . 227 20Filipendula ulmaria 33 34 46 29 38 29 5 3 37 99 . . 58 . 40 . . . 23 448Geranium palustre 6 4 18 2 3 . 1 . 5 28 . . 112 . . . . . . 306Lysimachia vulgaris 13 27 10 18 21 4 . 1 31 44 . 84 . . 39 . . . 114 232

Table 2. Comparison of diagnostic species for releve clusters defined in Table 1 (columns)and diagnostic species for phytosociological alliances as defined in the national list (Chy-try & Tichy 2003; rows). The upper part of the table shows numbers of common diagno-stic species; values in brackets next to alliance names are total numbers of diagnosticspecies for particular alliances as given in the national list. The lower part of the tablereports Srensen similarity (multiplied by 100) between groups of diagnostic species foreach of the ten clusters and groups of diagnostic species for each alliance.

Cluster number 1 2 3 4 5 6 7 8 9 10Number of releves 475 268 130 435 410 85 360 384 265 290Total number of diagnostic species 7 20 2 22 25 7 44 37 5 3for cluster

Number of common diagnostic speciesArrhenatherion (48) 0 0 0 4 0 0 29 18 0 0Polygono-Trisetion (10) 0 0 0 0 1 4 0 7 0 0Cynosurion (5) 0 0 0 0 0 0 1 2 0 0Alopecurion (5) 0 0 0 0 0 0 0 0 0 0Calthion (54) 1 2 2 6 17 1 0 3 5 3Cnidion (19) 1 0 0 2 0 0 0 0 0 0Molinion (27) 0 0 0 17 2 0 0 2 0 0

Srensen similarityArrhenatherion 0 0 0 11 0 0 63 42 0 0Polygono-Trisetion 0 0 0 0 6 47 0 30 0 0Cynosurion 0 0 0 0 0 0 4 10 0 0Alopecurion 0 0 0 0 0 0 0 0 0 0Calthion 3 5 7 16 43 3 0 7 17 11Cnidion 8 0 0 10 0 0 0 0 0 0Molinion 0 0 0 69 8 0 0 6 0 0

37


-1.0 1.0

-0.7

0.7

light

temperaturecontinent.

moisture

reaction

nutrients

1

2

3

45

6

7

8

9

10

Fig. 2. Principal components analysis (PCA) of ten releve clusters, based on average Ellen-berg indicator values for each cluster. The clusters are numbered as in Table 1.

3.0-5.0 5.0-6.0 6.0-7.0 7.0-8.0 8.0-9.4

Ellenberg indication value for moisture

58

60

62

64

66

68

70

72

74

76

78

Mea

n S

ren

sen

diss

imila

rity

Fig. 3. Beta-diversity pattern in meadows along the soil moisture gradient. Higher valuesof mean Srensen dissimilarity indicate higher beta-diversity, i.e. a higher mean degree ofchange in species composition among different sites. Boxes and whiskers show mediansand 50% and 95 % percentiles.

38


Discussion

Detrended correspondence analysis of a geographically stratified data setof 3102 releves of Czech hay meadows and mesic pastures revealed mois-ture as the main gradient and nutrients, correlated with soil base status, asa secondary gradient controlling species composition (Fig. 1). This result isin accordance with classical expert-based ordination of meadow types ofCentral Europe, presented in the form of ecograms by Ellenberg(1996), as well as with the results of recent studies based on numericalordination of phytosociological data sets and measured environmental vari-ables (Losvik 1993, Schaffers & Sykora 2002, Hajek & Hajkova 2004).These major gradients identified for Central European meadows are iden-tical with major gradients recognized for Central European forests (El-lenberg 1996, Wohlgemuth et al. 1999). Light availability is a less impor-tant factor, and is negatively correlated with moisture, possibly due to thedevelopment of higher and denser stands of broad-leaved herbs in moisthabitats, which decrease light availability near the soil surface.

Table 3. Syntaxonomical and ecological interpretation of the releve clusters identified bycluster analysis. Cluster numbers correspond to those used in Table 1 and Fig. 2.

Cluster Syntaxonomy Habitatnumber

1 Alopecurion, Cnidion wet meadows of inundated flood-plains of lowland rivers

2 Calthion: Cirsietum rivularis wet meadows of nutrient-rich habi-tats in the Western Carpathians

3 Calthion: Angelico-Cirsietum oleracei wet meadows of nutrient-rich habi-and Caricetum cespitosae tats, mostly in the Bohemian Massif

4 Molinion unmanured intermittently wetmeadows

5 Calthion: Angelico-Cirsietum palus- wet meadows of nutrient-poor habi-tris, Polygono-Cirsietum palustris tats, mostly in the Bohemian Massif

6 Polygono-Trisetion: Polygono- wet meadows of montane belt in theCirsietum heterophylli Bohemian Massif

7 Arrhenatherion (nutrient-rich types) mesic meadows of nutrient-richhabitats at lower altitudes

8 Arrhenatherion (nutrient-poor types), mesic meadows of nutrient-poor habi-Polygono-Trisetion, Cynosurion tats in submontane and montane belt

9 Calthion: Scirpetum sylvatici species-poor wet meadows dominatedby Scirpus sylvaticus

10 Calthion: Filipendulenion unmown wet meadows dominated byFilipendula ulmaria

39


Cluster analysis of our meadow data set (Table 1) more or less repro-duced the alliances traditionally recognized in phytosociological literature(Balatova-Tulackova et al. in Mucina & Maglocky 1985, Oberdorfer1993, Ellmauer & Mucina 1993, Ellmauer 1994, Dierschke 1995, 1997,Blazkova & Balatova in Moravec et al. 1995, Zuidhoff et al. 1996,Kucera & Sumberova 2001). Syntaxonomical interpretations of clustersare presented in Table 3.

Mesic meadows and pastures are included in clusters 7 and 8. Cluster 7comprises a widespread type of nutrient-rich meadows of the Arrhena-therion alliance (association Arrhenatheretum elatior is sensu lato),which is rich in diagnostic species. Cluster 8 combines submontane typesof the Arrhenatherion, often found on nutrient-poor soils, with mesicpastures of the Cynosurion and montane meadows of the Polygono-Trisetion. In the Czech Republic, mesic pastures often contain severalspecies typical of meadows while indicators of grazed habitats are few. Thisis perhaps due to frequent changes of meadows into pastures and vice versa,intermittent abandonment of pastures in rotational grazing systems (Pavluet al. 2003) or due to combined management with one hay-cutting andaftermath grazing (Krahulec et al. 2001). Therefore the boundary betweensubmontane mesic meadows, which lack several thermophilous species ofthe lowland Arrhenatherion, and pastures is rather fuzzy, as evidentfrom the combination of these grassland types into a single cluster.

Drier types of montane meadows of the Polygono-Triset ion are alsoincluded in cluster 8, but wetter types with broad-leaved herbs such asCirsium heterophyllum and Geranium sylvaticum form separate cluster 6.Many releves assigned to cluster 6 were originally assigned to the Poly-gono-Cirs ietum heterophyll i association, which is transitional be-tween the Polygono-Triset ion and Calthion alliances. Although themedium-high Hercynic ranges of the Czech Republic do not harbour manyspecies of high-mountain meadows, which makes them poorer in diagnosticspecies when compared with the meadows of the Alps or the Carpathians(Ellmauer 1994, Kliment 1994, Studer-Ehrensberger 2000), our resultssupport the concept of the separate Polygono-Triset ion alliance in theseHercynic ranges.

There is a single distinct cluster that includes the Molinion alliance(cluster 4). This cluster has several diagnostic species, of which some areshared with Nardus grasslands of the Violion caninae alliance (e.g. Nar-dus stricta, Potentilla erecta, Viola canina, Danthonia decumbens). Ellen-berg values indicate an intermediate position of this cluster on the mois-ture gradient between the mesic meadows of the Arrhenatherion andthe wet meadows of the Calthion. This cluster has also the lowest nutri-ent requirements of all clusters, which is in accordance with the low pro-ductivity of the Molinion meadows. These meadows have been tradi-tionally unmanured, mown only once a year or every second year in Julyor August (Ellenberg 1996, Ellmauer & Mucina 1993, Kucera &Sumberova 2001). Nowadays they are largely abandoned due to their lowhay yields.

40


Unlike the other alliances, Calthion wet meadows were divided amongfive clusters. We suspected that this might be an artifact of the rather highproportion of releves of this alliance included in the data set, which was noteliminated even by the geographically stratified selection of releves prior tothe analysis. Such influence of the data set structure on classification resultsis an inherent property of unsupervised classification methods such as clus-ter analysis (Bruelheide & Chytry 2000, Koc et al. 2003). However, thesecond classification of reduced data set with equal numbers of releves thatwere originally assigned to the alliances Calthion, Arrhenatherion andMolinion, also produced a partition with five of ten clusters correspond-ing to the Calthion. This result suggests that the overrepresentation ofthe Calthion clusters reflects a real pattern existing in the nature ratherthan just the unbalanced structure of our data set. The Calthion clustersrevealed in the analysis (Table 1) correspond to major associations such asCirsietum rivularis (cluster 2), Angel ico-Cirs ietum oleracei andCaricetum cespitosae (cluster 3), Angelico-Cirsietum palustrisand Polygono-Cirs ietum palustr is (cluster 5), Scirpetum sylvatici(cluster 9), and the Fil ipendulenion suball iance (cluster 10).

The high degree of splitting of the Calthion alliance is consistent withexpert knowledge summarized in the Central European phytosociologicalliterature, which recognizes more associations within the Calthion thanin any other alliance of meadow vegetation. Table 4 shows numbers of asso-ciations within different alliances of meadow vegetation, extracted fromnational lists and monographs of vegetation units from wider CentralEurope. Except for Hungary, where the environment is perhaps too dryfor the development of diverse Calthion vegetation (Borhidi 2003),rather high numbers of associations within the Calthion are consistentlydistinguished in all of these publications. Our analysis of beta-diversity inmeadows along the moisture gradient (Fig. 3) is consistent with this trend,showing that meadow vegetation in wet habitats exhibits a higher degreeof change in species composition among different sites. In wet meadows,there are several tall, broad-leaved herbs with a strong competitive ability,namely Cirsium species, Filipendula ulmaria, Scirpus sylvaticus and Carexcespitosa, which become dominants in habitats that correspond to their eco-logical requirements. Once becoming dominants, these species may alterecological conditions within their stands and influence species composition.Consequently, phytosociological classification tends to recognize more as-sociations within wet meadows. By contrast, mesic meadows usually con-tain several co-dominant species, in particular medium-tall grasses, ratherthan a single dominant.

Wet meadows of lowland river floodplains, traditionally assigned to thealliances Alopecurion and Cnidion, were merged in cluster 1. Thispoints out to the high similarity of both alliances, however, it can also bean artifact of a low number of Cnidion releves in our data set (1 % accord-ing to the original authors assignment) and marginal geographical locationof the Czech Republic with respect to the putative geographical distribu-tion of this alliance (Balatova-Tulackova 1969).

41


Table 4. Numbers of vegetations units (associations or association-level communities) dis-tinguished in some alliances of the Molinio-Arrhenatheretea class in Central Europe.

Alli

ance

Arr

hena

ther

ion

Pol

ygon

o-Tr

iset

ion

Phy

teum

o-Tr

iset

ion

Cyn

osur

ion

Cal

thio

n

Fili

pend

ulio

n

Mol

inio

n

Alo

pecu

rion

Cni

dion

Des

cham

psio

n

Ver

onic

olo

ngif

olia

e-Ly

sim

achi

onvu

lgar

is

Reference country/ Arrhenatheretalia Molinietaliaregion

Blazkova & Czech 5 7 5 16 10 7 7 6 3Balatova in RepublicMoravec et al.(1995)

Zuidhoff et al. The 1 2 6 1 2 (1996) Netherlands

Pott (1995) Germany 8 3 3 10 5 8 2

Date post:	01-Feb-2017
Category:	Documents
Upload:	dangtuyen
View:	221 times
Download:	0 times

Variabilita luk svazu Molinion v České a Slovenské republice

Documents