UNIVERSITY OF LATVIA
Liene Salmiņa
Limnogenous mire vegetation of Latvia
Summary of the PhD thesis
for promotion to the degree of Doctor of Biology
(specialization - botany)
Rīga, 2006
Summary
The PhD thesis Limnogenous mire vegetation of Latvia" comprises studies on limnogenous mire communities. The mire plant communities formed in the process of lake terrestrialization including the first succession stage represented by tall-sedge stands and Myricetum gale un Schoenetum ferruginei found in the drained part of lakes in Latvia were considered as limnogenous mire communities in this study.
For data analysis 1582 relevés were used. By using the cluster analysis and manual re-grouping of four groups, twenty end-clusters were created and according to the Central European vegetation classification twenty plant communities belonging to three classes and seven alliances were distinguished. Six new plant associations and two new plant communities for Latvia were described and for the first time in Latvia variants and subassociations were distinguished for seven mire and tall sedge associations. For the first time in Latvia diagnostic species for mire plant communities and alliances and differential species as well were distinguished by means of indicator species analysis. Synoptic tables were used for vegetation data display. The phytogeographical spectrum of limnogenous mires was characterised. The Ellenberg and Dull indicator values for vascular plants and bryophytes were used for plant community ecology analyses. This study also include the distribution analysis of limnogenous mire communities, analysis of rare plant species found in limnogenous mires of Latvia, and the study on factors influencing the distribution of Cladietum marisici in Latvia.
The PhD thesis consists of 140 pages and there are 163 references. The PhD thesis has seven chapters: Summary, Introduction, Materials and methods, Results and discussion, Conclusions, Annexes. There are 24 tables, 21 figures and 21 synoptic tables in the thesis. The thesis was prepared at the Faculty of Biology, Department of Botany and Ecology University of Latvia during the time period from 1997 till 2006. The supervisor was Dr. biol. Māra Pakalne from the Department of Botany and Ecology, Faculty of Biology of University of Latvia.
Aim of the study
The aim of the study is to prepare the syntaxonomical structure of limnogenous
mires of Latvia. To reach the aim the following tasks were set:
•Classify and analyse limnogenous mire vegetation,
•Determine diagnostic species for alliances and associations of limnogenous mires,
•Give the phytosociological description of the limnogenous mire communities,
•Characterise the phytogeographical spectra of limnogenous mires,
•Clarify ecology and distribution of the limnogenous mire communities in Latvia.
Many plant species found in limnogenous mires, including Cladium mariscus, have an uneven distribution in Latvia. Therefore, another task was set:
•Determine factors, influencing the distribution of the association Cladietum
marisci Allorg. 22 in Latvia.
Introduction
The Braun-Blanquet approach lies upon the theory that plant communities can
be distinguished on the basis of species composition, species frequency and quantity in the homogenous vegetation, and that plant communities can be united in the higher classification units following the same principles mentioned above. The diagnostic species characterise all the syntaxonomical units (Braun-Blanquet 1964). Vegetation studies give an opportunity to compare vegetation of different regions, determine distribution area of syntaxa. In Latvia vegetation studies, including mire studies is not completed and the majority of mire vegetation data has not been published. The list of syntaxa of mire vegetation of Latvia includes nine alliances belonging to four classes and 26 associations are distinguished (Bambe 1994, Eņģele 1998, Jermacāne 1998, Salmiņa 1998, Pakalne 1994 a, b, 1998, Pakalne & Čakare 2001, Jermacāne & Laiviņš 2001, Salmiņa 2003, 2005).
Despite the long history of vegetation studies in Europe (Braun - Blanquet 1921, Tüxen 1928, 1930), the need to distinguish the diagnostic species using statistically based data analysis were recognised only recently (Chytry et al. 2002 a, b, Chytry Tichy 2003, Knollova Chytry 2004, Roleček 2005). They were taken from the literature automatically and the data set and the origin of the field studies often have not been taken into the account leading to misunderstanding of the given sintaxa. Indicator species analysis (Dufrēne & Legendre 1997) by means of PC ORD or calculation of Ф-coefficient or u-value (Bruelheide 2000) by means of JUICE is the most common methods used when diagnostic species need to be distinguished. The diagnostic species can differ for one syntaxon located in different regions and they depend on the data set structure and size (Diekmann 1995, Bruun Ejrnaes, 2000, Chytry et al. 2002). Therefore, it is recommended to include in the data analysis also
data from closely related syntaxa and data from other regions when diagnostic species of higher syntaxa are to be distinguished. There are four types of diagnostic species (Dierschke 1994). Local diagnostic species are diagnostic only in a small part of the distribution area of the given syntaxon, regional diagnostic species are diagnostic in one climate region or one physiogeographical region, and there are diagnostic species of many regions, and there are also absolute diagnostic species.
Recently indicator species analysis was used to distinguish diagnostic species for mire communities (Salmiņa 2005) and species fidelity (u-value) to determine diagnostic species of higher syntaxa of natural grasslands of Latvia (Rūsiņa 2005).
Vegetation analysis is often followed by analysis of environmental factors on ground, for example, soil reaction or other soil chemical characteristics. However, it is impossible by large-scale studies, because of the time and finances required for this type of analysis. Therefore, Ellenberg indicator values for vascular plants are often used to characterise ecology of plant communities (Ellenberg et al. 1992, Diekmann Dupré 1997, Dupre Diekmann 1998, Schaffers, Sykora 2000, Exner el al. 2002, Chytry et al. 2003). According to Hawkes and Lawesson (Hawkes et al. 1997, Lawesson et al. 2003), Ellenberg indicator values can only be used in the Central Europe and calibration with measurements on ground is needed in other parts of Europe. And what is more, Wamelink (Wamelink et al. 2002) suggests using Ellenberg indicator values only for comparison of plant communities of one vegetation type. In Latvia Ellenberg indicator values so far have been used to characterise forest and grassland communities (Bambe 2002, Jermacāne 2002, Kreile 2002).
Methods
Phytosociological studies were carried out according to the Braun-Blanquet approach from 1997 until 2006.
For data ordination the detrended correspondence analysis (DCA) by means of PC ORD 4.0 was used (Hill Gauch 1980, McCune Grace 2002). For data ordination species recorded only in one releve were deleted (if n>l, then n=235). Downweighting for rare species was applied. Cluster analysis (Sorensen distance measure, β=-0.25) by means of PC ORD 4.0 was used for releve classification (McCune, Grace 2002).
Cluster analysis and manual re-grouping of four clusters ended in 20 end clusters that were assigned to 20 associations or plant communities according to the Central European vegetation classification approach (Pott 1995, Dierssen 1996, Ellenberg 1996). For seven associations cluster analysis were applied in order to reveal lower syntaxonomical units (Caricetum lasiocarpae, Caricetum rostratae, Caricetum elatae, Chrysohypno-Trichophoretum alpini, Caricetum limosae, Rhynchosporetum albae, Cladietum marisci). Synoptic tables were created using JUICE (Tichy 2001) and were used for vegetation data display.
Indicator species analysis by means of PC ORD 4.0 was used to identify diagnostic species of alliances and plant communities and to distinguish differential species of variants. Indicator value ranges from 0 to 100 (perfect indicator species) (Dufrēne & Lesendre 1997).
N - number of releves in the data set;
Np- number of releves in the given syntaxonomical unit;
n - species frequency in the data set;
np- species frequency in the given syntaxonomical unit.
Species, which indicator values exceeded 10 (IV>10) and had statistically high significance (p>0.001) were considered to be diagnostic species of alliances. Species, which IV-20 were considered to be characteristic species of plant communities, but species, which IV>50 had statistically high significance (p>0.001) were distinguished as differential species of variants. By evaluation of lists of diagnostic species, species ecology in Latvia and presence in other vegetation types, not included in this data set, were also taken into account. Species present in more than 40% of the releves in the syntaxon were distinguished as constant species.
Shannon's' diversity index (H), evenness (E) and the mean number of species in releve (S) were calculated for plant communities. Calculation of Ellenberg indicator values and Ellenberg and Dull indicator values were used to characterise ecology of plant communities (Ellenberg et al. 1992). Range of Ellenberg and Dull indicator values for 20 plant communities were calculated (minimum, maximum, mean, outliers) by means of SPSS 13.0. Spearmen's rank correlation was used in order to clarify what environmental factors according to Ellenberg and Dull indicator values
determine the species number in plant community. The phytogeographical spectra of limnogenous mire communities were analysed using species grouping in zonal, oceanic-continental and sectoral divisions (Meusel et at. 1965, 1978, 1992, Hulten Fries 1986) and modified for species occurring in Latvia by Fatare (Fatare 1986), Species found in more than one zone were considered as polyzonal.
Selected climate parameters, such as mean temperature of July, mean temperature of February, and the number of frost-free days between 45 nature regions with and without Cladium mariscus in Latvia was compared using Mann-Whitney U-test (Sokal Rohlf, 1995).
Nomenclature: vascular plants (Pētersone & Birkmane 1980, Gavrilova & Šulcs 2000), bryophytes (Āboliņa 2001), phytosociological nomenclature and syntaxonomy of plant communities (Pott 1995, Dierssen 1996, Ellenberg 1996, Weber etal. 2000).
Main results and discussion
The study included 82 study objects (Fig.l), the majority of which are located
in the Coastal Lowland (22 lakes), the Central Vidzeme (19), the Northern Vidzeme (14), and South-east Latvia geobotanical regions (11). The distribution of the studied limnogenous mires reflected the lake distribution in Latvia. Most of the studied limnogenous mires are formed due to filing-in or overgrowing of small and shallow lakes (Tab. 1), which are the most common lakes in Latvia.
Table 1
The studied lakes by size
Size, ha / Number %of lakes
0,1 -20
21 - 100
>100 / 60 73,2 13 15,9
9 10,9
Figure 1. The study objects of limnogenous mires
Legends
1 - Papes, 2 - Liepājas, 3 - Mazezers, 4 - Būšnieku, 5 - Pētera, 6 - Dūmezers, 7- Silkalēju, 8 - Garezers, 9 - Klāņu, 10 - Pelcenes, 11 - Vellacs, 12 - Sūnezers, 13 - Asarītis, 14-Vienīts, 15 - Zāļezers, 16 - Baltezers, 17 - Lielauces, 18 - Pušezers, 19 - Kaņieris, 20 - Slokas, 21 - Aklais, 22 -Dūņieris, 23 - Kūdraines, 24 - Engures, 25 - Maku, 26 - Dūņu, 27- Ķīšezers, 28 - Vidus, 29 - Ummis, 30 - Garezers, 31 - Linezers, 32 - Sāls, 33 - Melnezers, 34 - Pideņu, 35 - Ninieris, 36 - Pūrics, 37 -Auciema dzelves, 38 - Tavainis, 39 - Vīņaudu, 40 - Raiskuma, 41 - m. Lagzdiņš, 42 - Oleru, 43 -Raunis, 44 - Aškiņa, 45 - Bezdibenis, 46 - Mellūzis, 47 - Zummers, 48 - Klievezers, 49 - Salainis, 50 - Niedrājs, 51 - Podiņu, 52 - Bābenis, 53 - Kalmodu, 54 - Teļa, 55 - Melnezers (Dziļais Janēlis), 56 -Bednes, 57 - Briežu, 58 - Kūriņu, 59 - Tauns, 60 - Slieķu, 61 - Pāvītes, 62 - L. Plencis, 63 -Graulītis, 64 - Pļaviņas, 65 - Ļaudonītis, 66 - Damenu, 67 - Aizdumbles, 68 - m. Kugru, 69 -Motrines, 70 - Obeļevas, 71 - Dzierkaļu, 72 - Meirauku, 73 - Pelēču, 74 - Asaru, 75 - bez nosaukuma, 76 - Koškina, 77 - Kurtoša, 78 - Mičūnu, 79 - Krugloje, 80 - Skumbiņu, 81 - L. Dolgoje, 82 - Pinta.
1. Ordination and classification
The data set consists of 1528 releves from 82 limnogenous mires of Latvia. The species list included 272 species (203 vascular plant, 66 bryophyte and 3 charophyte species).
The data ordination revealed the heterogeneity of the data set and indicated floristical differences among limnogenous mire communities. And what is more, it indirectly reflected also ecological differences among plant communities. Cluster analysis (Fig. 2) to the level of 15 groups was carried out, as it was the maximum possible number of division of groups. The first division separated releves from transitional mires (Fig. 2 clusters 13-15), fens where at least some Sphagnum are present (Fig. 2 clusters 12) and fens and tall sedge communities (Fig. 2 clusters 1-11). Further divisions of transitional mires followed the dominant Sphagnum species in a moss layer (Sphagnum flexuosum, Sph.fallax and Sph. teres), and further divisions of fens and tall sedge communities followed the dominant bryophyte species or sedge species. Seven of the groups were manually re-grouped by dominant vascular plant species in order to characterise plant communities according to the Central European vegetation classification approach. They were as follows: Carex rostrata group, Carex lasiocarpa, Carex limosa, Carex elata, Trichophorum alpinum, and Rhynchospora alba group. Cluster analysis and manual re-grouping of some clusters yielded to 20 end clusters named according to the dominant vascular plant species of the group.
Figure 2. Results of cluster analysis of 1528 releves
2. Syntaxonomy of limnogenous mire vegetation of Latvia
According to the Central European vegetation classification system (Pott 1995, Dierssen 1982, 1996, Ellenberg 1996), the studied limnogenous mire vegetation belongs to three classes and seven alliances, representing fen, transitional mire, fringe, tall sedge, and shrub vegetation. Ecologically different plant communities dominated by the same vascular plant species were assigned to one syntaxon. Subassociations and variants reflected local ecological differences and in some cases also plant community succession.
Twenty plant communities were assigned to 18 associations, and two plant communities, such as Thelypteris palustris and Eriophorum vaginatum communities were distinguished without association rank (Tab. 2). Eight new plant communities for Latvia were distinguished. They were as follows: Caricetum buxbaumii, Chrysohypno - Trichophoretum alpini, Calletum palustris, Caricetum magellanici, Carici - Menyanthetum, Eleocharitetum quinqueflorae, and Thelypteris palustris community and Eriophorum vaginatum community.
Table 2
Syntaxonomy of limnogenous mire vegetation of Latvia
Class Phragmiti - Magnocaricctea Klika in Klika et Novak 1941
Order Phragmitetalia Koch 1926
Alliance Phragmition communis Koch 1926
Ass. Cladietum marisci Allorge 1922
Subass. Cladietum marisci typicum Pfeifer 1961
Var. typicum
Var. Myrica gale
Var. Thelypteris palustris
Var. Sphagnum
Subass. Cladietum marisci scorpidietosum Segal Westhoff 1969
Var. Scorpidium scorpioides
Var. Schoenus ferrugineus
Alliance Magnocaricion elatae Koch 1926
Ass. Caricetum elatae W. Koch 1926
Subass. Caricetum elatae typicum Jeschke 1963
Subass. Caricetum elatae scorpidietosum (nom.prov.)
Var. Scorpidium scorpioides
Var. Campylium stellatum
Ass. Caricetum paniculatae Wangerin ex von Rochow 1951
Ass. Caricetum gracilis (Almquist 1929) Graebner et Hueck 1931
Ass. Caricetum ripariae (Soó 1928) Balátovā-Tuláčková et al. 1993
Ass. Caricetum distichae (Steffen 1931) Jonas 1933
Ass. Carici-Menyanthetum Sod 1955
Ass. Calletum palustris (Osvald 1923) Van den Berghen 1952
Thelypteris palustris community
Table 2 (continued)
Class Scheuchzerio - Caricetea nigrae (Nordhagen 1936) R. Tx. 1937
Order Scheuchzerietalia palustris Nordhagen 1937
Alliance Rhynchosporion albae W. Koch 1926
Ass. Caricetum limosae Paul 1910 em. Osvald 1923
Subass. Caricetum limosae var. Sphagnum flexuosum
Subass. Caricetum limosae scorpidietosum (пот. prov.)
Ass. Rhynchosporetum albae W. Koch 1926
Subass. Rhynchosporetum albae var. Sphagnum flexuosum
Subass. Rhynchosporetum albae scorpidietosum (пот. prov.) Eriophorum vaginatum community
Alliance Caricion lasiocarpae Van den Berghen in Lebrun et al. 1949
Ass. Caricetum lasiocarpae Osvald 1923 em. Dierssen 1982
Subass. Caricetum lasiocarpae typicum
Subass. Caricetum lasiocarpae scorpidietosum Van den Berghen 1952
Var. Scorpidium scorpioides
Var. Drepanocladus revolvens
Var. Cinclidium stygium
Subass. Caricetum lasiocarpae sphagnetosum Duvigneaud et Van den Berghen 1945
Var. Sphagnum teres
Var. Sphagnum flexuosum
Ass. Caricetum rostratae Osvald 1923 em. Dierssen 1982
Subass. Caricetum rostratae typicum
Var. typicum
Var. Calliergonella cuspidata
Subass. Caricetum rostratae sphagnetosum (пот. prov.)
Var. Sphagnum flexuosum
Var. Sphagnum fallax
Var. Sphagnum squarrosum
Var. Sphagnum riparium
Var. Sphagnum teres
Order Caricetalia nigrae Koch 26 em. Br.-Bl. 1949
Alliance Caricion nigrae Koch 26 em. Klika 1934
Ass. Caricetum magellanicae Osvald 1923
Order Caricetalia davallianae Br.-Bl. 1949
Alliance Caricion davallianae Klika 1934
Ass. Eleocharitetum quinqueflorae Lüdi 1921
Ass. Caricetum buxbaumii Issl. 1932
Ass. Schoenetum ferruginei Du Rietz 1925
Ass. Chrysohypno - Trichophoretum alpini Hadač 1967
Chrysohypno - Trichophoretum alpini var. typicum Chrysohypno - Trichophoretum alpini var. Sphagnum
Class Alnetea glutinosae Br. - Bl. et R. Tx. 1943
Order Alnetalia glutinosae R. Tx. 1937 em. Th. Müller et Görs 1958 Alliance Salicion cinerea Th. Müller et Görs 1958
Ass. Myricetum gale Jonas 1932
3. Description of limnogenous mire vegetation
3.1. Diagnostic species of higher syntaxa
Few species had indicator value higher than 20 indicating that these species
are either present in other alliances or are weakly represented within the given alliance and only part of them are mentioned as diagnostic species of the alliances Magnocaricion elatae, Rhynchosporion albae, Caricion davallianae, and Caricion lasiocarpae in the Central and Northern Europe (Dierssen 1982, Pott 1995, Chytry Tichy 2003, Matuszkiewicz 2005). Mainly they are species, which can be considered as diagnostic species of the syntaxa in a regional scale (Dierschke 1994). The differences between the list of diagnostic species in Latvia and in the Central and Northern Europe can be explained by the differences between the flora of Eastern Europe and the rest of Europe, and by the fact, that before 2000 diagnostic species in Europe were mainly distinguished without using statistically based data analysis. And what is more, the diagnostic species depend on the data set size and structure used in data analysis (Chytry et al. 2002 a, b). Consequently, the list of diagnostic species for mire and tall sedge vegetation alliances of Latvia can be improved by expanding the data set with tall sedge, all mire and wet grassland vegetation of Latvia, and also adding vegetation data of wet forests. However, this data analysis can not be performed at the moment due to lack of vegetation data for Latvia.
Other factors, such as species competition and ecological requirements of species also must be considered by interpretation of results of indicator species analysis. Therefore, despite the fact that Primula farinosa (IV=3.2) and Pinguicula vulgaris (IV=0.4) did not have high indicator values for Caricion davallianae, they were identified as diagnostic species of this alliance, because rich fens of Caricion davallianae are the most important natural habitat for Pinguicula vulgaris and Primula farinosa in Latvia. The low indicator values (IV) of these species were due to the data set structure and the fact that species were also present in communities of Magnocaricion elatae. Filpendula ulmaria, Potentilla reptans, Sphagnum magellanicum, Eriophorum vaginatum, and Andromeda polifolia were not distinguished as diagnostic species of alliances because in Latvia they reach their ecological optimum in other plant communities.