Long-Term Changes in the Estonian Flora and Measures for Conservation

LONG-TERM CHANGES IN THE ESTONIAN FLORA AND MEASURES FOR CONSERVATION

PIKAAJALISED MUUTUSED EESTI TAIMESTIKUS JA FLOORA KAITSEKS SOBIVAD LOODUSKAITSEMEETMED

KAIRE LANNO

A Thesis for applying for the degree of Doctor of Philosophy in Ecology

Väitekiri filosoofiadoktori kraadi taotlemiseks ökoloogia erialal

Tartu 2014

Eesti Maaülikooli doktoritööd

Doctoral Thesis of the Estonian University of Life Sciences

LONG-TERM CHANGES IN THE ESTONIAN FLORA AND MEASURES FOR CONSERVATION

PIKAAJALISED MUUTUSED EESTI TAIMESTIKUS JA FLOORA KAITSEKS SOBIVAD LOODUSKAITSEMEETMED

KAIRE LANNO

A Thesis for applying for the degree of Doctor of Philosophy in Ecology

Väitekiri ﬁ losooﬁ adoktori kraadi taotlemiseks ökoloogia erialal

Tartu 2014 Institute of Agricultural and Environmental Sciences Estonian University of Life Sciences

According to the verdict No 198 of July 14, 2014 the Doctoral Commitee of the Agricultural and Natural Sciences of the Estonian University of Life Sciences has accepted the thesis for the defence of the degree of Doctor in Philosophy in Ecology

Opponent: Dr. Petr Macek Centre for Polar Ecology Faculty of Science, University of South Bohemia

Supervisor: Dr. Marek Sammul Institute of Agricultural and Environmental Sciences Estonian University of Life Sciences

Defence of the thesis: Estonian University of Life Sciences, room 1A5, Kreutzwaldi 5, Tartu, on August 28, 2014, at 12.00.

The English language was edited by Roger Evans, and Estonian by Piret Kruuspere.

Publication of the thesis is supported by the Estonian University of Life Sciences and by the Doctoral School of Earth Sciences and Ecology created under the auspices of the European Social Fund

LIST OF ORIGINAL PUBLICATIONS ...... 7 ABBREVIATIONS ...... 8 1. INTRODUCTION ...... 9 2. REVIEW OF THE LITERATURE ...... 10 2.1. Long-term changes of fl ora ...... 10 2.2. Reproductive success in remnant populations ...... 11 2.3. Restoration of plant populations ...... 11 3. AIMS OF THE STUDY ...... 13 4. MATERIAL AND METHODS ...... 14 4.1. Study of species persistence (I) ...... 14 4.2. Study of declining species of alluvial meadows (II) ...... 15 4.3. Studies of reproductive success and restoration of a declining species Ligularia sibirica (III, IV) ...... 15 5. RESULTS ...... 19 5.1. Persistence of fl ora of different habitat types...... 19 5.2. Persistence of species with different characteristics ...... 20 5.3. The occurrence and characteristics of declining species of alluvial meadows ...... 22 5.4. Reproductive success of the declining species ...... 23 5.5. The effectivity of reinforcement in populations of Ligularia sibirica ...... 24 6. DISCUSSION ...... 27 6.1. Persistence of fl ora of different habitat types and of species with different characteristics ...... 27 6.2. The occurrence of declining species of alluvial meadows ...... 28 6.3. Reproductive success of the declining species ...... 30 6.4. The effectiveness of reinforcement of populations of Ligularia sibirica ...... 31 6.5. Future prospects ...... 32 7. CONCLUSIONS ...... 33 REFERENCES ...... 35 SUMMARY IN ESTONIAN ...... 43 ACKNOWLEDGEMENTS...... 47 PUBLICATIONS ...... 49 CURRICULUM VITAE...... 110 ELULOOKIRJELDUS ...... 112 LIST OF PUBLICATIONS ...... 114

6 LIST OF ORIGINAL PUBLICATIONS

This thesis is based on the following papers, which are referred to by their Roman numerals in the text. The papers are reproduced with kind permission from Springer Science+Business Media B.V.

I Sammul, M., Kull, T., Lanno, K., Otsus, M., Mägi, M., Kana, S. 2008. Habitat preferences and distribution characteristics are indicative of species long-term persistence in the Estonian ﬂ ora. Biodiversity and Conservation 17: 3531-3550.

II Lanno, K., Sammul, M., Luuk, O., Mesipuu, M. 2014. Persistence of declining species on alluvial meadows of Estonia (Submitted).

III Ilves, A., Lanno, K., Sammul, M., Tali, K. 2013. Genetic variability, population size and reproduction potential in Ligularia sibirica (L.) populations in Estonia. Conservation Genetics 14: 661 – 669.

IV Lanno, K., Sammul, M. 2014. The survival of transplants of rare Ligularia sibirica is enhanced by neighbouring plants. Folia Geobotanica 49: 163-173.

The contributions of the authors to the papers: Paper Idea and study Data collection Data Manuscript design analysis preparation IMS, TK, KL MM, TK, SK KL, MS MS, TK, KL, MO, SK II KL, MS OL, KL, MMe MS, KL KL, MS III MS, AI KL, AI AI AI, MS, KL, KT IV MS, KL KL, MS KL KL, MS AI – Aigi Ilves, KT – Kadri Tali, KL – Kaire Lanno, MS – Marek Sammul, MMe – Meeli Mesipuu, MM – Merike Mägi, MO – Merit Otsus, OL – Ott Luuk, SK – Silja Kana, TK – Tiiu Kull

The participation of the author in preparing the listed publications is as follows: Paper I – 40%; Paper II – 80%; Paper III – 25%; Paper IV – 90%

7 ABBREVIATIONS

DCA – Detrended correspondence analysis EIV – Ellenberg indicator value

8 1. INTRODUCTION

Human infl uence, leading to habitat loss and fragmentation, poses a threat to plant species populations and has caused a global decline in plant species richness in the last century (Hobbs and Yates, 2003; Fischer and Lindenmayer, 2007; Brook et al., 2008). Often this infl uence expresses itself in changed land use, including the cessation of traditional management or intensifi cation of agriculture. There have also been considerable changes in land use in Estonia since the end of the 19th century. The largest of which were the restructuring of agriculture during collectivisation following World War II, increased utilisation of hay-fi elds and cessation of traditional management of semi-natural grasslands since the 1950s (Sammul et al., 2000), the draining of wetlands (Paal et al., 1998) and since the 1990s, the intensifi cation of forestry. It may be assumed that these changes have affected the local fl ora.

Considering that changes in human infl uence have been most pronounced in semi-natural, wetland and also in some forest habitats it can be presumed that the distribution of species adapted to ecological conditions prevailing in these habitats has declined the most. Using species distribution data from different periods, this study estimates the persistence of fl ora of different habitat types and the dependence of the persistence on different characteristics of species and habitats. As the loss and decline of populations is often associated with reduced fi tness and genetic diversity (Ellstrand and Elam, 1993) and implies the need for restoration, these aspects are also addressed using an example of a rare species. The results of the study can be applied in planning nature conservation actions and help to evaluate current conservation priorities.

9 2. REVIEW OF THE LITERATURE

2.1. Long-term changes of ﬂ ora

Information about plant species abundance and distribution can be obtained from different sources, which include surveillance records, monitoring data and historical ﬂ oras. If such datasets contain information from various time periods for one region, as some national atlases do (e.g. Preston et al., 2002; Kukk and Kull, 2005), estimations of large- scale changes in ﬂ ora are possible. Depending on the data available, such estimations can be obtained on a scale of the whole country (Smart et al., 2005; Tamis et al., 2005) or of a smaller region (Fischer and Stöcklin 1997; van der Veken et al., 2004; Walker and Preston, 2006; van Calster et al., 2008).

The habitats and species found to be most threatened according to these studies depend on the region and its land-use history. Due to changes in land use and increased nutrient deposition, the area of nutrient-poor calcareous grasslands has declined considerably in Western Europe, and particularly since the 1940s and 1950-s (e.g. Ratcliffe, 1984). The same applies to species-rich dry to mesic semi-natural grasslands in Sweden (e.g. Bernes, 1994). Eutrophication and land conversion have also caused a greater decline of species in wet/aquatic, arable and saline habitats compared to other habitats in regions across Europe during the 20th century (McCollin et al., 2000; Tamis et al., 2005; Stehlik et al. 2007; van Calster et al., 2008).

The species most often found declining in abundance or losing populations in these studies are stress tolerators, adapted to nutrient- poor soils (McCollin et al., 2000; van der Veken et al., 2004; Smart et al., 2005; Tamis et al., 2005; Bennie et al., 2006; Piessens and Hermy 2006; Walker and Preston, 2006; Stehlik et al., 2007; van Calster et al., 2008; Ozinga et al., 2009), and species with shorter life spans (Fischer and Stöcklin 1997; Stehlik et al., 2007; Hannus and von Numers 2010; Saar et al., 2012). Decline or local extinction is also more likely for rare species and for habitat specialists (Fischer and Stöcklin, 1997; McCollin et al., 2000; van der Veken et al., 2004; Lavergne et al., 2006; Walker and Preston, 2006; van Calster et al., 2008; Peter et al., 2009; Hannus and von Numers, 2010). The declining species are usually preserved in 10 habitats that are still of good ecological quality, i.e. they have remained in a similar ecological condition, under which they have developed for a long period of time (e.g. Aavik et al., 2008). Sometimes however, the loss of long-lived perennial species does not occur until some time after the quality and/or quantity of a habitat has declined, leaving an impression of long-term maintenance of species richness (the effect is called extinction debt: Tilman et al., 1994; Helm et al., 2006; Piessens and Hermy, 2006).

2.2. Reproductive success in remnant populations

Population decline can have a strong negative effect on plant fi tness. Small remnant populations often enter an extinction vortex where population decline leads to a reduced seed production and a drop in seed quality, and hence to a decreased germination success (e.g. Vergeer et al., 2003; Dostálek et al., 2010), which then may lead to a further decline in a population size. The decline in reproductive success can be related to a decline in genetic variability within a population. It has been found, that smaller populations are often genetically less diverse (Fischer and Matthies, 1998; Hensen and Oberprieler, 2005; Lienert et al., 2002), but not always (Kuss et al., 2008; Klank et al., 2012). This relationship can be affected by whether or not a species is self-compatible and other factors (Leimu et al., 2006), such as increased genetic drift and inbreeding with a declining population size and increased isolation between populations (Ellstrand and Elam, 1993; Hobbs and Yates, 2003). Since fi tness and genetic diversity are shown to not always be infl uenced by population size, but may be affected by other factors, most notably habitat quality, the effect of population size on reproductive success of declining species needs to be tested before choosing appropriate conservation measures.

2.3. Restoration of plant populations

If the populations of species have declined or even become extinct, then various restoration programmes may be undertaken, for example reintroduction, introduction or reinforcement, depending on the conservation goal and situation with a population as well as habitat conditions (e.g. Bottin et al., 2007; Hölzel et al., 2012). Where the problem lies in a decline in population size reinforcement may be used, which

11 means addition of individuals to the population, aiming to increase only population size or also the population’s viability by increasing its genetic variability (Falk et al., 1996). Depending on species, habitat type etc., restoration can be launched in many different ways either by sowing or planting, using wild collected or off site propagated material and possibly applying different habitat preparation methods (Milton et al., 1999; Drayton and Primack, 2000; Bottin et al., 2007; Guerrant and Kaye, 2007; Menges, 2008; Godefroid et al., 2011). The removal of neighbours is one common habitat preparation method and in several studies it has been shown that survival of transplants is reduced by competition from neighbouring plants (van der Wal et al., 2000; Midoko-Iponga et al., 2005; Kull et al., 2011). However, for some species shelter from neighbouring plants may be essential (e.g. Ryser, 1993; Callaway et al., 2002). In restoration projects, different methods are often used in combination, but to gain information factors inﬂ uencing population dynamics and restoration success the most, their effects should be studied separately.

12 3. AIMS OF THE STUDY

The study addresses the decline in species distributions and corresponding conservation measures on three levels: the whole country, one habitat type and the species level. Based on the Atlas of Estonian Flora, the changes in Estonian fl ora are assessed, estimating the persistence of species of different habitat types and of species with different characteristics. Based on changes occurred in land use and information from other works, we hypothesise, that the persistence of fl ora of semi-natural and wet habitats is the most affected (Bernes, 1994; McCollin et al., 2000; Stehlik et al. 2007) and that persistence of species depends on their characteristics related to competitive ability and habitat preferences (van der Veken et al., 2004; Smart et al., 2005; Walker and Preston, 2006). The work then concentrates on one habitat type, alluvial meadows, and its declining1 species, as this is an example of a habitat decreased in area because of changed human infl uence. We expect that declining species of alluvial meadows are characterised by small statue and strong habitat specifi city and that such species have remained on larger and managed (grazed or mown) meadows. Thirdly, reproduction success and restoration issues of one declining species inhabiting alluvial meadows, Ligularia sibirica, are addressed. We expect, that the decline in population size causes the decrease in reproduction success and that the effectiveness of reinforcement depends on habitat quality and interspecifi c interactions.

The following speciﬁ c questions are addressed:

1. What is the persistence of the ﬂ ora of different habitat types in Estonia (Paper I)? 2. Is the long-term persistence of species dependent on species characteristics related to habitat preference and distribution (Paper I)? 3. Is the occurrence of declining species related to site quality in alluvial meadows (Paper II)? 4. Has the reproduction success of the declining species Ligularia sibirica decreased as a result of the decline of the population size (Paper III)? 5. How effective is the reinforcement of declining plant species and how does its success depend on habitat characteristics (Paper IV)?

1 Declining species hereafter denote species whose distribution in Estonia has decreased according to the Atlas of the Estonian Flora (Kukk and Kull, 2005). 13 4. MATERIAL AND METHODS

4.1. Study of species persistence (I)

Species persistence was estimated using the Atlas of Estonian Flora (Kukk and Kull, 2005), comparing the distribution ranges of native terrestrial ﬂ ora in 2004 with the distribution ranges before 1970 in quadrats of the Atlas (each about 100 km2 in size). The study comprises data of 1031 species. For each species the persistence value was calculated as the percentage ratio of the number of quadrats occupied by the species between 1970 and 2004 from the total number of quadrats occupied by the species in the Atlas.

Using various literature sources (see Appendix 1 in Paper I), the typical ﬂ ora of different habitat types in Estonia as well as the data about different characteristics related to species habitat preference and distribution were compiled (see details in Paper I). One-way analysis of variance was used to test for differences in the mean persistence between species belonging to different groups of species commonness, occurrence at the border of the global distribution range, main habitat type, sensitivity to human impact, life-form, conservation category and Red List category and correlation analysis was used to test the relationship between persistence and Ellenberg indicator values (EIVs) of species requirements.

Based on their persistence value, species were divided into three groups (with persistence value >70%, 40-70% and <40%, respectively) and it was tested whether the proportion of species belonging to each group in the ﬂ ora of certain habitat type differed between different habitat types (see Paper I for details), using one-way analysis of variance.

For the ﬂ ora of each habitat type the weighted average of persistence value (PV) was calculated as follows:

PVh = (∑ PVi * Wih) / ∑ Wih

Here PVi is the persistence of species i and Wih is the importance weight of species i in habitat h (an importance weight of 2 was given to dominant and characteristic species of a habitat and of 1 to other species in that habitat type). The difference in weighted average of

14 persistence between different groups of habitat types (mires, outcrops, dunes and sandy plains, coastal habitats, grasslands, shrublands, forests, other forest-related habitats) was tested and the relationship between the weighted average of persistence of species in a habitat and the number of species in a ﬂ ora of this habitat was tested with correlation analysis.

4.2. Study of declining species of alluvial meadows (II)

In this study the persistence of species of alluvial meadows and present- day information about ecological conditions of alluvial meadows (including their species composition) were used. The species persistence data for the species belonging to the ﬂ ora of alluvial meadows, as well as information about species traits was acquired from Paper I. The species with persistence value of ≤70% (see 4.1 and Paper I) were treated as declining species.

Data about alluvial meadows, their characteristics and species composition was retrieved from the Database of Estonian Seminatural Communities. Altogether, the extracted data contained information about 859 alluvial meadows. Based on the species lists of the sites the occurrence or absence of declining species in a particular meadow was determined. The proportion of declining species and total number of species were calculated for every site. Additionally, for every site the average EIVs for light, moisture, reaction and nitrogen were calculated. The difference between EIVs and average height of the species between the group of declining species and other species was tested with one- way ANOVA. A general linear model was built to explain variation in the proportion of declining species on a site assuming binomial distribution of the dependent variable and using total number of species per site as a weight as suggested for analysis of proportions (Wilson and Hardy, 2002). Detrended correspondence analysis (DCA) was used to describe the variation in vegetation composition.

4.3. Studies of reproductive success and restoration of a declining species Ligularia sibirica (III, IV)

Ligularia sibirica (Asteraceae) was used as a model of a declining species in order to evaluate the relationships between size of population,

15 reproductive success and genetic diversity of a population. This is a rare and endangered species in Europe, growing mainly in alluvial and paludiﬁ ed grasslands and in spring fens. In Estonia, its distribution has decreased considerably, both in number and size of populations (Kukk, 2003), the main reason being a decline in habitat quality caused primarily by drainage and cessation of mowing and grazing, leading to changed light and soil conditions. Thus we also evaluated the effect of reinforcement in populations of Ligularia and estimated the effect of habitat degradation on the success of restoration efforts.

The reproductive success and genetic diversity were studied in seven populations differing largely in the number of individuals of Ligularia (Fig. 1). Seeds were collected in 2007 and 2008 to estimate the number of viable seeds per fl ower stem and germination rate, the latter was estimated using seeds collected in 2007. For genetic analysis the leaves were collected from populations in 2009 and the amplifi ed fragment length polymorphism (AFLP) technique was used (see Paper III for technical details of the AFLP analysis). Within populations, genetic diversity was estimated as average gene diversity over loci. Differences in the amount of viable seeds between populations and years were analysed using two-way analysis of variance and the correlation was used to test the relationship between population size, genetic diversity and plant fi tness estimates.

16 Fig. 1. Locations of Ligularia sibirica populations studied (Õ – Õisu, T – Tagula, P – Pressi, A – Anne, S – Sootaga, V – Väägvere, J – Jõhvi). The reproductive success and genetic diversity was studied in all of those populations and a reinforcement study was conducted in four populations (Sootaga, Väägvere, Tagula and Õisu).

The study estimating the effectiveness of reinforcement and inﬂ uence of overgrowing and competition on transplants of Ligularia was conducted in four populations (Fig. 1). In 2008, laboratory grown transplants from seeds collected the previous year from those four populations were planted back to their original populations in plots arranged in a two × two (vegetation intact or removed × open or overgrown habitat) factorial experimental design. Open habitats were representative of good and typical conditions of the habitat. The overgrown areas, with the exception of the presence of trees and shrubs, were as similar to the open areas in site conditions as possible and were situated immediately adjacent to the open areas. Eight plots 0.25 m2 in size were randomly placed within these habitats. To test the competitive effect of neighbouring plants, vegetation was manually removed from half of the plots in both habitats. There were four replicate plots per treatment and 16 plots in total, with four plants per plot (except in Õisu population, where not enough seedlings emerged and three plants were used per plot). The survival of plants was followed for three years. The proportion of surviving plants in each plot 17 was calculated on every monitoring occasion and arcsine square root transformed and repeated measures analysis of variance was used to test for the effect of removal of neighbouring vegetation, overgrowing and population on the survival of plants.

18 5. RESULTS

5.1. Persistence of ﬂ ora of different habitat types

The frequency distribution of plant species persistence shows, that 11.6% of species have persisted in less than 40% of their original quadrats (1.8% of species have a persistence value lower than 10%), 25.4% of species have persistence value between 40-70% and for almost half of the species in the Estonian ﬂ ora the distribution range has remained within 80% of the original range (Fig. 1 in Paper I).

The weighted averages of persistence were calculated for 46 habitat types, belonging to ten broader habitat groups (Appendix 2 in Paper I). The average persistence was the lowest in ﬂ ora of shrublands (Corylus avellana habitats), and the highest in dry boreal forests. At the level of broader groups of habitat types, the ﬂ ora of bedrock outcrops and shrublands showed the lowest average persistence (69.6% and 70.1%, respectively) and forest habitat types the highest (76.6%).

In Estonian fl ora, the proportion of species with persistence over 70% was signifi cantly smaller in mire habitat types than in forest habitat types (p<0.01) (Fig. 2). In mire habitat types there was a signifi cantly higher number of species with persistence values between 40% and 70% than in forest habitat types (p<0.001) or in other forest-related habitats (p<0.02). Species with persistence under 40% tended to be most abundant in cliff habitat types and in sandy and dune habitats, although this difference was not signifi cant.

The species richness of the ﬂ ora of a habitat type was negatively correlated with the weighted average of persistence of species from a habitat type (Fig. 5 in Paper I).

19 Fig. 2. Proportion of species with different persistence values in various groups of habitat types.

5.2. Persistence of species with different characteristics

Correlations between species persistence and EIVs of species ecological requirements showed that light-demanding species were less persistent than shade-tolerant species, and species preferring alkaline soils and nutrient-rich habitats were more persistent than species of more acidic habitats and nutrient-poor conditions (see Table 4 and Fig. 3 in Paper I).

All the factors studied, with the exception of endemism affected species persistence (Table 1, see also Tables 1, 2 and 3 in Paper I). Species of mires had signifi cantly lower mean persistence than species of forests and grasslands (p<0.002 and p<0.016, respectively). Although bedrock outcrop or dune species also had a low average persistence, they did not differ signifi cantly from other groups due to high within-group variation. Comparison of species with different levels of habitat specifi city 20 revealed that habitat specialists had a signifi cantly lower persistence than habitat generalists.

Table 1. The effect of different species characteristics on species persistence.

Factor d.f F - value p

Main habitat type 7 3.319 0.0017

Sensitivity to human impact 3 29.054 <0.001

Endemism 1 0.135 0.71

Border of the distribution range 1 90.309 <0.001

Life-form 4 6.972 0.00002

Habitat speciﬁ ty 1 143.04 <0.001

Commonness 6 84.106 <0.001

Conservation category 3 16.012 <0.0001

Red List category 5 35.462 <0.001

Analysis of species tolerance to human inﬂ uence showed that hemerophob species and anthropophytes showed the greatest decrease in range and apophytes had the highest persistence. Species at the border of their distribution range were less persistent compared to species well within their distribution range and from various life-form groups phanerophytes were the most and therophytes the least persistent.

The persistence of species belonging to different groups of commonness, conservation and Red List categories also differed signiﬁ cantly and the results showed that persistence declined with increasing level of rarity, the most strictly protected species showed lower persistence and of the Red List categories, species with “indeterminate” status were the least persistent (see Table 3 and Fig. 2 in Paper I).

21 5.3. The occurrence and characteristics of declining species of alluvial meadows

The declining species had higher EIVs for light (p<0.02) and moisture (p<0.001) compared to other species on alluvial meadows (Fig. 1 in Paper II). These species groups did not differ in their EIVs for soil reaction, nitrogen and also in average height of the species.

The results of the general linear model (Table 2) show that the proportion of declining species occurring on a site was greater in larger alluvial meadows. There were also more declining species in sites with higher light and moisture availability and on more alkaline soil types. The nonlinear relationship with nitrogen availability indicated the occurrence of the lowest proportion of declining species in meadows at medium values of nitrogen EIVs.

Table 2. The ﬁ nal general linear model of the effect of characteristics of the meadow on the proportion of declining species. The light, soil moisture, soil reaction and nitrogen are Ellenberg indicator values calculated for the sites. Binomial distribution of the proportion of declining species was deﬁ ned in the model and total species number found on a meadow was used as a weight.

Factor Estimate Std. error z value Pr(>|z|) (Intercept) -13.546 2.246 -6.032 <0.001 Area 0.266 0.072 3.724 <0.001 Light 1.150 0.242 4.740 <0.001 Soil moisture 0.442 0.057 7.690 <0.001 Soil reaction 0.361 0.110 3.301 <0.001 Nitrogen -1.576 0.551 -2.862 0.004 Nitrogen2 0.152 0.054 2.831 <0.005

The results of the DCA ordination showed, that there is no differentiation of vegetation between sites containing declining species and those without (Fig. 2 in Paper II). The ﬁ rst DCA axis was positively correlated with the total number of species, but also with presence of recent mowing and negatively with EIV for soil moisture. The second DCA axis was positively correlated with EIVs for nitrogen and soil reaction,

22 and negatively correlated with EIV for light and with the area of the meadow.

5.4. Reproductive success of the declining species

The production of viable seeds differed signiﬁ cantly between populations of the declining species Ligularia sibirica (p<0.001) and was at its lowest in the smallest population (Fig. 3; Table 1 in Paper III). There was also a difference between years (p<0.001), resulting from seeds being largely damaged and eaten in three populations in 2007 but much less so in 2008. There was no correlation between mean viable seed production and population size (Fig. 4 a, b). The germination percentage varied between 41% and 73% and correlated positively with the population size (p<0.001) (Fig. 4 c). There was also a positive relationship between the population size and genetic variability measured as average gene diversity over loci (Fig. 4 d).

Fig. 3. Mean number of viable seeds (with standard error) in the studied populations of Ligularia in 2007 and 2008. Tagula and Sootaga are large, Pressi, Väägvere and Õisu medium-sized, and Anne and Jõhvi small populations.

23 Fig. 4. The relationship between population size and (a) mean amount of viable seeds per ﬂ ower stem in 2007; (b) mean amount of viable seeds per ﬂ ower stem in 2008; (c) germination %; and (d) average gene diversity over loci in seven populations of Ligularia sibirica.

5.5. The effectivity of reinforcement in populations of Ligularia sibirica

The survival of transplants of the declining species Ligularia sibirica was signiﬁ cantly affected by the removal of neighbouring vegetation, overgrowing, time, and it also differed among populations (Fig 5, see also Table 2 in Paper IV).

24 Fig. 5. The dynamics of survival of transplants of Ligularia sibirica in the four populations in Estonia (Sootaga, Väägvere, Tagula, Õisu) in intact plots (squares) and plots with vegetation removed (rhombs) in open (a) and overgrown habitats (b). In x-axes is time (1 - 2008 Jun, 2 – 2008 Aug, 3 – 2009 Jun, 4 – 2009 Aug, 5 – 2010 Jun, 6 – 2010 Aug). The vertical bars denote 0.95 conﬁ dence intervals.

25 Mortality was up to 20% higher in plots where the vegetation was previously removed compared to intact plots, when data from all populations was pooled. As the interaction of removal with time shows (p<0.0001), the described effect appeared in the second year and decreased slightly in the third year (Fig. 5). The survival was also inﬂ uenced by the interaction between removal, time and population (p<0.0001), revealing the lower mortality in intact plots in all populations except Väägvere (although according to the post-hoc test, this difference was not signiﬁ cant).

Comparison of open and overgrown habitats revealed a higher survival rate in open habitats throughout the duration of the experiment. The interaction between time, population and overgrowing was also statistically signiﬁ cant (p<0.001). Although survival in both habitats was separated in each population, the higher survival in open habitats was apparent only in Väägvere on the ﬁ nal monitoring occasion.

In overgrown habitats at the end of the third year, the survival of transplants was similarly low in all plots irrespective of removal treatment in most populations, so it was clear that in open habitats, plant survival was higher in intact plots than in plots where neighbours had been removed.

26 6. DISCUSSION

6.1. Persistence of ﬂ ora of different habitat types and of species with different characteristics

The persistence of species of different habitats can be compared to changes occurring in the area of those habitat groups. Since 1950 the area of grasslands has decreased by almost 90%, and of mires by two- thirds (mainly through the loss of fens) (Fig. 6 in Paper I). At the same time the area of forests has nearly doubled. The results show a somewhat greater persistence of grassland species (73.1%) than of mire species (72.4%), despite the larger decline in grassland area. The ﬂ ora of forests has decreased the least (76.6%).

The decrease in the area of grasslands is often caused by abandonment which is followed by overgrowth, and hence the change in ecological conditions is not as quick as for mires, where the drainage or partial destruction of the area leads to rapid change in the quality of the whole habitat area (Fojt and Harding, 1995; Minkkinen et al., 1999). Mire species also have a high habitat specifi city and they cannot fi nd refuge in some other locations such as fi eld boundaries or road verges, as can some grassland species (Smart et al., 2002; Huhta and Rautio, 2007). However, the relative persistence of species in grasslands may result from the delay before the species become extinct i.e. there is the extinction debt (e.g. Helm et al., 2006; Kuussaari et al., 2009). In the case of forests, although the fl ora of forests (and shade-tolerant species overall) has declined the least, there is an evident decline in the fl ora of certain forest types such as fresh boreal forests and boreal heath forests, mainly caused by high cutting pressure and scarcity of high-value, old-growth forests (Lõhmus et al., 2004). The higher persistence of forest species comes mainly from the high persistence of species belonging to dry boreal forests and fl oodplain willow shrublands.

The average persistence turned out to be lower in habitats that contain more species. This may be because species with low persistence are more likely to inhabit more diverse habitats or it could be a sign of the greater effect of eutrophication on nutrient poor species-rich communities on calcareous soils. The latter is also supported by the fact that nutrient- demanding species have higher persistence (Fig. 3 in Paper I). A similar 27 negative effect of eutrophication on diversity has been noted in a number of studies in other regions (e.g. Tamis et al., 2005; Walker and Preston, 2006; van Calster et al. 2008). Nutrient enrichment affects grasslands directly because of the fertilisation, but may also occur in other habitats through atmospheric deposition (e.g. Bobbink et al., 1998).

It has been observed that rare species as well as habitat specialists are more likely to become extinct than generalist species (e.g. Fischer and Stöcklin, 1997; McCollin et al., 2000; Lavergne et al., 2006; Hannus and von Numers, 2010). This was also the case in our study, where persistence decreased with decreasing species commonness and was smaller for habitat specialists compared to generalists, thus emphasising the need for concentrating conservation actions on sustaining habitats providing speciﬁ c environmental conditions for such species. Persistence was also lower for species at the border of their distribution range, presumably because of unsuitable climatic conditions for those species.

Of the species with different life-forms, therophytes have declined the most. This ﬁ nding is in agreement with the results of several other studies which showed that the species with short life-span are more endangered (Fischer and Stöcklin, 1997; Stehlik et al., 2007; Saar et al., 2012). Comparison of species with different sensitivity to human impact revealed a bigger decline in hemerophobs, showing the impact of an increase in negative human activities such as drainage and eutrophication, but also anthropophytes, which show the impact of a decrease in positive human activities such as the management of grasslands.

The comparison of species in different legal protection and Red List categories shows that high priority is given to the species that have declined the most, as they belong to the strictest categories and so the problem is being addressed, although it does not ensure the persistence of species.

6.2. The occurrence of declining species of alluvial meadows

The results revealed, that declining species were more light-demanding and needed higher soil moisture content than other species in alluvial meadows. This conﬁ rms the negative effects of overgrowing and drainage on species of such wet habitats (e.g. Truus and Tõnisson, 1998).

28 These results are also in accordance with several studies that have found a decline in light demanding species (e.g. Hannus and von Numers, 2010), resulting from loss of mowing or grazing and subsequent competitive exclusion. Often short statured species are more prone to extinction (e.g. Saar et al., 2012), but this study did not ﬁ nd any difference in the height of declining species compared to other species. It might be presumed that dependence on light availability can be mediated by germination if litter and shade are preventing the germination (e.g. Huhta et al., 2001). In this case the typical height of the species is not important.

Sites which have a higher probability of containing declining species tended to be larger, and this could be expected, because large habitats tend to be more heterogeneous, providing different microhabitats for species. Additionally sites with a higher proportion of declining species tended to have higher moisture and light availability, neutral or calcareous soils and either low or high, but not medium productivity. These traits point towards the importance of local variability of environmental conditions and the need to preserve natural ﬂ ooding regime which is the key factor creating local environments and variability. There is however a discrepancy between the declining species needing wetter conditions whereas continuity of alluvial meadows depends on mowing, which cannot take place, if the habitat is too wet. Contemporary mowing with large machines also serves as a factor in homogenizing habitats (destroying hummocks, always cutting at the same height, etc). Thus, it could be that grazing may suit the needs of declining species better, even though there was no effect of presence of grazing on the probability of ﬁ nding the declining species.

The results of the DCA showed that the species composition of alluvial meadows varied on the one hand on an axis of the total number of species (and mowing) vs. moisture regime, and on the other hand mostly on an axis of productivity and soil reaction vs. light availability and area. However, the sites containing species with declining distribution were distributed randomly across the ordination plane and did not differentiate from other sites. This could indicate that the occurrence of declining species is not predictable based on species composition of the meadows.

29 It is important to preserve the habitats which for alluvial meadows usually mean the regular mowing and removal of hay. However, our study indicates that this standard view is not sufﬁ cient for the preservation of species with declining distribution (mowing for example did not explain the existence of declining species). Thus, the conservation of declining species of alluvial meadows may need some additional species-speciﬁ c measures.

6.3. Reproductive success of the declining species

The results with the declining species Ligularia sibirica showed the negative effect of a small population size on the germination percentage, indicating the decreased fi tness of the populations due to the decrease in population size. The number of viable seeds produced did not reveal any positive correlation with population size, differing between populations and years. However, in the smallest population seed production was similarly low in both years of the study. The declined population size is often related to reduced genetic variation (Lienert et al., 2002; Hensen and Oberprieler, 2005), which may be the cause of the decline in plant fi tness. However, some studies have not found any relationship between population size and genetic variation (e.g. Leimu and Mutikainen, 2005; Honnay et al., 2007; Kuss et al., 2008) and the small size of the population does not automatically mean low fi tness and genetic variability.

As the results showed a positive relationship between population size and within-population genetic diversity, the reduced genetic diversity seemst to be one cause of impaired reproductive output. The genetic diversity was relatively low (slightly lower than the average for mixed- mating species according to Nybom (2004)), and indicates the existence of self-pollination or inbreeding in populations of Ligularia. Based on the fl uctuations in viable seed production between populations and years it can be presumed that while in large populations the plants can produce more seeds in good years, the small populations with low genetic variability may not be able to do so. In bad years seed production seems to be low in both large and small populations, irrespective of genetic diversity. On the one hand, unsuitable environmental factors certainly play a role in low reproductive fi tness in smallest populations (see also paper IV), but there can also be a relationship between genetic factors and reproductive fi tness of this species.

30 Consideration of such reproductive and genetic aspects is important when planning conservation actions like reinforcement of species. Although it is very important to retain the genetic material of individual populations, the use of local plant material may not be sufﬁ cient, if the small isolated populations are genetically impoverished and this is a cause of reduced ﬁ tness.

6.4. The effectiveness of reinforcement of populations of Ligularia sibirica

Our hypothesis was that reinforcement is enhanced by removal of neighbouring vegetation around transplants of Ligularia and it is less effective in overgrown areas. It has been shown, that the effect of reducing competition can vary by species and site (e.g. Kaye and Brandt, 2005). We found, however, that the survival of transplants was lower in plots where the neighbouring plants were removed. This can be explained by the facilitative effect of neighbours offering shelter against animals (the damage from animals has also been described for example in McKendrick (1995) and Jusaitis (2005)) and preventing soil erosion. This does not mean that the role of competition between Ligularia and other species is minor. Seedlings are especially vulnerable to competition as shown for example in Harper (1977), Grace and Tilman (1990). The positive effect of neighbours however in the case of Ligularia outweigh the negative effect resulting from competition. The inﬂ uence of such positive effects and the balance between facilitation and competition has also been discussed in different reviews (e.g. Callaway and Walker, 1997; Brooker et al., 2008).

The effect of overgrowing on transplant survival was tested because although Ligularia prefers semi-open spots and open areas (Kukk, 2003), in some cases Ligularia also grows under a relatively dense forest and shrub cover, in other cases it suffers severely in such overgrown areas. The results showed that in most populations the survival was indeed lower in overgrown areas. In addition, under the tree cover the mortality was high irrespective of having neighbours or not, stressing the importance of restoring or preserving habitat quality. The overgrowing changes environmental conditions, most notably those of light and moisture. These changed conditions primarily harm seedlings and juvenile plants, whereas older individuals may resist some time in spite of unsuitable

31 conditions, causing the time delay in extinction and extinction debt as described above.

In case of such projects, there are different criteria for measuring the success of restoration (Pavlik, 1996; Menges, 2008), for example the start of ﬂ owering and sexual reproduction. In our study, ﬂ owering was rare after three years of the study. Therefore it is advisable to monitor the plants until the viability of seeds and seedlings can be assessed. The longer monitoring period is also necessary because the short-term survival can be misleading (e.g. Kaye, 2002), especially for species with long life spans.

It can be concluded that the effectiveness of reinforcement of this declining species is limited, and it can only be used if habitat quality is good. In this case, the addition of individuals may primarily help to go through the early critical life stages of species. Additional planting site preparation for this species is not needed, however the need for some preparation method depends on species and also can differ between the populations of the same species.

6.5. Future prospects

The study assessed the causes of decline in ﬂ ora of Estonia from different aspects and treated the issue at different levels, describing the changes in the distribution of species, but also in reproductive success of populations and the effectiveness of species restoration. However, there are several issues where conclusive evidence was not obtained and further studies are required. For example the study showed that species at the border of their distribution range had low persistence in the Estonian ﬂ ora. Further analysis of species range shifts would be informative in the context of global climate change and species protection. Secondly, declining species occurrence could also be studied in some other habitat types in addition to alluvial meadows. Moreover, the dynamics of vegetation and populations of individual (rare, declining or threatened) species in restored alluvial meadows deserves special attention in order to estimate the success of large-scale restorations currently taking place in Estonia.

32 7. CONCLUSIONS

This study assessed the changes in plant species distribution and respective conservation measures on various scales ranging from broad estimations at the country level down to the level of species, dealing with the reproductive success and restoration of the populations. The results of the studies show the negative effect on species distribution of, on the one hand increased human activities (eutrophication, drainage) and on the other, the cessation of management of semi-natural habitats. Since the decline of species at different scales is caused by different reasons, these reasons have to be determined for applying appropriate conservation measures. It appears that the methods for preservation of specifi c habitat types might not fully coincide with the specifi c needs of individual species. Thus conservation planning needs to address such cases and evaluate what is an appropriate balance between preservation of habitat types in general and the threatened fl ora of these habitats.

Based on the results of this thesis the following conclusions can be drawn: The fl ora of mire habitats has declined more than the fl ora of forests and grasslands in the second half of the 20th century in Estonia. Considering the large simultaneous decrease in the area of grasslands this could indicate the existence of an extinction debt in grassland communities. There were also large differences in persistence of different habitat types inside those broader habitat groups. The persistence of species of species-rich habitats was lower than persistence of the species in habitats containing less species. Habitat specialists have declined more compared to habitat generalists, emphasising the need to preserve specifi c environmental conditions. Lower persistence was also found for species at the border of their distribution range, hemerophobs and therophytes. Species of alluvial meadows with declining distribution had higher preference for light and moisture than other species and the proportion of declining species depended on the area of the meadow, light availability, moisture conditions, soil reaction and productivity of the site. At the same time, the occurrence

33 of declining species is not predictable based on the species composition of the meadows studied. The reproductive success, as well as genetic variability, of the declining species studied (Ligularia sibirica) was lower in smaller populations, indicating that genetic factors can affect plant ﬁ tness of this species besides environmental factors. Reinforcement of populations of declining species may be effective if the habitat quality is preserved or restored. The effect of additional site preparation methods such as removal of competitors are described to depend on species and site. Ligularia does not need such efforts; instead the neighbouring plants provide shelter for it. Considering the small genetic diversity and low reproductive success of small populations of Ligularia the use of only local plant material for reinforcement may not be sufﬁ cient.

34 REFERENCES

Aavik, T., Jõgar, Ü., Liira, J., Tulva, I., Zobel, M. 2008. Plant diversity in a calcareous wooded meadow – the signifi cance of management continuity. Journal of Vegetation Science 19: 475-484. Bennie, J., Hill, M.O., Baxter, R., Huntley, B. 2006. Infl uence of slope and aspect on long-term vegetation change in British chalk grasslands. Journal of Ecology 94: 355-368. Bernes, C. (ed). 1994. Biological diversity of Sweden – a country study. Swedish Environmental Protection Agency, Monitor 14. Bobbink, R., Hornung, M., Roelofs, J.G.M. 1998. The effects of air-borne nitrogen pollutants on species diversity in natural and semi-natural European vegetation. Journal of Ecology 86: 717-738. Bottin, L., Le Cadre, S., Quilichini, A., Bardin, P., Moret, J., Machon, N. 2007. Re-establishment trials in endangered plants: a review and the example of Arenaria grandifl ora, a species on the brink of the extinction in the Parisian region (France). Ecoscience 14: 410-419. Brook, B.W., Sodhi, N.S., Bradshaw, C.J.A. 2008. Synergies among extinction drivers under global change. Trends in Ecology and Evolution 23: 453-460. Brooker, R.W., Maestre, F.T., Callaway, R.M., Lortie, C.L., Cavieres, L.A., Kunstler, G., Liancourt, P., Tielbörger, K., Travis, J.M.J., Anthelme, F., Armas, C., Coll, L., Corcket, E., Delzon, S., Forey, E., Kikvidze, Z., Olofsson, J., Pugnaire, F., Quiroz, C.L., Saccone, P., Schiffers, K., Seifan, M., Touzard, B., Michalet, R. 2008. Facilitation in plant communities: the past, the present, and the future. Journal of Ecology 96: 18-34. Callaway, R.M., Walker, L.R. 1997. Competition and facilitation: a synthetic approach to interactions in plant communities. Ecology 78: 1958-1965. Callaway, R.M., Brooker, R.W., Choler, P., Kikvidze, Z., Lortie, C.J., Michalet, R., Paolini, L., Pugnaire, F.I., Newingham, B., Aschehoug, E.T., Armas, C., Kikodze, D., Cook, B.J. 2002. Positive interactions among alpine plants increase with stress. Nature 417: 844-848.

35 Dostálek, T., Münzbergova, Z., Plačková, I. 2010. Genetic diversity and its effect on fi tness in an endangered plant species, Dracocephalum austriacum L. Conservation Genetics 11: 773-783. Drayton, B., Primack, R.B. 2000. Rates of success in the reintroduction by four methods of several perennial plant species in eastern Massachusetts. Rhodora 102: 299-331. Ellstrand, N.C., Elam, D.R. 1993. Population genetic consequences of small population size: Implications for plant conservation. Annual Review of Ecology and Systematics 24: 217-242. Falk, D.A., Millar, C.I., Olwell, M. 1996. Restoring diversity. In Falk, D.A., Millar, C.I., Olwell, M. (eds). Strategies for reintroduction of endangered plants. Island Press, Washington DC, pp. XIII-XXII. Fischer, J., Lindenmayer, D.B. 2007. Landscape modifi cation and habitat fragmentation: a synthesis. Global Ecology and Biogeography 16: 265-280. Fischer, M., Matthies, D. 1998. RAPD variation in relation to population size and plant fi tness in the rare Gentianella germanica (Gentianaceae). American Journal of Botany 85: 811-819. Fischer, M., Stöcklin, J. 1997. Local extinctions of plants in remnants of extensively used calcareous grasslands 1950-1985. Fojt, W., Harding, M. 1995. Thirty years of change in the vegetation communities of three valley mires in Suffolk, England. Journal of Applied Ecology 32: 561-577. Godefroid, S., Piazza, C., Rossi, G., Buord, S., Stevens, A.-D., Aguraiuja, R., Cowell, C., Weekley, C.W., Vogg, G., Iriondo, J.M., Johnson, I., Dixon, B., Gordon, D., Magnanon, S., Valentin, B., Bjureke, K., Koopman, R., Vicens, M., Virevaire, M., Vanderborght, T. 2011. How successful are plant species reintroductions? Biological Conservation 144: 672-682. Grace, J.B., Tilman, D. 1990. Perspectives in plant competition. Academic Press, San Diego. Guerrant, E.O., Kaye, T.N. 2007. Reintroduction of rare and endangered plants: common factors, questions and approaches. Australian Journal of Botany 55: 362-370.

36 Hannus, J.-J., von Numers, M. 2010. Temporal changes in the island fl ora at different scales in the archipelago of SW Finland. Applied Vegetation Science 13: 531-545. Harper, J. 1977. The population biology of plants. Academic Press, London.. Helm, A., Hanski, I., Pärtel, M. 2006. Slow response of plant species richness to habitat loss and fragmentation. Ecology Letters 9: 72-77. Hensen, I., Oberprieler, C. 2005. Effects of population size on genetic diversity and seed production in the rare Dictamnus albus (Rutaceae) in central Germany. Conservation Genetics 6: 63-73 Hobbs, R.J., Yates, C.J. 2003. Impacts of ecosystem fragmentation on plant populations: generalising the idiosyncratic. Australian Journal of Botany 51: 471-488. Honnay, O., Adriens, D., Coart, E., Jacquemyn, H., Roldan-Ruiz, I. 2007. Genetic diversity within and between remnant populations of the endangered calcareous grassland plant Globularia bisnagarica. Conservation Genetics 8: 293-303. Huhta, A.-P., Rautio, P., Tuomi, J., Laine, K. 2001. Restorative mowing on an abandoned semi-natural meadow: short-term and predicted long-term effects. Journal of Vegetation Science 12: 677-686. Huhta, A.-P., Rautio, P. 2007. A case with blue gentian blues: roadside- cutters creating neo grasslands as refugia for endangered Gentianella campestris. Nordic Journal of Botany 25: 372-379. Hölzel, N., Buisson, E., Dutoit, T. 2012. Species introduction – a major topic in vegetation restoration. Applied vegetation Science 15: 161-165. Jusaitis, M. 2005. Translocation trials confi rm specifi c factors affecting the establishment of three endangered plant species. Ecological Management & Restoration 60: 61-67.

37 Kaye, T.N. 2002. Reintroducing the endangered pink Sand-verbena to Pacifi c Coast beaches: direct seeding and out-planting. In: Brooks, M.B., Carothers, S.K., LaBanca, T. (eds). The ecology and management of rare plants of northwestern California. Proceedings from Symposium North Coast Chapter California Native Plant Society, California Native Plant Society, Sacramento, pp. 131-140. Kaye, T.N., Brandt, A. 2005. Seeding and transplanting rare WillametteValley prairie plants for population restoration. Technical report prepared for the Eugene District, USDI Bureau of Land Management and Institute of Applied Ecology, Corvalis, Oregon. Klank, C., Ghazoul, J., Pluess, A.R. 2012. Genetic variation and plant performance in fragmented populations of globefl owers (Trollius europaeus) within agricultural landscapes. Conservation Genetics 13: 873-884. Kukk, Ü. 2003. The distribution of Ligularia sibirica (L.) Cass. In Estonia and changes in its population. Biuletyn Ogrodów Botanicznych 12: 11-22. Kukk, T., Kull, T. (eds). 2005. Atlas of the Estonian fl ora. Institute of Agricultural and Environmental Sciences of the Estonian University of Life Sciences, Tartu. [In Estonian]. Kull, T., Kull, T., Sammul, M. 2011. Reduced light availability and increased competition diminish the reproductive success of wet forest sedge Carex loliacea L. Plant Species Biology 26: 84-92. Kuss, P., Pluess, A.R., Ægisdóttir, H.H., Stöcklin, J. 2008. Spatial isolation and genetic differentiation in naturally fragmented plant populations of the Swiss Alps. Journal of Plant Ecology 1: 149-159. Kuussaari, M., Bommarco, R., Heikkinen, R., Helm, A., Krauss, J., Lindborg, R., Öckinger, E., Pärtel, M., Pino, J., Rodá, F., Stefanescu, C., Teder, T., Zobel, M., Steffen-Dewenter, I. 2009. Extinction debt: a challenge for biodiversity conservation. Trends in Ecology and Evolution 24: 564-571.

38 Lavergne, S., Molina, J., Debussche, M. 2006. Fingerprints of environmental change on the rare mediterranean fl ora: a 115- year study. Global Change Biology 12: 1466-1478. Leimu, R., Mutikainen, P. 2005. Population history, mating system, and fi tness variation in a perennial herb with a fragmented distribution. Conservation Biology 19: 349-356. Leimu, R., Mutikainen, P., Koricheva, J., Fischer, M. 2006. How general are positive relationships between plant population size, fi tness and genetic variation? Journal of Ecology 94: 942-952. Lienert, J., Fischer, M., Schneller, J., Diemer, M. 2002. Isozyme variability of the wetland specialist Swertia perennis (Gentianaceae) in relation to habitat size, isolation, and plant fi tness. American Journal of Botany 89: 801-811. Lõhmus, A., Kohv, K., Palo, A., Viilma, K. 2004. Loss of old-growth, and the minimum need for strictly protected forests in Estonia. Ecological Bulletins 51: 401-411. McCollin, D., Moore, L., Sparks, T. 2000. The fl ora of a cultural landscape: environmental determinants of change revealed using archival sources. Biological Conservation 92: 249-263. McKendrick, S.L. 1995. The effects of herbivory and vegetation on laboratory-raised Dactylorhiza praetermissa (Orchidaceae) planted into grassland in southern England. Biological Conservation 73: 215-220. Menges, E.S. 2008. Restoration demography and genetics of plants: when is a translocation successful? Australian Journal of Botany 56: 187-196. Midoko-Iponga, D., Krug, C., Milton, S.J. 2005. Competition and herbivory infl uence growth and survival of shrubs on old fi elds: implications for restoration of renosterveld shrubland. Journal of Vegetation Science 16: 685-692. Milton, S.J., Bond, W.J., du Plessis, M.A., Gibbs, D., Hilton-Taylor, C., Linder, H.P., Raitt, L., Wood, J., Donaldson, J.S. 1999. A protocol for plant conservation by translocation in threatened lowland Fynbos. Conservation Biology 13: 735-743.

39 Minkkinen, K., Vasander, H., Jauhiainen, S., Karsisto, M., Laine, J. 1999. Post-drainage changes in vegetation composition and carbon balance in Lakkasuo mire, Central Finland. Plant and Soil 207: 107-120. Nybom, H. 2004. Comparison of different nuclear DNA markers for estimating intraspecifi c genetic diversity in plants. Molecular Ecology 13: 1143-1155. Ozinga, W.A., Römermann, C., Bekker, R.M., Prinzing, A., Tamis, W.L.M., Schaminée, J.H.J., Hennekens, S.M., Thompson, K., Poschlod, P., Kleyer, M., Bakker, J.P., van Groenendael, J.M. 2009. Dispersal failure contributes to plant losses in NW Europe. Ecology Letters 12: 66-74. Paal, J., Ilomets, M., Fremstad, E., Moen, A., Borset, E., Kuusemets, V., Truus, L., Leibak, E. 1998. Estonian Wetlands Inventory 1997. Publication of the project “Estonian Wetlands Conservation and Management”. Eesti Loodusfoto, Tartu, 166 pp. Pavlik, B.M. 1996. Defi ning and measuring success. In: Falk, D.A., Millar, C.I., Olwell, M. (eds). Strategies for reintroduction of endangered plants. Island Press, Washington DC, pp. 127-155. Peter, M., Gigon, A., Edwards, P.J., Lüscher, A. 2009. Changes over three decades in the fl oristic composition of nutrient-poor grasslands in the Swiss Alps. Biodiversity and Conservation 18: 547-567. Piessens, K., Hermy, M. 2006. Does the heathland fl ora in north-western Belgium show an extinction debt? Biological Conservation 132: 382-394. Preston, C.D., Pearman, D.A., Dines, T.D. 2002. New atlas of the British and Irish fl ora. Oxford University Press, Oxford. Ratcliffe, D.A. 1984. Post-medieval and recent changes in British vegetation: the culmination of human infl uence. New Phytologist 98: 73-100. Ryser, P. 1993. Infl uences of neighbouring plants on seedling establishment in limestone grassland. Journal of Vegetation Science 4: 195-202.

40 Saar, L., Takkis, K., Pärtel, M., Helm, A. 2012. Which plant traits predict species loss in calcareous grasslands with extinction debt? Diversity and Distributions 18: 808-817. Sammul, M., Kull, K., Kukk, T., 2000. Natural grasslands in Estonia: evolution, environment and economic roles. In: Viiralt, R., Lillak, R., Michelson, M. (eds). Conventional and ecological grassland management. Estonian Grassland Society, Tartu. Smart, S.M., Bunce, R.G.H., Firbank, L.G., Coward, P. 2002. Do fi eld boundaries act as refugia for grassland plant species diversity in intensively managed agricultural landscapes in Britain? Agriculture, Ecosystems and Environment 91: 73-87. Smart, S.M., Bunce, R.G.H., Marrs, R., LeDuc, M., Firbank, L.G., Maskell, L.C., Scott, W.A., Thompson, K., Walker, K.J. 2005. Large-scale changes in the abundance of common higher plant species across Britain between 1978, 1990 and 1998 as a consequence of human activity: tests of hypothesised changes in trait representation. Biological Conservation 124: 355-371. Stehlik, I., Caspersen, J.P., Wirth, L., Holderegger, R. 2007. Floral free fall in the Swiss lowlands: environmental determinants of local plant extinction in a peri-urban landscape. Journal of Ecology 95: 734-744. Tamis, W.L.M., van`t Zelfde, M., van der Meijden, R., Groen, C.L.G., Udo de Haes, H.A. 2005. Ecological interpretation of changes in the dutch fl ora in the 20th century. Biological Conservation 125: 211-224. Tilman, D., May, R.M., Lehman, C.L., Nowak, M.A. 1994. Habitat destruction and the extinction debt. Nature 371: 65-66. Truus, L., Tõnisson, A. 1998. The ecology of fl oodplain grasslands in Estonia. In: Joyce, C.B., Wade, P.M. (eds). European wet grasslands: biodiversity, management and restoration. Wiley, New York, NY, pp. 49–60. Van Calster, H., Vandenberghe, R., Ruysen, M., Verheyen, K., Hermy, M., Decocq, G. 2008. Unexpectedly high 20th century fl oristic losses in a rural landscape in northern France. Journal of Ecology 96: 927-936.

41 Van der Wal, R., Egas, M., van der Veen, A., Bakker, J. 2000. Effects of resource competition and herbivory on plant performance along a natural productivity gradient. Journal of Ecology 88: 317-330. Van der Veken, S., Verheyen, K., Hermy, M. 2004. Plant species loss in an urban area (Turnhout, Belgium) from 1880 to 1999 and its environmental determinants. Flora 199: 516-523. Vergeer, P., Rengelink, R., Copal, A., Ouborg, N.J. 2003. The interacting effects of genetic variation, habitat quality and population size on performance of Succisa pratensis. Journal of Ecology 91: 18-26. Walker, K.J., Preston, C.D. 2006. Ecological predictors of extinction risk in the flora of lowland England, UK. Biodiversity and Conservation 15: 1913-1942. Wilson, K., Hardy, I.C.W. 2002. Statistical analysis of sex ratios: an introduction. In: Hardy, I.C.W. (ed). Sex ratios: Concepts and research methods. Cambridge University Press, Cambridge, pp. 48-92.

42 SUMMARY IN ESTONIAN

PIKAAJALISED MUUTUSED EESTI TAIMESTIKUS JA FLOORA KAITSEKS SOBIVAD LOODUSKAITSEMEETMED

Üha suurenev inimtegevuse mõju, mis põhjustab liikide kasvukohtade kadu või nende killustumist, on viinud erinevate taimeliikide leviku vähenemiseni. Kui mingi piirkonna kohta on olemas liikide levikuinfo erinevatest ajaperioodidest, siis on võimalik toimunud muutuste põhjuseid hinnata. Nii on mitmed Euroopa erinevates piirkondades tehtud uuringud näidanud, et 20. sajandi jooksul on enam vähenenud näiteks toitainetevaestele kasvukohtadele omased taimeliigid.

Liikide leviku ja arvukuse täpsemate muutuste ning sobivate looduskaitsemeetmete väljaselgitamiseks on lisaks suures mastaabis toimuvatele muutustele oluline teada ka muutuste põhjuseid väiksemal, näiteks liigi või populatsiooni tasemel. Viimane aitab kindlaks teha muid liike ohustavaid tegureid, nagu näiteks paljunemisedukuse vähenemine või geneetilise mitmekesisuse kadu. Antud töö uuriski liikide levikut ja looduskaitsemeetmeid kolmes erinevas mastaabis: terve Eesti, ühe kasvukohatüübi ja liigi tasandil. Antud töö uurimisküsimusteks olid: 1) milline on erinevate kasvukohatüüpide taimeliikide püsivus Eestis?; 2) kas liikide püsivus sõltub erinevatest taimeliikide kasvukohaeelistuste ja leviku tunnustest?; 3) kas väheneva levikuga taimeliikide levik lamminiitudel on seotud niitude kvaliteediga?; 4) kas väheneva leviku ja arvukusega taimeliigi Ligularia sibirica paljunemisedukus on populatsiooni suuruse vähenemise tõttu langenud?; 5) kui efektiivne on vähenenud taimeliigi tugiasustamine ja kuidas selle edukus sõltub kasvukoha omadustest?

Üle-Eestilise taimede leviku muutumise uurimiseks kasutati Eesti taimede levikuatlast, mis sisaldab infot liikide leviku kohta kahest erinevast ajaperioodist, 1970. aasta ja 2004. aasta seisuga. Liikidele arvutatud püsivusväärtuste põhjal (see näitab, kui suures osakaalus levikuatlase ruutudes on liik praegusel ajal säilinud võrreldes 1970. aasta seisuga) hinnati ühelt poolt erinevate kasvukohatüüpide ﬂ oora püsivust ja teiselt poolt seda, kas liikide püsivus sõltub erinevatest liigi kasvukohaeelistuse ja levikuga seotud tunnustest.

43 Erinevate tunnustega liikide vähenemist analüüsides selgus, et enam on vähenenud liigid, mis on valgusnõudlikumad, samuti aluselistel muldadel ja toitainetevaestes tingimustes kasvavad liigid. Liikide püsivus sõltus ka sellest, milline on liigi põhikasvukoht (rohkem olid vähenenud liigid, mille põhikasvukohaks on kaljud või sood), tundlikkusest inimmõju suhtes (enam olid vähenenud hemerofoobid, kuid ka antropofüüdid) ning eluvormist (enim on vähenenud terofüüdid). Suuremal määral on vähenenud ka elupaigaspetsialistid, Eestis levikupiiril olevad liigid ning liigid, mis on haruldased, rangemas kaitsekategoorias ning ohustatud või määratlemata staatusega Punases nimestikus. Tulemused näitasid, et erinevate kasvukohatüüpide ﬂ oora püsivus on erinev, jäädes vahemikku 62%–91%. Lisaks näitasid tulemused, et mida liigirikkam on kasvukohatüüp, seda enam on selle liigid vähenenud, mis võib olla märk toitainerikkaid kasvukohti eelistavate liikide suuremast püsivusest ning toitainetevaeste liigirikaste kasvukohtade eutrofeerumisest. Kõrvutades tulemusi muutustega, mis on toimunud kasvukohatüüpide pindalas alates 1950. aastast, ilmnes, et soode liigid on rohkem vähenenud kui niitude liigid, kuigi niitude pindala on kahanenud suuremal määral. Metsade pindala on suurenenud ning metsade liigid on ka suurima püsivusega. Samas on metsade puhul samuti tunduvalt vähenenud teatud kasvukohatüüpide, näiteks laanemetsade ﬂ oora.

Kasutades nimetatud uurimusest pärinevat lamminiitude ﬂ oora liikide nimekirja koos nende püsivustunnustega ning lamminiitude andmeid Eesti pärandkoosluste andmebaasist, analüüsiti, milliste omadustega on need lamminiidud, kus vähenevad liigid (püsivusväärtusega ≤70%) on säilinud, ja millised on nende liikide tunnused. Selgus, et vähenevad liigid on võrreldes muude lamminiitudel kasvavate liikidega suurema valgus- ja niiskusnõudlusega. Vähenevate liikide osakaal oli suurem nendel niitudel, mis olid suuremad, paremate valgustingimustega ja niiskemad, neutraalsete ja/või aluseliste muldadega. Samas näitas ordinatsioonanalüüs, et niidud, kus vähenenud liigid esinesid, ei eristunud muudest niitudest taimestiku poolest – seega pole vähenevate liikide esinemine taimestiku koosseisu järgi lihtsasti ennustatav.

Üksikute liikide puhul võivad nende vähenemise põhjuseks olla lisaks halvenenud keskkonnateguritele ka muud tegurid, nagu näiteks paljunemisedukuse ja/või geneetilise mitmekesisuse vähenemine. Seetõttu tuleks olulisematele või haruldasematele/ohustatumatele

44 liikidele kaitsemeetmete väljatöötamiseks nende vähenemise põhjused eraldi kindlaks teha. Antud töös uuritigi üht niiskete kasvukohtade ohustatud liiki, harilikku kobarpead (Ligularia sibirica), mille nii populatsioonide arv kui ka suurus on Eestis vähenenud. Täpsemalt analüüsiti liigi paljunemisedukust, geneetilist mitmekesisust ja tugiasustamise efektiivsust. Paljunemisedukuse uuring viidi läbi kahel aastal, millest üks oli kobarpea jaoks soodne, kuid teine mitte. Kui suurtes populatsioonides oli heal aastal kobarpeal palju seemneid ja halval aastal vähe, siis väiksemates populatsioonides oli kobarpea seemnete hulk väike mõlemal uuringuaastal, st ka muidu soodsal aastal. See tähendab, et populatsioonis ei teki enam piisavalt juveniilseid taimi ning seda ähvardab hääbumine. Seemnete idanemisprotsent oli positiivses korrelatsioonis populatsiooni suurusega võimendades veelgi paljunemisedukuse erinevust suurte ja väikeste populatsioonide vahel. Madalam paljunemisedukus väikestes populatsioonides võib olla seotud vähenenud geneetilise mitmekesisusega, mida toetab tulemus, et populatsiooni geneetiline mitmekesisus oli positiivses korrelatsioonis populatsiooni suurusega.

Sama liigi neljas populatsioonis uuriti ka populatsiooni tugiasustamise efektiivsust ja selle sõltuvust kasvukoha omadustest. Täpsemalt uuriti kasvukoha kinnikasvamise ning naabertaimede hajusa konkurentsi mõju. Selleks istutati laboris ettekasvatatud taimed loodusesse ja jälgiti nende ellujäämist. Taimedest olid pooled istutatud kasvukoha avatud ning pooled kinnikasvanud osasse ning mõlemal juhul olid osadel ruutudel naabertaimed eemaldatud ning osadel mitte. Istutatud taimede suremus oli oluliselt suurem neis ruutudes, kust naabrid olid eemaldatud (mis näitab, et naabertaimed pakkusid kaitset ja varju), ja kinnikasvanud kasvukohas – see rõhutab omakorda kasvukoha kvaliteedi säilitamise või taastamise vajadust tugiasustamise efektiivsuse jaoks.

Kokkuvõtteks võib töö tulemustest järeldada, et toimunud muutused näitavad ühelt poolt inimmõju suurenemise (kuivendamine, eutrofeerumine) negatiivset mõju, teisalt ka selle puudumise ehk koosluste hooldamata jätmise negatiivset mõju. Vähenevate liikide esinemine kindlas kasvukohatüübis seostub mõningate kasvukoha kvaliteeti näitavate tunnustega, eelkõige on tähtis loodusliku niiskusrežiimi säilitamine. Samas näiteks oodatud majandamise positiivset mõju vähenevate liikide leidumisele tuvastada ei õnnestunud.

45 Kui eesmärgiks on kindlate liikide kaitse, siis peaks nende vähenemise põhjuseid eraldi vaatama (sest neis võib rolli mängida ka näiteks vähenenud paljunemisedukus ning geneetiline mitmekesisus) ja kaitseks lisaks tavapärastele koosluse hooldusvõtetele võib-olla rakendama muid lisameetmeid. Sellised meetmed, nagu näiteks tugiasustamine, võivad olla efektiivsed, kuid eeltingimuseks on kasvukoha sobiv kvaliteet; ning väga väikestes populatsioonides ei pruugi kohaliku geneetilise materjali kasutamine olla piisav populatsiooni elujõu piisavaks suurendamiseks.

46 ACKNOWLEDGEMENTS

First of all I would like to express my gratitude to my supervisor Marek Sammul. Marek, thank you for your support, help and constructive guidance through all these years.

I warmly thank all my colleagues in the Department of Botany of Institute of Agricultural and Environmental Sciences for great working atmosphere. Especially Karin, Silja, Vivika, Kaili and Lauri – thank you for your company during work and in also in free time. I also thank all the co-authors of the papers and the people helping in ﬁ eldwork.

I am thankful to Aveliina Helm for being the opponent of my predefence and for her comments about my thesis.

I would like to thank Piret Kruuspere for improving Estonian and Roger Evans for improving English.

I also thank my family for their support and patience during all these years. And Indrek – thank you for your support, faith and for reminding me, which things are really important and which are not.

The studies were supported by grants no 6048, 7567, 7631 and 8745 from Estonian Science Foundation, the EU FP6 project EuMon, target ﬁ nancing project no 0362481s03, grant no P6062PKPK06 from Estonian University of Life Sciences and institutional research funding IUT (21-1) of the Estonian Ministry of Education and Research.

PUBLICATIONS Sammul, M., Kull, T., Lanno, K., Otsus, M., Mägi, M., Kana, S. 2008. Habitat preferences and distribution characteristics are indicative of species long-term persistence in the Estonian ﬂ ora. Biodiversity and Conservation 17: 3531-3550. Biodivers Conserv (2008) 17:3531–3550 DOI 10.1007/s10531-008-9393-5

ORIGINAL PAPER

Habitat preferences and distribution characteristics are indicative of species long-term persistence in the Estonian ﬂora

Marek Sammul Æ Tiiu Kull Æ Kaire Lanno Æ Merit Otsus Æ Merike Ma¨gi Æ Silja Kana

Received: 5 October 2007 / Accepted: 1 April 2008 / Published online: 18 April 2008 Ó Springer Science+Business Media B.V. 2008

Abstract Large-scale changes in regional floras provide direct information about changes in biodiversity through time and enable the evaluation of conservation targets. We compared the distribution ranges in 2004 of Estonian native terrestrial flora with the distribution ranges before 1970, using the Atlas of Estonian Flora. Relative persistence was related to species endemism, commonness, occurrence at its border of the global distribution range, main habitat type, sensitivity to human impact, life-form, conservation category, and Red List category. A literature-based database of the flora of Estonian habitat types was used to evaluate relative persistence of the flora of different habitats. Changes in the flora are largely dependent on human activities. The decrease in mire and grassland habitats and the increase in forests are reflected in the persistences of related species. Flora of mire habitats decreased the most. The fact that an almost ten-fold decrease of grasslands has not resulted in as large a decrease in the ranges of grassland species could serve as evidence of the extinction debt of these habitats. We also found a greater decrease among habitat specialists than habitat generalists and lower average persistence of the species of species-rich habitats. Our data show that current prioritization of species for conservation is in concordance with needs, as reflected in the changes in the range of species. However, conservation has not been entirely successful: the decrease of protected species continues. Our simple method for summarizing large databases was effective for the evaluation of large scale effects of conservation actions.

Keywords National ﬂora Distribution ranges Eutrophication Land-use change Large-scale changes Long-term changes Monitoring Species persistence

Introduction

There is general acknowledgment of large scale decline in global species diversity (Myers 1979; Wilson 1992; May et al. 1995; Brooks et al. 2002 etc). While the main reason for

M. Sammul (&) T. Kull K. Lanno M. Otsus M. Ma¨gi S. Kana Department of Botany, Institute of Agricultural and Environmental Sciences, Estonian University of Life Sciences, Riia str 181, Tartu 51014, Estonia e-mail: [email protected]

51 123 3532 Biodivers Conserv (2008) 17:3531–3550 such mass extinction is human impact, the causes can range from extirpation of populations to the effects of pollution and climate change. However, the main factor causing current species loss is the destruction of habitats (Soule´ 1987; Tilman et al. 1994; Fischer and Sto¨cklin 1997; Riis and Sand-Jensen 2001 etc.), driven mostly by changing practices of land-use (including abandonment, and intensification of use of natural resources) (Chapin et al. 2000; Sala et al. 2000). Habitat loss is often accompanied by other changes such as fragmentation and isolation of the remaining habitat patches. These processes have negatively influenced many plant populations (Henle et al. 2004; Honnay et al. 2005), and greater negative effects of the change in quality and quantity of habitat on specialist species than on generalists would be expected (Dupre´ and Ehrleń 2002). There may be a threshold for the area of habitat occupied by a species below which the probability of its extinction increases drastically (Tilman et al. 1994; Hanski 2000; Hanski and Ovaskainen 2002). Moreover, some long- lived perennial species can persist despite declining population size for a relatively long time. This is sometimes responsible for the apparent long-term maintenance of species richness even after quality and/or quantity of habitat has fallen (Eriksson 2000; Hedin 2003; Ikonen 2004; Helm et al. 2006). Analysis of long-term change in species distributions in relation to dynamics of habitats might reveal which species decrease in distribution mainly because of changed habitat quantity; appropriate measures for their conservation could then be undertaken. To acquire long-term data on species abundance most countries have established some form of biodiversity monitoring. This is often accompanied by mapping of biodiversity or inventories of conservationally important species (e.g. Andersson 2002). Inventories and surveys provide important information regarding the spatial distribution of species, even though their usability for monitoring may be limited (e.g. Pavlik and Barbour 1988; Hutchings 1991). National atlases of plant species are becoming increasingly available (e.g. the Atlas of European flora, Jalas and Suominen 1972). Quite often these atlases provide information from various time periods (e.g. Preston et al. 2002; Kukk and Kull 2005) or are a repetition of an earlier similar study, enabling large-scale changes in the flora of a specific region to be estimated. The most threatened habitat types and their flora vary depending upon the region and its land-use history. In the UK, orchids of calcareous grasslands and woodlands have suffered the greatest decline in range (Kull and Hutchings 2006), while on arable land rare plants have decreased the most and the largest changes in abundance of plants are related to increased nutrient load (Smart et al. 2005). Similarly, eutrophication, along with decrease of flora of saline habitats, and of aquatic and wet habitats has driven changes in the flora of the Netherlands (Tamis et al. 2005). In Sweden, species-rich dry to mesic semi-natural grasslands have decreased from 2 million ha to 200,000 ha (Bernes 1994). Their flora contains of a large number of habitat specialists (Cousins and Eriksson 2001), and the habitats are considered strongly threatened, probably possessing a large extinction debt (Eriksson et al. 2002). In the Finnish archipelago, forest plants have increased at the expense of grassland species due to depopulation and consequent abandonment of stock farming (von Numers and Korvenpaä¨ 2007). To date, the Estonian native flora is thought to have suffered only a modest decline of diversity. However, Estonia has undergone large changes in the dominant practice of land use since World War II. These include the complete restructuring of agriculture due to enforced collectivisation (complete prohibition of private farms and establishment of collective farms) after invasion by the Soviet Union (Viiralt and Lillak 2006), massive drainage of wetlands (incl. various peatlands) (Ilomets 2005), abandonment of semi-natural

123 52 Biodivers Conserv (2008) 17:3531–3550 3533 grasslands and increased utilisation of hay-fields since the 1960s, a dramatic drop in agricultural activities in the early 1990s due to the economic downturn (Sammul et al. 2000), and intensification of forestry. It is reasonable to expect that these changes have affected the flora of Estonia. Earlier studies in Estonia have shown an increased human impact on several forest and mire habitats as well as decreased use of semi-natural habitats (Sammul et al. 2000; Kukk and Sammul 2006). Hence, we presume that most species with decreased distribution are associated with these habitat types. In this paper we estimate long-term distribution trends of vascular plant species and compare their persistence to changes in area of their habitats. Our objectives are (1) to compare the persistence of the flora of different habitat types, (2) to determine whether the species characteristics related to species habitat preference and distribution are related to long-term persistence, (3) to compare current species conservation priorities with species persistence, and (4) to provide recommendations for monitoring and conservation of plant species.

Methods

Mapping of species distribution

The compilation of the database for the Atlas of Estonian vascular plants (see Kull et al. 2002 for details) was undertaken at the beginning of the 1970s. Estonia was divided into 494 quadrats, each about 100 km2 (11.1 9 9.45 km) in size, using the Central European grid system (60 9 100). The database includes the list of species in all 494 quadrats. All quadrats were inspected several times mainly by professional botanists. The help of amateurs has been used for some groups of species (e.g. orchids), in cases where qualiﬁed people were available. In each quadrate various habitats were visited and species presence was recorded. Addi- tionally, records on species ﬁnds from herbaria, from reliable data from literature, and from different projects, have been included. Literature sources, herbaria, older vegetation analyses, older species counts in quadrats, and various sources of historical data on species distribution were used for mapping species presence in quadrats prior to 1970. The apomictic genera Alchemilla, Crataegus, Euphrasia, Hieracium, Pilosella and Taraxacum, hydrophytes and most invasive species, were excluded from current analysis since data on their distribution are scarce and not readily comparable to that of the other taxa. Data about the distribution of 1,031 species was included in the current analyses. The taxonomy follows Kukk 1999a.

Species characteristics and habitat preferences

We used literature sources (Flora of Estonian SSR 1953–1984; Kukk 1999a; Kukk and Kull 2005) to compile a list of traits indicating commonness of a species (seven ordinal classes provided by the Flora of Estonian SSR 1953–1984 and Kukk 1999a: Very Rare— 1–3 findings during the last 50 years; Rare—4–10 findings during the last 50 years; Uncommon—20–30 regionally restricted populations; Scattered—sparse distribution across Estonia; Occasional—common species within a restricted region; Common— abundant in suitable habitats, but regionally restricted; Frequent—abundant and common across Estonia); endemism in the Baltics and Fennoscandia (endemic or not); whether the species reaches the border of its global distribution range in Estonia (yes or no); sensitivity to human impact (anthropophyte—dependent on human activities, apophyte—benefiting

53 123 3534 Biodivers Conserv (2008) 17:3531–3550 from human activities, hemeradiaphor—indifferent to human influence, hemerophob— species harmed by human influence) (Kukk 1999a); life-form (according to Raunkiaer’s classification). We also added indicator values of species requirements (environmental preferences for light, temperature, continentality, soil moisture, soil acidity, nutrients, and salinity) (Ellenberg 1974; Karrer and Killian 1990; Ellenberg et al. 1991; Englisch et al. 1991; Karrer 1992) as well as species conservation category according to Estonian legislation (three ordinal categories defining level of conservation priority: 1st—strictly protected, 2nd—moderately protected, 3rd—category of weakest protection) (Kukk 1999b) and species status in the Estonian Red List which does not imply legal obligations in Estonia (five ordinal categories: Endangered—species under strong threat of becoming extinct; Vulnerable—species whose populations are quickly declining; Rare—species with restricted distribution; Care Demanding—relatively common species whose status requires attention; Indeterminate—species whose degree of being endangered cannot be specified due to insufficient data) (Lilleleht 1998). The main habitat of each species was determined using the following general categories of habitat types (Flora of Estonian SSR 1953–1984; Kukk 1999a; Leht 1999): dunes, shores (coastal habitats), bedrock outcrops and rocks, grasslands, forests, mires (bogs and fens), cultural habitats, no preference.

Typical ﬂora of habitat types

Using various sources of literature (see Appendix 1) a database of typical species lists for habitat types found in Estonia was compiled. We used a habitat classification by Paal (1997) with some additional well-defined ecotone habitat types (a total of 46 habitat types; Appendix 2). According to the published phytosociological descriptions of the flora of habitat types (Appendix 1), the presence of each species in the typical flora of a particular habitat type was recorded on two levels: (1) dominant or characteristic species of a habitat; (2) other species commonly found in that particular type of habitat. (A single species can belong to a group of dominant species in several habitat types.) Only natural and semi- natural mainland habitats were included. Species that are found only occasionally in one or the other habitat type were omitted from the flora of that habitat type. Species habitat specificity was estimated at two levels. Species limited to (i.e. occurring in the list of) less than 15% of habitat types (472 species) were considered to be habitat specialists while species occurring in more than 15% of habitat types were considered to be habitat generalists.

Data analysis

We calculated the persistence for each species as the percentage ratio of the number of quadrats occupied by the species between 1970 and 2004 in the Atlas of Estonian vascular plants (Kukk and Kull 2005) from the total number of quadrats occupied by the species in the Atlas (as in Kull et al. 2002 and Kull and Hutchings 2006). Differences in mean persistence between species belonging to different groups of endemism, commonness, occurrence at the border of the global distribution range, main habitat type, sensitivity to human impact, life-form, conservation category, and Red List category were tested with one-way type III analysis of variance. Insufﬁcient overlap between different groups prevented testing of interactions between factors. Correlation analysis was used to seek relationships between persistence and indicator values of species requirements.

123 54 Biodivers Conserv (2008) 17:3531–3550 3535

Using the species’ persistences three groups of species were selected: (a) species with no decrease or only a slight decrease in distribution range (persistence 70–100%); (b) species with intermediate decrease in range (persistence 40–70%); (c) species that had suffered a large decrease in range (persistence 0–40%). We used one-way type III analysis of variance to test whether the proportion of species belonging to each persistence group in the typical ﬂora of a particular habitat type differed between the following groups of habitat types: forest habitats, other forest-related habitats (clear-cuts, forest edges, forest survey lines, electricity and other tracks, and burnt-over areas), shrubland habitats, mires, grasslands, dunes and sandy plains, and coastal habitats (see Appendix 2 for a complete list). Bedrock outcrop habitats had to be omitted from this analysis due to lack of replicates. To correct for the mass effect in all results of analyses of variance we employed the Bonferroni-type correction with the Dunn-Sˇida´k method (Sokal and Rohlf 1995) and obtained the critical experiment-wise error rate using the following equation:

1=k pcritical ¼ 1 ðÞ1 0:05 ð1Þ Here 0.05 is the original level of probability of type I error and k stands for the number of statistical tests. The weighted average of persistence value (PV) was calculated for the ﬂora of each habitat type (h) as follows: X X PVh ¼ PVi Wih = Wih ð2Þ

Here PVi is the persistence of species i and Wih is the importance weight of species i in habitat h. An importance weight of 2 was given to dominant and characteristic species of a habitat, and an importance weight of 1 was given to other species in that habitat type. Correlation between the weighted average of persistence of a habitat and the number of species in a ﬂora of a habitat was tested, along with a test of the difference in weighted average of persistence between different groups of habitat types.

Results

Traits of decreased species

The distribution range of almost half (49% of the total) of the species in the Estonian flora has remained within 80% of the original range in the post-1970 period (Fig. 1). There are 18 species in which the persistence is lower than 10%, 120 species (11.6%) persisted in less than 40% of their original quadrats and 261 species (25.4%) persisted in 40–70% of quadrats. Different groups of main habitat type, life-form, border of the distribution range, and sensitivity to human impact all differed in their persistence (Tables 1 and 2). Endemic species had lower mean persistence than those with wider distribution, but this difference was not significant. There was a mixed trend regarding species tolerance to human influence. Hemerophob species showed the strongest decrease in range, but this group did not differ significantly from anthropophytes (Tukey HSD test, P = 0.26). Apophytes had the highest persistence and formed a single homogeneous group (P \ 0.001). Hemeradi- aphor species, which were second in persistence, differed significantly from hemerophob species (P \ 0.003), but did not differ from anthropophytes. Species at their border of the distribution range were less persistent than species within their distribution range. Even

55 123 3536 Biodivers Conserv (2008) 17:3531–3550

)

% 20

Frequency ( Frequency 10

0 0 10 20 30 40 50 60 70 80 90 100 Persistence value

Fig. 1 Frequency distribution of persistences of plant species in the Estonian ﬂora

though we did not separate between various borders, species at their south-western or western border were least persistent. Of the various life-form groups, phanerophytes were the most persistent and therophytes the least persistent. There was no difference in persistence rate between chamaephytes, hemicryptophytes and geophytes. There were only minor variations in persistence between groups of species with different main habitat preference. We found that species of mires exhibited significantly lower mean persistence than forest and grassland species (P \ 0.002 and P \ 0.016 respectively). Species inhabiting either bedrock outcrops or dunes had on average the lowest persistence, but due to the very high level of within-group variation, there was no statistically significant difference between these and other species groups. However, there was considerable difference in average persistence between species with different habitat specificity. Habitat generalists had significantly higher persistence than habitat specialists. Species belonging to various groups of commonness, conservation, and Red List categories also differed significantly in their persistence (Table 3). Persistence declined with increasing level of rarity. The highest persistence was observed for ‘frequent’ species (Fig. 2a), followed by ‘common’ species. ‘Occasional’ and ‘scattered’ species formed one homogeneous group, while ‘uncommon’, ‘rare’ and ‘very rare’ species formed another homogeneous group with the lowest persistence. The conservation status of species was also consistent with the persistence of the species (Fig. 2b): those most strictly protected are species with the lowest persistence. However, there is no significant difference between the persistence of species of two most strict conservation categories (1st and 2nd categories). Moreover, the lowest conservation category species (3rd category) exhibited the same persistence as unprotected species. Of the Red List categories, species with ‘indeterminate’ status (see Lilleleht 1998 for definitions) were least persistent, followed by groups in decreasing threat categories (Fig. 2c). We found three significant correlations between species persistence and indicator values of species’ requirements (Table 4, Fig. 3). Light-demanding species were less persistent than shade-tolerant species; species preferring alkaline soils were more persistent than species of more acidic habitats, and species preferring nutrient-rich habitats were more persistent than species of nutrient-poor conditions.

123 56 Biodivers Conserv (2008) 17:3531–3550 3537

Table 1 Averages, standard Species category Persistence N deviations and sample size of persistences of categories of Average SD species distribution and habitat preference Main habitat type Mires 66 24 122 Bedrock outcrops and rocks 62 31.6 14 Dunes 67 23.8 28 Coastal habitats 72 20.4 68 Shores 71 24.7 110 Grasslands 75 23.8 309 Forests 77 22.2 229 Cultural habitats 72 25.2 140 Sensitivity to human impact Hemerophob 61 25.4 99 Hemeradiaphor 70 23.8 490 Apophyte 81 19.8 362 Anthropophyte 67 27.8 77 Endemism Nonendemic 73 23.8 1,016 Endemic 71 27.1 15 Border of distribution range Not at the border of range 78 21.6 617 At the border of range 65 24.8 413 Life-form Therophytes 70 23.2 156 Chamaephytes 76 18.3 57 Geophytes 77 20.5 137 Hemicryptophytes 76 20.9 426 Phanerophytes 87 17.6 59 Habitat speciﬁcity Generalist 81 18 507 Specialist 65 25.7 472

Changes in ﬂoristic composition of habitat types

The weighted averages of persistence of species from most habitats were between 70 and 75%. The lowest average persistence was observed in flora of deciduous shrublands (mostly Corylus avellana habitats) while the highest average persistence was found in flora of dry boreal forests and in floodplain willow shrublands (Appendix 2). Of the groups of habitat types, the flora of bedrock outcrops had the lowest average persistence (69.6%), followed by habitats of shrublands (70.1%), while forest habitat types had the highest average persistences (76.6%). We divided the flora of Estonia into three groups according to their persistence. Species with persistence under 40% were most abundant in cliff habitat types and in sandy and dune habitats (Fig. 4). However, this difference was not significant (Table 5). There was a

57 123 3538 Biodivers Conserv (2008) 17:3531–3550

Table 2 Differences in persistence between species groups with various traits as tested with one-way ANOVA Factor Source of SS d.f. MS F-value P variation

Main habitat type Intercept 2,052,703 1 2,052,703 3,658 \0.0001 Factor 13,035 7 1,862 3.32 0.0017 Error 567,841 1,012 561 Sensitivity to human impact Intercept 2,798,578 1 2,798,578 5,306 \0.001 Factor 45,972 3 15,324 29 \0.001 Error 540,090 1,024 527 Endemism Intercept 303,721 1 303,720.9 533 \0.0001 Factor 77 1 77 0.14 0.71 Error 586,609 1,029 570.1 Border of the distribution range Intercept 5,054,831 1 5,054,831 9,644 \0.001 Factor 47,335 1 47,335 90 \0.001 Error 539,351 1,029 524 Life-form Intercept 2,938,437 1 2,938,437 6,734 \0.0001 Factor 12,170 4 3,042 6.97 0.00002 Error 362,199 830 436 Habitat speciﬁcity Intercept 5,480,158 1 5,480,158 10,948 \0.001 Factor 71,600 1 71,600 143 \0.001 Error 515,086 1,029 501

Differences between groups are statistically signiﬁcant (after Bonferroni-type correction) at the P- level \ 0.0085

Table 3 Differences in persistence between groups with different levels of commonness in the ﬂora, conservation priority and threat (red list) as tested with one-way ANOVA Factor Source of variation SS d.f. MS F-value P

Commonness Intercept 3,701,073 1 3,701,073 9,703 \0.001 Factor 192,487 6 32,081 84 \0.001 Error 388,684 1,019 381 Conservation category Intercept 1,047,640 1 1,047,640 1,920 \0.0001 Factor 26,214 3 8,738 16 \0.0001 Error 560,472 1,027 545 Red List category Intercept 856,866 1 856,866 1,768 \0.001 Factor 85,919 5 17,184 35 \0.001 Error 496,205 1,024 485

Differences between groups are significant (after Bonferroni-type correction) at the P-level \ 0.017 significantly higher number of species with persistence values between 40 and 70% in mire habitat types than in forest habitat types (Tukey HSD test, P \ 0.001) or in other forest- related habitats (P = 0.02). The proportion of species with persistence over 70% was significantly smaller in mire habitat types than in forest habitat types (P = 0.01). We found a significant non-linear negative correlation between species richness in the flora of a habitat type and the weighted average of persistence of species from a habitat type (R2 = 0.28; P \ 0.05) (Fig. 5).

123 58 Biodivers Conserv (2008) 17:3531–3550 3539

(a) D 100 C 90 80 B B A A A 70 60

Persistence 50 40 30

Rare Very rare Scattered Common Frequent Uncommon Occasional Commonness (b) 100 90 B B 80 A A 70 60

Persistence 50 40 30 1st category 2nd category 3rd category Not protected Conservation category

Persistence 40 30

Rare

Endangered Vulnerable Indeterminate Care demandingNot in Red Book Red Data Book

Fig. 2 Mean persistence for species with different levels of commonness (a), conservation category (b; 1st category is the most strictly protected one), and Red List category (c). The same letters above columns denote homogeneous groups while bars indicate 95% conﬁdence intervals of means

Discussion

Effects of changing land use

Due to large changes in land-use the Estonian land cover has experienced considerable change during the last 50 years. The total area of grasslands has decreased by about 90%; the area of mires has decreased by two-thirds while that of forest habitats has doubled (Fig. 6). Obviously, corresponding changes in the flora would be expected. Increase of forest area has been used as an indicator of positive changes in forest biodiversity, which in some cases is supported with evidence (e.g. von Numers and Korvenpaä¨ 2007). However, as shown by our data, such a simplified approach is not

59 123 3540 Biodivers Conserv (2008) 17:3531–3550

Table 4 Spearman rank order correlations between persistence and indicator values of species requirements

Ellenberg indicator value Valid N Spearman R t(N–2) P-level

Light 919 -0.147 -4.51 \0.001 Temperature 705 -0.007 -0.17 0.862 Continentality 784 -0.037 -1.03 0.301 Soil moisture 867 -0.039 -1.13 0.257 Soil acidity 737 0.115 3.12 \0.002 Nutrients demand 845 0.216 6.41 0.001 Salinity 863 0.063 1.85 0.064

100 100

75 75

50 50

25 25 Persistence Persistence 0 0 0 5100510 Light Temperature

100 100

75 75

50 50

Persistence 25 25 Persistence

0 0 0246810 036912 Continentality Soil moisture

100 100

75 75

50 50

25 25 Persistence Persistence 0 0 0246810 0510 Soil acidity Nutrients demand

100

Persistence 25

0 0510 Salinity

Fig. 3 Correlations between persistence and Ellenberg indicator values of species requirements. Only statistically signiﬁcant trendlines (Table 4) have been plotted

123 60 Biodivers Conserv (2008) 17:3531–3550 3541

Fig. 4 Proportion of species 0.75 0.9 with different persistence values in various groups of habitat types 0.6 0.8

0.45 0.7

0.3 0.6

0.15 0.5

Proportion of species 0.4 0 0.3

0.2

0.1

Table 5 Difference between different groups of habitat types in relative number of species of various ranges of persistence (PV) Source of variation SS d.f. MS F-value P

PV \ 40% Intercept 0.168 1 0.168 134 \0.0001 Habitat type 0.007 6 0.001 0.98 0.45 Error 0.051 41 0.001 PV [ 40–70% Intercept 1.45 1 1.45 510 \0.0001 Habitat type 0.078 6 0.013 4.55 0.0013 Error 0.117 41 0.003 PV [ 70–100% Intercept 20.0 1 20.0 3,272 \0.0001 Habitat type 0.10 6 0.017 2.81 0.022 Error 0.25 41 0.006

Fig. 5 Relationship between 100 weighted average of persistence y = -3.97Ln(x) + 92.2 value of the ﬂora of a habitat type in relation to number of species 90 in the ﬂora of a habitat type 80

Weighted average of persistence of average Weighted 50 0 100 200 300 400 500 600 Number of species

61 123 3542 Biodivers Conserv (2008) 17:3531–3550

Fig. 6 Changes in area of mires, 2500 1950 forests, and grasslands in Estonia from 1950 to 2005 (based on 2000 2005 Laasimer 1965; Metsakaitse 1999; Sammul et al. 2000; Kukk 1500 and Sammul 2006; Statistics Estonia 2007). 1000

Area (*1000 ha) 500

0 fens forests grasslands justified, and increase in forest area does not necessarily indicate a shift towards higher forest biodiversity. Nor does it reflect the increased intensity of use of forest resources, the resulting scarcity of high quality, old-growth forests (Andersson et al. 2003;Lo˜hmus et al. 2004), and decrease of biodiversity. Decrease of old-growth forests in Estonia is approaching a point at which conservation targets (Lo˜hmus et al. 2004) cannot be met. In time, major landscape- and habitat-changing processes, such as drainage, creation of fields, increased construction of roads, abandonment of managed land, etc, will also have an increasingly negative effect on the fauna and flora of forests. So far, the flora of forests, as well as shade-tolerant plants, has in total suffered least, but fresh boreal forests and boreal heath forests, which are among habitats that provide the best wood and thus suffer highest cutting pressure, already demonstrate strong decline in species richness (Appendix 2). The decrease in area of grasslands includes a decrease of high-diversity and conservationally important semi-natural grasslands (Kukk and Sammul 2006). Estonia is known for its high diversity grasslands (Kull and Zobel 1991; Kukk and Kull 1997;Sammuletal.2000). However, even though our results indicate moderate persistence of grassland species, we have detected a relatively higher decline in diversity of species-rich communities (mostly grasslands). Hence, the preservation of high diversity grasslands is not guaranteed and degradation, analogous to most other European countries (e.g. Willems 1983;Bakker1989; Smart et al. 2003) could continue, partly also due to loss of species and decrease in size of species pool. The decrease of mires has been habitat type-dependent (see Ilomets 2005) and has resulted in loss of almost all Estonian fens while the reduction in area of raised bogs has been much smaller (Pajula 2006). We detected lower persistence among mire species than among grassland species. Thus, local extinctions seem to be faster in mires than in grasslands, despite the greater decrease of grassland area. Partial habitat destruction in mires is usually coupled with wide-reaching impacts on the quality of the remaining habitat patches. The rapid change in ecological conditions in degraded mires leads to the rapid floristic changes revealed by our data (see also Fojt and Harding 1995). The loss of grassland habitats is primarily due to direct destruction, or abandonment and subsequent overgrowth. Therophytes respond quickly to such successional changes while clonality helps to resist the habitat change. Still, the relatively persistent floristic composition of grasslands may only be transient (von Numers and Korvenpaä¨ 2007) and may mask a high extinction debt (see also Helm et al. 2006). It should also be noted that mire flora has much higher habitat specificity than grassland flora. It has been shown before that grassland species can sometimes find refuges in new, man-made habitats such as road verges (e.g. Tikka et al. 2000; Cousins and Eriksson 2001). These habitats are not nearly equal in quality to the true grasslands (Tikka et al. 2000), but may sometimes serve as migration corridors and provide a transient habitat. The very specific habitat conditions of wetlands, however, are hardly ever replicated by humans. Quite the

123 62 Biodivers Conserv (2008) 17:3531–3550 3543 contrary: draining wet areas and turning them into utilisable land has been a major under- taking (e.g. Krug 1993; Ilomets 2005). We found that in Estonia the average persistence of species is lower in habitats that contain more species. This could be because more diverse habitats are more likely to contain species with low persistence. However, the pattern of the relationship (Fig. 5) implies that the relationship is primarily caused by low diversity habitats where ﬂora has not declined. Alternatively, it could be an effect of eutrophication, which has a relatively stronger negative effect on nutrient-poor communities. The latter, especially on calcareous soils, are among the most diverse communities inEstonia. This hypothesis is supported by similar observations from other countries, where negative effect of eutrophication has been detected (McCollin et al. 2000; Tamis et al. 2005; Piessens and Hermy 2006; Smart et al. 2006;Ro¨mermann et al. 2008), and by higher persistence of nutrient-demanding species in our data (Table 4, Fig. 3). The issue demands further detailed study, because if the stronger decline of the ﬂora of diverse communities is a general trend it has serious implications for conservation strategies. Persistence of species with different characteristics

Species with specific ecological requirements are often restricted to few habitat types and are therefore more likely to be rare (Cousins and Eriksson 2001; Dupre´ and Ehrleń 2002). The persistence of rare species is always influenced by stochasticity of the environment. Loss of one local population has a relatively larger negative impact for habitat specialists than for ubiquitous species. Our data showed decreased persistence with decreasing species commonness (Fig. 2), as well as strong negative effect of habitat specificity on persistence. Often, changed landscapes act as a dispersal barrier for habitat specialists and their negative growth rate is not balanced by immigration from other propagule sources. It is thus important to concentrate conservation efforts on sustaining habitats which provide specific environmental conditions and habitats for rare and specialized species. Our results revealed that species at the border of their global distribution range had low persistence in the Estonian flora. Such populations are mostly limited by unsuitable climatic conditions and the implications of the lower persistence probability of this group of species should be considered in planning both national and international conservation networks. Further analysis of range shifts might provide valuable information in the context of the global climate change and species protection at a pan-European scale. Higher persistence of nutrient-demanding species indicates the widely recorded trend of eutrophication of ecosystems (McCollin et al. 2000; Smart et al. 2005; Piessens and Hermy 2006;Ro¨mermann et al. 2008). Above all the enrichment concerns grasslands, many of which have been fertilized for agriculture, but probably also other habitats due to atmospheric deposition. Together with species at the border of their distribution range, hemerophobs, and of habitat specialists, all of which have suffered at least 35% decline, a set of vulnerable species is formed, that deserves further attention both from a conservation point of view and also scientifically.

Data restrictions

The data-set that was used in this study is a result of rather large generalizations. First, the mapping of species distributions was carried out over a long time interval (35 years is considered as ‘‘current’’), during which Estonian habitats have undergone a series of huge changes. Secondly, the mapping of the Estonian ﬂora has been carried out by counting species in quadrats that have an area of about 100 km2. Hence, there could be considerable changes

63 123 3544 Biodivers Conserv (2008) 17:3531–3550 occurring within these quadrats both spatially and temporally (including changes in area of various habitat types located within quadrat) that are overlooked. Thirdly, the amount of data acquired before 1970 is relatively scarce. Hence, we are only able to estimate the decline of species range; an estimated increase in species range since 1970 may be caused by the lack of information from previous periods. However, considering such restrictions only strengthens our results as all the points raised above reduce the detectability of changes.

Monitoring methodology

Due to the large effort needed for monitoring general trends in biodiversity as well as the fulfillment of national and international conservation targets, recent efforts have concentrated primarily on changes in biodiversity. Persistence could be used for evaluation of large-scale conservation efforts which are rarely documented (Pullin et al. 2004; Sutherland et al. 2004). The 2010 target (Balmford et al. 2005) requires information about dynamics of the complete biota. Plants are among the few species groups for which this task is actually realistic. Surveys like mapping of flora (combined with various inventories) are the only mechanisms that provide data on the condition of the entire flora and could alert us to a decline of a species that is not currently protected (e.g. see Kull et al. 2002) as species that are not protected are usually not monitored. Monitoring of change in common species is an important but rather neglected task in conservation planning. It is best to know when a species starts to decline and to take action before it has become rare. Moreover, such warnings are not just indicative of changes in individual species, but if several species show a consistent pattern of change, it also signals changes in the quality of habitats or environmental conditions. Large scale floristic censuses are quite rare. There are a few atlases of flora of whole countries which include data from various time periods (e.g. Flora of UK, Preston et al. 2002) and few similar databases, such as the flora of the Netherlands (Tamis et al. 2005). Even though such species lists lack the precision that comes with detailed and repeated vegetation analyses or even with simple addition of abundance data (Balmer 2002), they are still useful for covering large areas and for estimation of coarse changes. In order to provide comparative material and to evaluate pan-European shifts in plant species richness, publication of similar data from other countries is badly needed.

Implications for plant conservation

Our analysis points out that despite conservation efforts plant diversity in Estonia continues to decline and this is mostly due to human influence. The decrease of hemerophobs, of mire species, and of species preferring conditions with low nutrient availability indicates increasing negative human influence. Drainage and eutrophication are probably amongst the most dangerous effects. At the same time, a decrease of anthropophytes, light- demanding species and grassland species, indicates a decrease in positive human influence (e.g. grassland management). The majority of plants now considered rare have reached that status because of changes in land use in Estonia (Fig. 2a; see also Pa¨rtel et al. 2005). Moreover, the number of rare species has not decreased (Kukk 2003); instead, our results suggest that it could continue to increase. Hence, it could be argued that conservation efforts have been only partially successful in preserving plant diversity. Still, our results show that, on average, both legal protection categorisation and Red Data Book categorisation give high priority to recently-decreasing species (Fig. 2) and, hence, are addressing the problems to same extent (but see Pa¨rtel et al. 2005). As an evidence for at least some success of conservation actions, most species that were rare at

123 64 Biodivers Conserv (2008) 17:3531–3550 3545 previous surveys are still extant: e.g. there are several species that remain viable despite miniscule populations in Estonia (Rytta¨ri et al. 2003).

Conclusions

Our results emphasize that habitat plant diversity could change asynchronously with changes in the habitat area and depends on habitat vulnerability. We don’t want to underestimate the importance of stopping decline in habitats, quite the contrary, but point out that separating quantity and quality could enhance conservation planning for fulfillment of large-scale and long-term conservation targets. Estonia, which is often lauded for well-preserved biodiversity, should also, at least partly, reconsider its plant conservation strategies. Our data reveals troubling trends in the continuing decline of rare species and even of species with legal protection, in less persistent flora of species-rich habitats, and in effects of eutrophication. Land-use changes (e.g. Bernes 1994; Henle et al. 2004; Honnay et al. 2005; von Numers and Korvenpaä¨ 2007) and eutrophication (e.g. McCollin et al. 2000; Van der Veken et al. 2004; Tamis et al. 2005; Piessens and Hermy 2006) seem to be the main driver behind changes in plant diversity throughout the Europe. Conservation specialists have recognized some of these negative developments and measures have been planned to prevent continuation of these processes (see e.g. Sammul and Lo˜hmus 2005 for an overview). A preservation of habitats for specialist species should be given highest priority in the near future.

Acknowledgements Michael Hutchings, Dirk Schmeller and anonymous reviewers provided valuable comments to this manuscript. This study is based on research ﬁnanced by EU FP6 project EuMon, target ﬁnancing project no 0362481s03, grants no 6048 and 7567 from Estonian Science Foundation and grant no P6062PKPK06 from Estonian University of Life Sciences. Marguerite Oetjen kindly corrected the English.

Appendix 1

List of sources used to compile a database of species presence in various habitats. Diekmann M (1994) Deciduous forest vegetation in Boreonemoral Scandinavia. Acta Phytogeographica Suecica 80:1–116 Eichwald K (1966) Eesti NSV floora. 10. Valgus, Tallinn Eichwald K, Eilart J, Kalda A et al (1969) Eesti NSV floora. 4. Valgus, Tallinn Eichwald K, Kalamees K, Kask M et al (1971) Eesti NSV floora. 8. Valgus, Tallinn Eichwald K, Kask M, Kuusk V et al (1978) Eesti NSV floora. 6. Valgus, Tallinn Eichwald K, Kask M, Talts S et al (1959) Eesti NSV floora. 3. Eesti Riiklik Kirjastus, Tallinn Eichwald K, Kukk E, Kuusk V et al (1984) Eesti NSV floora. 9. Valgus, Tallinn Eichwald K, Talts S, Vaga A et al (1956) Eesti NSV floora. 2. Eesti Riiklik Kirjastus, Tallinn Eilart J, Kask M, Kuusk V et al (1973) Eesti NSV floora. 5. Valgus, Tallinn Kalda A (1960) Eesti NSV laialehiste lehtmetsade taimkate. TRU¨ toimetised 83. Botaanika-alased toö¨d IV:123–155 Krall H, Pork K, Rebassoo H (1973) Eesti niitude floora. Floristilised ma¨rkmed I(5): 315–337 Kukk T (1999) Eesti taimestik. Teaduste Akadeemia Kirjastus, Tartu-Tallinn Kukk T (2004) Eesti taimede kukeaabits. Varrak, Tallinn Kukk T, Kull K (1997) Puisniidud. Estonia Maritima 2:1–249 Kuusk V, Talts S, Viljasoo L (1979) Eesti NSV floora.11. Valgus, Tallinn Kuusk V, Tabaka L, JankeviIien JR (eds) (1996) Flora of the Baltic Countries. Compen- dium of Vascular Plants. 2. Estonian Academy of Sciences, Institute of Zoology and

65 123 3546 Biodivers Conserv (2008) 17:3531–3550

Botany, Latvian Academy of Sciences, Institute of Biology. Lithuanian Academy of Sciences, Institute of Botany. Tartu Kuusk V, Tabaka L, JankeviIien JR (eds) (2003) Flora of the Baltic Countries. Compen- dium of Vascular Plants. 3. Estonian Academy of Sciences, Institute of Zoology and Botany, Latvian Academy of Sciences, Institute of Biology. Lithuanian Academy of Sciences, Institute of Botany. Tartu Laasimer L (1965) Eesti NSV taimkate (Flora of the Estonia). Valgus, Tallinn Laasimer L, Kuusk V, Tabaka L et al (ed) (1993) Flora of the Baltic Countries. Com- pendium of Vascular Plants. 1. Estonian Academy of Sciences, Institute of Zoology and Botany, Latvian Academy of Sciences, Institute of Biology, Lithuanian Academy of Sciences, Institute of Botany, Tartu Lo˜hmus E (2004) Eesti metsakasvukohatuü¨bid. EPMU¨ Metsanduslik Uurimisinstituut. Teine tru¨kk. Eesti Loodusfoto, Tartu Ma¨gi M, Lutsar L (2001) Inventory of semi-natural grasslands in Estonia 1999–2001. Estonian Fund for Nature and Royal Dutch Society for Nature Conservation Paal J (1997) Eesti taimkatte kasvukohatuü¨pide klassifikatsioon. Classification of Estonian vegetation site types. Tartu U¨ likooli Botaanika ja O¨ koloogia Instituut, Tallinn Paal J (2000) ‘‘Loodusdirektiivi’’ elupaigatuü¨pide ka¨siraamat. Eesti Natura 2000. Tartu Paal J, Rooma I, Turb M (2004) Kas Karula kuplitel kasvab su¨rjametsi? Eesti Lood- useuurijate Seltsi aastaraamat 82:90–131 Pa¨rtel M, Kalamees R, Zobel M et al (1999) Alvar grasslands in Estonia: variation in species composition and community structure. J Veg Sci 10:561–570 Rebassoo H (1973) Huvitavamaid taimeleide Vaïnamere laidudelt. Floristilised ma¨rkmed I (5):306–309 Trass H (1960) Laä¨ne-Eesti madalsoode floora analuü¨s. TRU¨ toimetised 83. Botaanika- alased toö¨d IV: 35–95 Vaga A, Eichwald K (1953) Eesti NSV floora. 1. Eesti Riiklik Kirjastus, Tallinn U¨ ksip A (1961) Eesti NSV floora. 7. Eesti Riiklik Kirjastus, Tallinn

Appendix 2

Habitat types and basic descriptive characteristics of their ﬂora used in this study.

Habitat type Habitat group Weighted average Number of persistence of species

Alvar forests and shrublands Forests 66 155 Boreal heath forests Forests 66 257 Dry boreal forests Forests 91 22 Fresh boreal forests Forests 65 278 Dry boreo-nemoral forests Forests 87 60 Fresh boreo-nemoral forests Forests 74 332 Floodplain forests Forests 78 169 Floodplain willow shrublands Forests 90 53 Rich paludiﬁed forests Forests 77 230 Poor paludiﬁed forests Forests 75 108 Minerotrophic swamp forests Forests 75 203

123 66 Biodivers Conserv (2008) 17:3531–3550 3547

Appendix continued Habitat type Habitat group Weighted average Number of persistence of species

Mixotrophic (transitional) Forests 78 120 bog forests Ombrotrophic bog forests Forests 74 59 Drained peatland forests Forests 75 118 Wooded meadows Grasslands 76 576 Alvar grasslands Grasslands 72 301 Boreal heath grasslands Grasslands 80 45 Boreal grasslands Grasslands 67 232 Boreo-nemoral grasslands Grasslands 70 550 Floodplain grasslands Grasslands 73 439 Coastal meadows Grasslands 77 341 Paludiﬁed grasslands Grasslands 71 460 Minerotrophic fens Mires 71 305 Floodplain swamps Mires 74 100 Mixotrophic (transitional) fens Mires 72 199 Spring fens Mires 68 90 Heath moors Mires 84 15 Treeless and treed ombrotrophic Mires 66 50 raised bogs Vegetation of bedrock outcrops Outcrops 70 124 Salt marshes Coastal habitats 74 27 Rubble, pebble and gravel ridges Coastal habitats 71 162 Fucous ridges Coastal habitats 77 96 Vegetation of coastal dunes Dunes and sandy plains 73 173 Depressions between dunes Dunes and sandy plains 76 36 Vegetation of inland dunes Dunes and sandy plains 67 200 and sandy plains Alvar juniper shrubs Shrublands 71 247 Juniper shrubs with deciduous Shrublands 82 75 species Juniper shrubs on sands Shrublands 65 61 Corylus avellana bushes Shrublands 62 123 Alnus incana bushes Shrublands 66 140 Willow bushes Shrublands 71 78 Forest margins on mineral soils Other forest-related habitats 65 476 Forest margins on swampy soils Other forest-related habitats 72 203 Forest survey lines, electricity Other forest-related habitats 63 253 and other tracks Burnt-over areas Other forest-related habitats 74 31 Cut-over areas Other forest-related habitats 64 226

67 123 3548 Biodivers Conserv (2008) 17:3531–3550

References

Andersson L (2002) Mapping nature protection values—a habitat-wise presentation of regional variation in rare and vulnerable species. Svensk Bot Tidskr 96:313–322 Andersson L, Martverk R, Ku¨lvik M, Palo A, Varblane A (2003) Woodland key habitat inventory in Estonia 1999–2002. Regio Publishing, Tartu Bakker JP (1989) Nature management by grazing and cutting. Kluver Academic Publishers, Dordrecht Balmer O (2002) Species lists in ecology and conservation: Abundances matter. Conserv Biol 16:1160–1161 Balmford A, Bennun L, ten Brink B, Cooper D, Coˆte IM, Crane P et al (2005) The Convention on Biological Diversity’s 2010 target. Science 307(5707):212–213 Bernes C (ed) (1994) Biological diversity of Sweden—a country study. Swedish Environmental Protection Agency, Monitor 14 Brooks TM, Mittermeier RA, Mittermeier CG, da Fonseca GAB, Rylands AB, Konstant WR, Flick P, Pilgrim J, Oldfield S, Magin G, Hilton-Taylor C (2002) Habitat loss and extinction in the hotspots of biodiversity. Conserv Biol 16(4):909–923 Chapin FS, Zavaleta ES, Eviner VT, Naylor RL, Vitousek PM, Reynolds HL, Hooper DU, Lavorel S, Sala OE, Hobbie SE, Mack MC, Dıáz S (2000) Consequences of changing biodiversity. Nature 405: 234–242 Cousins SA, Eriksson O (2001) Plant species occurrences in a rural hemiboreal landscape: effects of remnant habitats, site history, topography and soil. Ecography 24:461–469 Dupre´ C, Ehrleń J (2002) Habitat configuration, species traits and plant distributions. J Ecol 90:796–805 Ellenberg H (1974) Zeigerwerte der Gefa¨sspflanzen Mitteleuropas. Scripta Geobot 9:1–122 Ellenberg H, Weber HE, Du¨ll R, Wirth V, Werner W, Paulisen D (1991) Zeigerwerte von Pflanzen in Mitteleuropa. Scripta Geobot 18:1–248 Englisch M, Karrer G, Wagner H (1991) Bericht u¨ber den Zustand des Waldbodens in Niedero¨sterreich. Forstl Bundesversuchsanst/Amt der Niedero¨sterr. Landesregierung, Wien Eriksson O (2000) Functional roles of remnant plant populations in communities and ecosystems. Global Ecol Biogeogr 9:443–449 Eriksson O, Cousins SA, Bruun HH (2002) Land-use history and fragmentation of traditionally managed grasslands in Scandinavia. J Veg Sci 13:743–748 Fischer M, Sto¨cklin J (1997) Local extinctions of plants in remnants of extensively used calcareous grasslands 1950–1985. Conserv Biol 11:727–737 Flora of Estonian SSR 1953–1984. Eesti NSV Floora, vol I-XI. Tallinn, Valgus Fojt W, Harding M (1995) Thirty years of change in the vegetation communities of three valley mires in Suffolk, England. J Appl Ecol 32:561–577 Hanski I (2000) Extinction dept and species credit in boreal forests: modelling the consequences of different approaches to biodiversity conservation. Ann Zool Fenn 37:271–280 Hanski I, Ovaskainen O (2002) Extinction debt at extinction threshold. Conserv Biol 16:666–673 Hedin J (2003) Metapopulation biology of Osmoderma eremita—dispersal, habitat quality and habitat history. PhD thesis. Department of Ecology, Lund University Helm A, Hanski I, Pa¨rtel M (2006) Slow response of plant species richness to habitat loss and fragmentation. Ecol Lett 9:72–77 Henle K, Lindenmayer DB, Margules CR, Saunders DA, Wissel C (2004) Species survival in fragmented landscapes: Where are we now? Biodiv Conserv 13:1–8 Honnay O, Jacquemyn H, Bossuyt B, Hermy M (2005) Forest fragmentation effects on patch occupancy and population viability of herbaceous plant species. New Phytol 166:723–736 Hutchings MJ (1991) Monitoring plant populations: census as an aid to conservation. In: Goldsmith FB (ed) Monitoring for conservation and ecology. Chapman & Hall, London Ikonen I (ed) (2004) Fogur er hlı´ðin. Fair is the blooming meadow. TemaNord 2004:564. Nordic Council of Ministers Ilomets M (2005) Eesti soode taastamine – vajadused, printsiibid, hetkeseis (Restoration of Estonian mires—needs, principles and present state). In: Sammul M, Lo˜hmus A (eds) Ecological Restoration. Year-Book of the Estonian Naturalists’ Society 83, pp 72–95 Jalas J, Suominen J (1972) Atlas Florae Europaeae. Distribution of vascular plants in Europe. 1. Pter- idophyta (Psilotaceae to Azollaceae). Committee for Mapping the Flora of Europe, Societas Biologica Fennica Vanamo, Helsinki Karrer G (1992) O¨ sterreichische Waldboden- Zustandsinventur. Teil VII: Vegetationso¨kologische Analysen. Mitt Forstl Bundesversuchsanst 168:193–242 Karrer G, Killian W (1990) Standorte und Waldgesellschaften im Leithagebirge Revier Sommerein. Mitt Forstl Bundesversuchsanst 165:1–244

123 68 Biodivers Conserv (2008) 17:3531–3550 3549

Krug A (1993) Drainage history and land use pattern of a Swedish river system – their importance for understanding nitrogen and phosphorus load. Hydrobiol 251:285–296 Kukk T (1999a) Eesti taimestik (Vascular plant flora of Estonia). Teaduste Akadeemia Kirjastus, Tallinn- Tartu (in Estonian with English summary) Kukk U¨ (1999b) Protected plants of Estonia. EPMU¨ Keskkonnakaitse Instituut, Tartu Kukk U¨ (2003) The distribution of Ligularia sibirica (L.) Cass. in Estonia and changes in its population. Biul Ogrodow Botanic 12:11–22 Kukk T, Kull K (1997) Puisniidud (Wooded Meadows). Estonia Maritima 2:1–249 Kukk T, Kull T (eds) (2005) Atlas of the Estonian flora. Institute of Agricultural and Environmental Sciences of the Estonian University of Life Sciences, Tartu Kukk T, Sammul M (2006) Loodusdirektiivi poollooduslikud kooslused ja nende pindala Eestis (Area of seminatural Natura 2000 habitat types in Estonia). In: Sammul M (ed) Year-Book of the Estonian Naturalists’ Society 84, pp114–158 Kull K, Zobel M (1991) High species richness in an Estonian wooded meadow. J Veg Sci 2:711–714 Kull T, Hutchings MJ (2006) A comparative analysis of decline in the distribution ranges of orchid species in Estonia and the United Kingdom. Biol Conserv 129:31–39 Kull T, Kukk T, Leht M, Krall H, Kukk U¨ , Kull K, Kuusk V (2002) Distribution trends of rare vascular plant species in Estonia. Biodivers Conserv 11:171–196 Laasimer L (1965) Eesti NSV taimkate (Flora of the Estonia). Valgus, Tallinn Leht M (ed) (1999) Eesti taimede maä¨raja (A guide of Estonian plants). EPMU¨ ZBI, Eesti Loodusfoto, Tartu Lilleleht V (ed) (1998) Eesti punane raamat. Ohustatud seened, taimed ja loomad (Red data book of Estonia. Threatened fungi, plants and animals). Teaduste Akadeemia looduskaitse komisjon, Tartu (in Estonian with English summary). Available online at http://www.zbi.ee/punane/ Lo˜hmus A, Kohv K, Palo A, Viilma K (2004) Loss of old-growth, and the minimum need for strictly protected forests in Estonia. Ecol Bull 51:401–411 May RM, Lawton JH, Stork NE (1995) Assessing extinction rates. In: Lawton JH, May RM (eds) Extinction rates. Oxford University Press, Oxford McCollin D, Moore L, Sparks T (2000) The flora of a cultural landscape: environmental determinants of change revealed using archival sources. Biol Conserv 92:249–263 Metsakaitse (1999) Metsakaitse- ja metsauueduskeskus. Aastaraamat Mets099. Yearbook: Forest 099. OU¨ Paar Myers N (1979) The shrinking ark: a new look at the problem of disappearing species. Pergamon Press, Oxford Paal J (1997) Eesti taimkatte kasvukohatuü¨pide klassifikatsioon (Classification of Estonian vegetation site types). Tartu U¨ likooli Botaanika ja O¨ koloogia Instituut, Tallinn (in Estonian) Pajula R (2006) Kui palju on Eestis soid (How much mires are there in Estonia)? Eesti Loodus 1:14–19 Pa¨rtel M, Kalamees R, Reier U¨ , Tuvi E-L, Roosaluste E, Vellak A, Zobel M (2005) Grouping and prioritization of vascular plant species for conservation: combining natural rarity and management need. Biol Conserv 123:271–278 Pavlik BM, Barbour MG (1988) Demographic monitoring of endemic sand dune plants, Eureka Valley, California. Biol Conserv 46:217–242 Piessens K, Hermy M (2006) Does the heathland flora in north-western Belgium show an extinction debt? Biol Conserv 132:382–394 Preston CD, Pearman DA, Dines TD (2002) New atlas of the British and Irish flora. Oxford University Press, Oxford Pullin AS, Knight TM, Stone DA, Charman K (2004) Do conservation managers use scientific evidence to support their decision-making? Biol Conserv 119:245–252 Riis T, Sand-Jensen K (2001) Historical changes in species composition and richness accompanying per- turbation and eutrophication of Danish lowland streams over 100 years. Freshwater Biol 46:269–280 Ro¨mermann C, Tackenberg O, Jackel A-K, Poschlod P (2008) Eutrophication and fragmentation are related to species’ rate of decline but not to species rarity: results from a functional approach. Biodivers Conserv 17:591–604 Rytta¨ri T, Kukk U¨ , Kull T, Ja¨ka¨la¨niemi A, Reitalu M (eds) (2003) Monitoring of threatened vascular plants in Estonia and Finland—methods and experiences. The Finnish Environment 659, Helsinki Sala OE, Chapin FSI, Armesto JJ et al (2000) Global biodiversity scenarios for the year 2000. Science 287:1770–1774 Sammul M, Lo˜hmus A (eds) (2005) Ecological restoration. Year-Book of the Estonian Naturalists’ Society 83, Estonian Naturalists’ Society, Tartu Sammul M, Kull K, Kukk T (2000) Natural grasslands in Estonia: evolution, environmental and economic roles. In: Viiralt R, Lillak R, Michelson M (eds) Conventional and ecological grassland management. Estonian Grassland Society, Tartu

69 123 3550 Biodivers Conserv (2008) 17:3531–3550

Smart SM, Clarke RT, van de Poll HM, Robertson EJ, Shield ER, Bunce RGH, Maskell LC (2003) National- scale vegetation change across Britain; an analysis of sample-based surveillance data from the Countryside Surveys of 1990 and 1998. J Environ Manag 67:239–254 Smart SM, Bunce RGH, Marrs R, LeDuc M, Firbank LG, Maskell LC, Scott WA, Thompson K, Walker KJ (2005) Large-scale changes in the abundance of common higher plant species across Britain between 1978, 1990 and 1998 as a consequence of human activity: tests of hypothesised changes in trait representation. Biol Conserv 124:355–371 Smart SM, Thompson K, Marrs RH, Le Duc MG, Maskell LC, Firbank LG (2006) Biotic homogenization and changes in species diversity across human-modified ecosystems. 273:2659–2665 Sokal RR, Rohlf FJ (1995) Biometry, the Principles and practice of statistics in biological research, 3rd edn. WH Freeman and Co, New York Soule´ ME (1987) Viable populations for conservation. Cambridge Unviersity Press, New York Statistics Estonia (2007) Metsavaru metsade inventeerimise statistilise valikmeetodi (SMI) alusel (Forest resources according to statistical inventory). Available: http://pub.stat.ee/px-web.2001/dialog/varval. asp?ma=KK51&ti=METSAVARU+METSADE+INVENTEERIMISE+STATISTILISE+VALIKMEET ODI+%28SMI%29+ALUSEL&path=../Database/Keskkond/06Loodusvarad_ja_nende_kasutamine/08 Metsavaru/&search=METS&lang=2. Cited 17.01.2007 Sutherland WJ, Pullin AS, Dolman PM, Knight TM (2004) The need for evidence-based conservation. TREE 19:305–308 Tamis WLM, Van’t Zelfde M, van der Meijden R, Groen CLG, de Haes HAU (2005) Ecological interpretation of changes in the dutch flora in the 20th century. Biol Conserv 125:211–224 Tikka PM, Koski PS, Kivela¨ RA, Kuitunen MT (2000) Can grassland plant communities be preserved on road and railway verges? Appl Veg Sci 3:25–32 Tilman D, May RM, Lehman CL, Nowak MA (1994) Habitat destruction and the extinction dept. Nature 371:65–66 Van der Veken S, Verheyen K, Hermy M (2004) Plant species loss in an urban area (Turnout, Belgium) from 1980 to 1999 and its envronmental determinants. Flora 199:516–523 Viiralt R, Lillak R (2006) Rohumaade po˜llumajandusliku kasutuse ajalooline areng. Rohumaaviljeluse ja rohumaateaduse areng ning edendajad Eestis. In: Bender A (ed) Erituü¨biliste rohumaade rajamine ja kasutamine I. Eesti Vabariigi Po˜llumajandusministeerium. Jo˜geva Sordiaretuse Instituut, Tartu U¨ likooli Kirjastus von Numers M, Korvenpaä¨ T (2007) 20th century vegetation changes in an island archipelago, SW Finland. Ecography 30:789–800 Willems JH (1983) Species composition and above ground phytomass in chalk grassland with different management. Vegetatio 52:171–180 Wilson EO (1992) The diversity of life. Allen Line, London

123 70 II Lanno, K., Sammul, M., Luuk, O., Mesipuu, M. 2014. Persistence of declining species on alluvial meadows of Estonia Submitted. Persistence of declining species on alluvial meadows of Estonia

Kaire Lanno, Marek Sammul, Ott Luuk & Meeli Mesipuu

Lanno, K. (Corresponding author, [email protected]): Institute of Agricultural and Environmental Sciences, Estonian University of Life Sciences, Kreutzwaldi 5, Tartu, EE-51014, Estonia.

Sammul, M. ([email protected]): Institute of Agricultural and Environmental Sciences, Estonian University of Life Sciences, Kreutzwaldi 5, Tartu, EE-51014, Estonia. Current address: European College, University of Tartu. Lossi 36, Tartu, Estonia

Luuk, O. ([email protected]): Institute of Ecology and Earth Sciences, University of Tartu, Lai 40, Tartu, EE-51005, Estonia.

Mesipuu, M. ([email protected]): Institute of Agricultural and Environmental Sciences, Estonian University of Life Sciences, Kreutzwaldi 5, Tartu, EE-51014, Estonia.

Abstract Question: Which traits characterize the declining species of alluvial meadows, and which characteristics indicate the meadows where these species still persist? Location: Alluvial meadows of Estonia. Methods: The information on the changes in species distribution was combined with information from the mapping of alluvial meadows. Traits of species which distribution has declined the most were compared to other species (not declining) found on alluvial meadows. Characteristics of meadows which inhabit declining species were detected using GLM and DCA. Results: The declining species had higher demand for moisture and light availability than species not declining. The proportion of declining species on a meadow was correlated with the area, moisture and light availability, soil reaction and productivity of the meadows. Surprisingly, the presence of declining species was not related to the management (time since last grazing or mowing) of the meadow. The vegetation of the meadows containing declining species did not differentiate according to the DCA analysis. Conclusions: The results highlight the importance of preserving the habitat quality, especially the moisture regime, for the conservation of declining species of alluvial meadows. However, the occurrence of declining species is not simply predictable from the site characteristics and for protection of declining species species-specific conservation measures may be needed.

Keywords: alluvial meadows; species persistence; semi-natural communities; habitat quality; species traits; nature conservation; species decline; wetlands

Running head: Declining species of alluvial meadows.

73 Introduction It is widely acknowledged, that the loss and fragmentation of habitats are the main factors which lead to decline in distribution range and/or abundance of species adapted to specific habitat conditions (Hobbs & Yates 2003; Fischer & Lindenmayer 2007; Brook et al. 2008). The changes in species distribution can be assessed using various datasets like surveillance records or national floras, which contain information for some region from various time periods (like e.g. in Tamis et al. 2005; van Calster et al. 2008; Sammul et al. 2008). Based on this and with the information about the change in habitat areas, the conclusions can be made about the vegetation types and species most endangered (e.g. Sammul et al. 2008). Among habitats which area has decreased the most in Europe are semi- natural communities (Eriksson et al. 2002; Poschlod et al. 2005), among them alluvial meadows (e.g. Prach 2007). In Estonia, the decrease is caused mainly by conversion of those meadows into cultivated grasslands and arable fields in the 1960s and abandonment since 1980s (Truus & Tõnisson 1998). The area of the alluvial meadows has decreased about 10-fold from 150 000 ha in the beginning of 20th century to about the 15 000 ha at present (Kukk & Sammul 2006). Without extensive management, consisting of mowing and hay removal or grazing, the overgrowth with shrubs and trees and occasionally with reed is quick. Abandoned nutrient rich habitats like alluvial meadows also accumulate large amount of litter which leads to suppression of seedling emergence (Clark & Tilman 2010; Neuenkamp et al. 2013). These processes cause decrease of plant populations, habitat specialists usually being the most affected (e.g. Fischer & Stöcklin 1997; van der Veken et al. 2004). The characteristics of species with declined distribution may provide valuable information about what are the causes of their decline (e.g. Musters 2013), where one could still presume to find them, and which conservation measures should be undertaken. Based on the information about species distribution in the Atlas of Estonian vascular plants (Kukk & Kull 2005), the persistence of distribution of different plant species, including the species belonging to the flora of the alluvial meadows has been evaluated (Sammul et al. 2008). The Database of Estonian Seminatural Communities (compiled and administered by Estonian Seminatural Communities Conservation Association) provides information about the present- time distribution of alluvial meadows in Estonia and their ecological conditions as well as species lists. Based on these sources we will evaluate the occurrence of species with declined distribution in Estonia on alluvial meadows, addressing the following questions: (1) what traits characterise declining species; (2) what traits characterise meadows where the declining species still persist?

Methods The declining species of alluvial meadows We used the data about habitat preference of plant species and persistence of their distribution provided by Sammul et al. (2008). Based on information from local flora (Flora of Estonian SSR 1953-1984) and other literature sources (for a complete list see Sammul et al. 2008) species were attributed to the typical habitats where they occur. From that database we retrieved information about species typically preferring alluvial grasslands as (one of) their primary habitat(s) (altogether 506 species). The persistence of species was assessed based on their distribution according to the Atlas of Estonian vascular plants (Kukk & Kull 2005). The Atlas has divided Estonia into approximately 10*10 km quadrats and has recorded the presence of each species in each quadrat both before 1970 and after that, with most of the

74 information gathered in years 1998-2000. The persistence for each species was calculated as the percentage ratio of the number of 10*10 km quadrats occupied by the species between 1970 and currently in the Atlas (see also Sammul et al. 2008). The species with persistence value of 70% or less (meaning a decline of 30% or more in the distribution of a species in Estonia from 1970 until the current period) were treated as declining species. There were altogether 81 declining species of alluvial meadows. Of these, 45 species were found on alluvial meadows included in current analysis (see below). The average height of the species occurring on the meadows was retrieved from Leht (1999), and plant strategy according to Grime from BiolFlor database (in order to obtain more balanced occurrence of species in different groups and to avoid groups with no species, species were recategorized into four categories: CSR, S, R and C (comprising the species with C, CS and CR strategies)).

Ecological characteristics of alluvial meadows The data about alluvial meadows was retrieved from the Database of Estonian Seminatural Communities. We extracted the data of the habitats with NATURA2000 type 6450 (Northern boreal alluvial meadows). In order to reduce the variability caused by temporal variation in detection probability of species as well as bias in species detection caused by unequal knowledge of species by the people who visited the sites, we selected only the sites which were visited in June, July, August or September (summer to early autumn), by selected list of professional botanists. For every site, the database includes the information about various characteristics (Figure 6 in Luhamaa et al. 2001 provides the full information about the information gathered during field visits). In this study the following traits were used: area, cover of forest patches, cover of bushes, area of open meadow, drainage (on the scale from 0 being absent to 3 being strong), time since last mowing (0 – never mown; 1 – mowing ended >10 years ago; 2 – mowing ended 4-10 years ago; 3 – mowed 1-3 years ago; 4 – mown in current year), also an expert estimation of a geobotanical condition of a community (typical and well-developed, relatively stable meadow being the best) and floristic value (the presence of rare and threatened species as well as typical flora of a habitat) (the latter two estimates are both at the scale from 0 being absent (totally changed community, i.e. not a meadow anymore) to 3 being high); and the species list of the site (the entries with less than 5 species recorded were excluded). Altogether, the compiled list contained the information about 859 alluvial meadows. Based on the species lists we determined the occurrence or absence of declining species in a particular meadow. For every site, the proportion of declining species and total number of species were calculated. Additionally, for every site the average Ellenberg indicator value (EIV) of all species recorded on a site was calculated for light, soil moisture, soil reaction and nitrogen (values derived from Ellenberg et al. 1991).

Statistical analyses Before the analyses, the area of sites was LOG10-transformed, cover of forest patches and cover of shrubs were LOG10(x+1)-transformed and the area of open meadow was squared. If the database contained several estimations for following characteristics for one site (due to variability of the conditions of the site): drainage, mowing regime, grazing, only the prevailing assessment was retained. The

75 assessments of the estimated geobotanical value and floristic value were also recoded to consist of a single estimate (prevailing assessment) each. Detrended correspondence analysis (DCA) was used in order to describe the variation in vegetation composition. Environmental vectors were passively fitted on the resulting ordination. The difference between declining species and other species (not declining) found on a meadow in Ellenberg indicator values and average height of species was tested using one-way analysis of variance, and in strategy type using Fisher's exact test for count data. The general linear model was built to explain variation in the proportion of declining species on a site assuming binomial distribution of the dependent variable and using total number of species per site as a weight, as suggested for analysis of proportions by Wilson & Hardy (2002). The continuous factors used in the model were area of the meadow, cover of forest patches, cover of bushes, area of open meadow, and Ellenberg indicator values, and categorical factor used was expert estimate about drainage. We included the power terms of all continuous factors for the test on non-linearity of the relationship and all second- order interactions between factors. Analysis of variance was performed using Statistica 12 (StatSoft Inc.) and other statistical analyses were performed using R vers. 3.0.1 (R Development Core Team, Vienna, Austria). DCA was performed with package VEGAN (version 2.0-10), and environmental vectors were fitted on a DCA ordination plane using the function ‘envfit’ with 999 permutations.

Results The declining species Altogether, 45 declining species out of the 81 species appeared on the alluvial meadows (Appendix 1). Those declining species had higher values of EIVs (Fig. 1) for light (P < 0.02) and moisture demand (P < 0.001). There was no significant difference in EIVs of soil reaction and nitrogen demand between these two groups of species. There was no difference in average height of the declining and other species (P = 0.56). The two groups of species did not differ in their strategy type (P = 0.39).

Fig. 1. The difference between the declining species and other (non-declining) species recorded on alluvial meadows based on their Ellenberg indicator values for light, moisture, reaction and nitrogen demand.

76 The characteristics of meadows containing declining species According to the general linear model, the factors presented in Table 1 proved to be significant influencing the occurrence of declining species. The proportion of declining species occurring on a site was bigger in larger alluvial meadows, in sites with higher light and moisture availability as well as on more alkaline soils. The relationship with nitrogen availability was nonlinear and lowest proportion of declining species was found at medium values of nitrogen EIVs of meadows.

Table 1. The final general linear model of the effect of characteristics of the meadow on the proportion of declining species. The light, soil moisture, soil reaction and nitrogen are Ellenberg indicator values calculated for the sites. Binomial distribution of the proportion of declining species was defined in the model and total species number found on a meadow was used as a weight. Factor Estimate Std. error z value Pr(>|z|) (Intercept) -13.546 2.246 -6.032 <0.001 Area 0.266 0.072 3.724 <0.001 Light 1.150 0.242 4.740 <0.001 Soil moisture 0.442 0.057 7.690 <0.001 Soil reaction 0.361 0.110 3.301 <0.001 Nitrogen -1.576 0.551 -2.862 0.004 Nitrogen2 0.152 0.054 2.831 <0.005

The results of the DCA ordination showed, that there is no differentiation of vegetation between sites containing declining species and those without declining species (Fig. 2). The first DCA axis was positively correlated with total number of species found at the meadow, but also with presence of mowing in recent history and negatively with EIV for soil moisture (the DCA scores for the axes are shown in Table 2). The second DCA axis was positively correlated with EIVs for nitrogen and soil reaction and negatively correlated with the EIV for light and with the area of the meadow.

Fig. 2. The result of the DCA ordination of the vegetation showing the distribution of meadows with (triangles) and without (circles) declining species.

Table 2. The DCA scores for the different characteristics of the meadows. The light, soil moisture, soil reaction and nitrogen are Ellenberg indicator values calculated for the sites. Factor DCA1 DCA2 r2 Pr(>r) Area -0.423 -0.906 0.192 0.001 Cover of forest patches 0.997 -0.082 0.023 0.040 Cover of bushes 0.175 -0.985 0.034 0.010 Area of open meadow 0.976 0.217 0.010 0.245 Drainage 0.398 -0.917 0.066 0.001 Time since last mowing 0.801 -0.599 0.148 0.001 Geobotanical value 0.192 -0.981 0.096 0.001 Floristic value 0.074 -0.997 0.059 0.001 Light -0.541 -0.841 0.261 0.001 Soil moisture -0.919 -0.393 0.841 0.001 Soil reaction -0.230 0.973 0.297 0.001 Nitrogen -0.252 0.968 0.737 0.001 Proportion of declining species -0.984 -0.178 0.174 0.001 Total number of species 0.934 -0.357 0.284 0.001

Discussion The area of alluvial meadows in Europe has declined and this has brought about a decline in the distribution of species associated with these habitats (Joyce & Wade 1998; Kukk & Sammul 2006; Prach 2007; Eriksson 2008; Sammul et al. 2008). The common causes of the degradation of alluvial meadows are conversion to an arable land, eutrophication, expansion of invasive species, and abandonment of traditional agricultural practices (see Joyce & Wade 1998 and Prach 2007 for a review). Often larger and competitively superior species expand at the expense of low-statue species- rich vegetation (Lepš 1999; Prach 2007). Thus, to attempt a knowledge-based protection of plant diversity of alluvial meadows, taking into account ecological demands of species which are suffering because of undesired human impact, the following analyses could help guide the decision making. Firstly, one needs to detect what species are being lost; secondly, the habitat requirements of these species need to be evaluated; thirdly, if possible, environmental conditions of areas where such declining species still remain could be evaluated; and fourthly, based on habitat requirements of declining species, priority areas for conservation could be estimated. In this study we attempted to follow the first three steps of this procedure and evaluated the traits and distribution of declining species of alluvial meadows in order to shed light into the causes of decline and provide information for the conservation about where the emphasis of restoration efforts could be directed. Declining species were characterized by higher light and moisture demand compared to other species in alluvial meadows. This confirms the negative effect of overgrowing and drainage on species of such wet habitats (e.g. Truus & Tõnisson 1998). The results are also in accordance with several studies that have found the decline in light demanding species (e.g. Hannus & von Numers 2010), resulting from the loss of mowing or grazing and subsequent competitive exclusion. Often also the short species are detected to be more prone to extinction (e.g. Saar et al. 2012). However, our study did not find the difference in height of the declining species

78 compared to other species. This could be because the plant species of alluvial meadows are quite high compared to some other grassland types, where after abandonment taller species gain in abundance. Moreover, it could be that the dependence on light availability is mediated by germination (shade and litter accumulation hindering the germination) (e.g. Huhta et al. 2001) in which case the typical size of a plant species is irrelevant. It has been shown in several analyses that the present-day vegetation of alluvial grasslands is dispersal-limited (e.g., Bakker & Berendse 1999; Prach 2007), and disturbances in vegetation that create gaps where seeds can germinate promote species richness (e.g., Donath et al. 2003). Our study could also not find the difference in strategy type in declining species compared to other species, although it could be assumed, that if the decline of species is associated with ceased habitat management, the group of declining species would contain more stress-tolerator and less competitor species (e.g. Hellstrom et al. 2003). The sites which have higher probability of containing declining species tended to be larger, have higher moisture and light availability, neutral or calcareous soils and either low or high, but not medium productivity. The presence of declining species in larger meadows could be expected. Large meadows also tend to be more heterogeneous, providing different microhabitats for species. However, large habitats could not remain without suitable management or favourable environmental conditions, which suggests that the presence of management should have had a positive impact on the presence of declining species (e.g. Kull & Zobel 1991; Metsoja et al. 2012). Surprisingly, we found no such relationship. This could mean that the management of meadows alone is not sufficient to preserve the declining plant species. The declining species may sometimes have rather specific habitat requirements or specific causes of decline, such as, e.g., isolation leading to a loss of genetic diversity and reproductive success (e.g. Ilves et al. 2013). Thus their preservation may demand special conservation measures in addition to ensured grazing or mowing (e.g. Lanno and Sammul 2014). The results of the DCA showed that the species composition of alluvial meadows varied on one hand on the axis of moisture – mowing. The total number of species was correlated with mowing and it is well known that the moistest alluvial meadows are very difficult to mow, quite often even impossible. The second DCA axis was more complicated. It was related to productivity and soil reaction on one end and light availability and area on other. The nature conservation estimations of the meadows were related to the latter, but presence of declining species was more strongly correlated with moisture conditions. However, the sites containing declining species were distributed randomly across the ordination plane and did not differentiate from sites without declining species on the basis of vegetation. This indicates that the occurrence of declining species is not predictable based on the species composition of the studied meadows. Unfortunately there is not enough information about on which specific sites declining plant species were present before the 1970. The Atlas of Estonian vascular plants does not allow to estimate this because of the large size of the quadrats. It would be very informative to compare the meadows that have lost the declining species with meadows where the latter are still present. This would allow more precise inference about the reasons of species loss based on the changes in management and vegetation. Based on the results of this study it can be concluded, that the main precondition for the presence of declining species are the moisture and light availability. In addition, declining species are more likely to persist on large meadows,

79 containing some alcaline features. However, the occurrence of declining species on a site is not related to the species composition. On one hand it means, that the presence of a declined species on a site is a chance event and is not strongly related to some specific type of a meadow. On the other hand, it could indicate, that when the purpose is to conserve the specific declining (hence, probably rare, or becoming rare) species, species-specific degradation causes need to be detected. Moreover, for a preservation of the populations of declining species also species-specific restoration methods need to be applied (e.g., Lanno & Sammul 2014), as only the general methods applied for habitat preservation (like mowing and grazing) may not be sufficient.

Acknowledgements We thank the Seminatural Community Conservation Association for the data of Alluvial meadows. This work was supported by institutional research funding IUT (21-1) of the Estonian Ministry of Education and Research.

References Aug, H & Kokk, R. 1983. Eesti NSV looduslike rohumaade levik ja saagikus. Eesti NSV Agrotööstuskoondise Informatsiooni ja Juurutamise Valitsus, Tallinn, EE Bakker, J.P., Berendse, F. 1999. Constraints in the restoration of ecological diversity in grassland and heathland communities. Trends in Ecology and Evolution 14: 63-68. Brook, B.W., Sodhi, N.S. & Bradshaw, C.J.A. 2008. Synergies among extinction drivers under global change. Trends in Ecology and Evolution 23: 453-460. Clark, C.M. & Tilman, D. 2010. Recovery of plant diversity following N cessation: effects of recruitment, litter and elevated N cycling. Ecology 91: 3620-3630. Donath, T.W., Hölzel, N., Otte, A. 2003. The impact of site conditions and seed dispersal on restoration success in alluvial meadows. Applied Vegetation Science 6: 13-22. Ellenberg, H., Weber, H.E., Düll, R., Wirth, V., Werner, W. & Paulissen, D. 1991. Zeigerwerte von Pflanzen in Mitteleuropa. Scripta Geobotanica 18: 1-248. Eriksson, O., Cousins, S.A.O. & Bruun, H.H. 2002. Land-use history and fragmentation of traditionally managed grasslands in Scandinavia. Journal of Vegetation Science 13: 743–748. Eriksson, M.O.G. 2008. Management of Natura 2000 habitats. 6450 Northern Boreal alluvial meadows. European Commission, Brussels, BE. Fischer, J. & Lindenmayer, D.B. 2007. Landscape modification and habitat fragmentation: a synthesis. Global Ecology and Biogeography 16: 265-280. Fischer, M. & Stöcklin, J. 1997. Local extinctions of plants in remnants of extensively used calcareous grasslands 1950-1985. Conservation Biology 11: 727-737. Flora of Estonian SSR 1953-1984. Eesti NSV floora, vol I-IX. Valgus, Tallinn, EE. Hannus, J.-J. & von Numers, M. 2010. Temporal changes in the island flora at different scales in the archipelago of SW Finland. Applied Vegetation Science 13: 531-545. Hansson, M. & Fogelfors, H. 2000. Management of a seminatural grassland; results from a 15-year-old experiment in southern Sweden. Journal of Vegetation Science 11: 31–38. Hellstrom, K., Huhta, A.P., Rautio, P., Tuomi, J., Oksanen, J. & Kari, L. 2003. Use of sheep grazing in the restoration of semi-natural meadows in northern Finland. Applied Vegetation Science 6: 45-52.

80 Hobbs, R.J. & Yates, C.J. 2003. Impacts of ecosystem fragmentation on plant populations: generalising the idiosyncratic. Australian Journal of Botany 51: 471-488. Huhta, A.-P., Rautio, P., Tuomi, J. & Laine, K. 2001. Restorative mowing on an abandoned semi-natural meadow: short-term and predicted long-term effects. Journal of Vegetation Science 12: 677-686. Ilves, A., Lanno, K., Sammul, M. & Tali, K. 2013. Genetic variability, population size and reproduction potential in Ligularia sibirica (L.) populations in Estonia. Conservation Genetics 14: 661 – 669. Joyce, C.B. & Wade, P.M. (eds.) 1998. European Wet Grasslands: Biodiversity, Management and Restoration. Wiley, Chichester, UK. Kukk, T. & Kull, T. 2005. Atlas of the Estonian flora. Institute of Agricultural and Environmental Sciences of the Estonian University of Life sciences, Tartu, EE. Kukk, T. & Sammul, M. 2006. Loodusdirektiivi poollooduslikud kooslused ja nende pindala Eestis. Eesti Looduseuurijate Seltsi Aastaraamat 84: 114–158. Teaduste Akadeemia Kirjastus, Tallinn, EE. Kull, K. & Zobel, M. 1991. High species richness in an Estonian wooded meadow. Journal of Vegetation Science 2: 711-714. Lanno, K. & Sammul, M. 2014. The survival of transplants of rare Ligularia sibirica is enhanced by neighbouring plants. Folia Geobotanica 49: 163-173. Leht, M. (ed.) 1999. Eesti taimede määraja [A guide of Estonian plants]. EPMÜ ZBI, Eesti Loodusfoto, Tartu, EE. Lepš, J. 1999. Nutrient status, disturbance and competition: an experimental test of relationships in a wet meadow. Journal of Vegetation Science 10: 219-230. Luhamaa, H., Ikonen, I. & Kukk, T. 2001. Läänemaa pärandkooslused. Seminatural Communities of Läänemaa County, Estonia. Pärandkoosluste Kaitse Ühing, Tartu-Turku, EE. Metsoja, J.-A., Neuenkamp, L., Pihu, S., Vellak, K., Kalwij, J.M. & Zobel, M. 2012. Restoration of flooded meadows in Estonia – vegetation changes and management indicators. Applied Vegetation Science 15: 231-244. Musters, C.J.M., Kalkman, V. & van Strien, A. 2013. Predicting rarity and decline in animals, plants, and mushrooms based on species attributes and indicator groups. Ecology and Evolution 3: 3401-3414. Neuenkamp, L., Metsoja, J.-A., Zobel, M. & Hölzel, N. 2013. Impact of management on biodiversity-biomass relations in Estonian flooded meadows. Plant Ecology 214: 845-856. Poschlod, P., Bakker, J.P. & Kahmen, S. 2005. Changing land use and its impact on biodiversity. Basic and Applied Ecology 6: 93–98. Prach, K. 2007. Alluvial meadows under changing management: their degradation and restoration. In: Okruszko, T., Maltby, E., Szatyłowicz, J., Świątek, D. & Kotowski, W. (eds.) Wetlands: monitoring, modelling and management, pp. 265- 271. Taylor & Frances Ltd., London, UK. Saar, L., Takkis, K., Pärtel, M. & Helm, A. 2012. Which plant traits predict species loss in calcareous grasslands with extinction debt? Diversity and Distributions 18: 808-817. Sammul, M., Kull, T., Lanno, K., Otsus, M., Mägi, M. & Kana, S. 2008. Habitat preferences and distribution characteristics are indicative of species long- term persistence in the Estonian flora. Biodiversity Conservation 17: 3531-3550.

81 Truus, L. & Tõnisson, A. 1998. The ecology of floodplain grasslands in Estonia. In: Joyce, C.B. & Wade, P.M. (eds.) European wet grasslands: biodiversity, management and restoration, pp. 49–60. Wiley, New York, NY, US. Van der Veken, S., Verheyen, K. & Hermy, M. 2004. Plant species loss in an urban area (Turnhout, Belgium) from 1880 to 1999 and its environmental determinants. Flora 199: 516-523. Wilson, K. & Hardy, I.C.W. 2002. Statistical analysis of sex ratios: an introduction. In: Hardy, I.C.W. (ed.) Sex ratios: Concepts and research methods, pp. 48- 92. Cambridge University Press, Cambridge, UK.

82 Appendix 1. List of declining species found in alluvial meadows according to the Database of Estonian Seminatural Communities.

Persistence Species Family value Acorus calamus Araceae 68 Bidens cernua Asteraceae 66 Carex buxbaumii Cyperaceae 65 Carex davalliana Cyperaceae 62 Carex demissa Cyperaceae 18 Carex diandra Cyperaceae 70 Carex hartmanii Cyperaceae 59 Carex hostiana Cyperaceae 64 Carex lepidocarpa Cyperaceae 56 Carex riparia Cyperaceae 63 Carex viridula Cyperaceae 69 Dactylorhiza maculata Orchidaceae 58 Epilobium roseum Onagraceae 66 Eriophorum latifolium Cyperaceae 64 Euphorbia palustris Euphorbiaceae 65 Galium rivale Rubiaceae 65 Gladiolus imbricatus Iridaceae 47 Holcus lanatus Poaceae 64 Iris sibirica Iridaceae 67 Juncus alpinus nodulosus Juncaceae 70 Luzula campestris Juncaceae 66 Nardus stricta Poaceae 54 Ophioglossum vulgatum Ophioglossaceae 67 Pedicularis sceptrum-carolinum Scrophulariaceae 34 Polemonium caeruleum Polemoniaceae 41 Polygala vulgaris Polygalaceae 58 Polygonum bistorta Polygonaceae 69 Polygonum hydropiper Polygonaceae 65 Polygonum persicaria Polygonaceae 64 Polygonum viviparum Polygonaceae 44 Ranunculus reptans Ranunculaceae 60 Rorippa amphibia Brassicaceae 63 Rumex pseudonatronatus Polygonaceae 54 Sanguisorba officinalis Rosaceae 67

83 Schoenoplectus lacustris Cyperaceae 67 Schoenus ferrugineus Cyperaceae 64 Scolochloa festucacea Poaceae 62 Scutellaria hastifolia Lamiaceae 67 Senecio paludosus Asteraceae 64 Stellaria crassifolia Caryophyllaceae 39 Swertia perennis Gentianaceae 47 Trifolium spadiceum Fabaceae 47 Veronica scutellata Scrophulariaceae 69 Viola persicifolia Violaceae 51 Viola uliginosa Violaceae 58

84 III Ilves, A., Lanno, K., Sammul, M., Tali, K. 2013. Genetic variability, population size and reproduction potential in Ligularia sibirica (L.) populations in Estonia. Conservation Genetics 14: 661 – 669. Conserv Genet (2013) 14:661–669 DOI 10.1007/s10592-013-0459-x

RESEARCH ARTICLE

Genetic variability, population size and reproduction potential in Ligularia sibirica (L.) populations in Estonia

Aigi Ilves • Kaire Lanno • Marek Sammul • Kadri Tali

Received: 28 March 2012 / Accepted: 31 January 2013 / Published online: 12 February 2013 Ó Springer Science+Business Media Dordrecht 2013

Abstract Ligularia sibirica (L.) Cass. (Asteraceae) is a Introduction EU Habitats Directive Annex II plant species that has suffered a lot from human-caused major changes in quality Populations of many endangered plant species have and availability of habitats in Estonia. The aim of this study decreased in size and become isolated because of the was to find out if the observed decline in population size is changes in land use during the last century (e.g. Brook et al. reflected in the amount of genetic variation and fertility in 2008; Hobbs and Yates 2003). However, the impairment of remnant populations of this species. AFLP technique was habitat conditions is often accompanied by changes in used for that purpose. Genetic diversity within populations demographic processes (Schleuning and Matthies 2009) and was assessed as the percentage of polymorphic loci in a increased genetic stochasticity, i.e. genetic drift (Ouborg given population and average gene diversity over loci. The et al. 2006). As the effective population size diminishes, degree of genetic differentiation among populations and random loss and fixation of alleles through drift result in the genetic differentiation between pairs of populations was loss of genetic variation and differentiation of populations estimated. The amount of viable seeds per flower stem was (Ellstrand and Elam 1993). This, in turn, could lead to compared among populations and between years (2007 and reduced fitness and further reduce the population size and 2008). Average genetic diversity over loci and proportion thus predestine a population into a vortex of extinction of polymorphic loci in L. sibirica populations were sig- (Ellstrand and Elam 1993; Gilpin and Soule´ 1986). nificantly correlated with population size, suggesting the The predicted relationship between plant population size action of genetic drift and/or inbreeding. No correlation and genetic diversity has been documented in numerous was found between genetic and geographic distances. studies (Dittbrenner et al. 2005; Fischer and Matthies 1998; Natural barriers like forests may have been efficiently Gaudeul et al. 2000; Hensen et al. 2005; Hensen and preventing seed migration even between geographically Oberprieler 2005; van Treuren et al. 1991; Vergeer et al. closer populations. Results of this study suggest that 2003), whereas some studies reveal no such correlation genetic erosion could be partially responsible for the lower (Bachmann and Hensen 2007; Hensen et al. 2010; Kahmen fitness in smaller populations of this species. and Poschlod 2000; Pluess and Sto¨cklin 2004; Tero et al. 2003). The negative effect of small population size on Keywords AFLP Genetic diversity Ligularia sibirica genetic variation is often stronger in predominantly out- Population genetic structure Seed production crossing and self-incompatible plant species (Honnay and Small population Jacquemyn 2007; Leimu et al. 2006), possibly in connec- tion with reduced pollinator activity with decreased number of flowering plants (Kwak et al. 1998). Genetically eroded populations can be ‘‘rescued’’ by efficient gene flow between them owing to migration of pollen and seeds & A. Ilves ( ) K. Lanno M. Sammul K. Tali (Ellstrand and Elam 1993; Loveless and Hamrick 1984). Institute of Agricultural and Environmental Sciences, Estonian University of Life Sciences, Riia 181, 51014 Tartu, Estonia Reduction in population size is often accompanied with e-mail: [email protected] frequent matings among relatives, leading to heightened 123

87 662 Conserv Genet (2013) 14:661–669 homozygosity and inbreeding depression (Frankham et al. Materials and methods 2010). Several traits related to plant fitness are shown to be affected by declining population size: e.g., seed production Study species decreased with population size of rare Gentianella germanica (Fischer and Matthies 1998) and Dracocephalum aus- Ligularia sibirica (L.) Cass. is a tall (50–150 cm) perennial triacum (Dosta´lek et al. 2010), plants produced fewer and herb, the life-span of its genet can be more than 10 years. It lighter seeds in smaller populations of Dictamnus albus has a short slowly growing rhizome, kidney-shaped long- (Hensen and Oberprieler 2005) and Pulsatilla vulgaris stalked lower leaves and much smaller stalkless upper (Hensen et al. 2005); germination was less successful in leaves. In Estonia, it flowers at the end of July and in smaller populations of Silene regia (Menges 1991) and August, forming 1–8 erect flowering stems with numerous Succisa pratensis (Vergeer et al. 2003). The negative effect (10–30; sometimes even 50) yellow flower heads. This of low genetic variation and small population size on plant species is mainly insect-pollinated but is capable of self- fitness and thus increased probability of extinction was fertilization (M. Sammul, pers. obs.). The seeds are sup- recently confirmed by meta-analysis based on studies pub- plied with pappus and mature in late Aug. or in Sept. lished between 1987 and 2005 (Leimu et al. 2006). There- The main distribution area of L. sibirica is in Russia, fore, assessing the survival probabilities of endangered plant extending from the European part of Russia over Siberia to species, it is important to carry out studies where both plant the Far East (Hulteń and Fries 1986). In Europe, this spe- performance and genetic aspects are analyzed. cies is considered to be a postglacial relict, occurring in Ligularia sibirica (L.) Cass. (Asteraceae) is a rare and isolated localities in Estonia, Latvia, Ukraine, Poland, endangered plant species everywhere it is found in Europe. Croatia, Romania, Czech and Slovak Republics, Hungary, In Estonia, its populations have primarily suffered from Bulgaria, Austria and France (Hegi 1987; Hulteń and Fries human-caused changes in quality and availability of habi- 1986). L. sibirica is listed in the Annex II of the EU tats (primarily overgrowth by trees and shrubs coinciding Habitats Directive (Council Directive 92/43/EEC, 1992). with worsened light conditions and increased competition, In Estonia, L. sibirica is at the north-western edge of its as well as the drainage of former wet habitats). The number main distribution range. Typical habitats here for this of localities of this rare plant has decreased from 18 in species are paludified grasslands, floodplain grasslands, 1969 to 9 (Kukk 2003; Kukk and Kull 2005) and most of forest plains and spring fens (Kukk 2003). its populations continue to decline. In several populations the decrease in the number of juveniles and generative Sampling of populations, seeds and germination plants has been reported (Kukk 2003). In some populations, there have been attempts to restore the habitat and in some Seven populations differing largely in number of L. sibirica others to reinforce the population size by introducing were included in the study: Anne, Jo˜hvi, O˜ isu, Sootaga, ex-situ nourished new individuals originating from the Tagula, Pressi and Vaä¨gvere (Fig. 1). Two other popula- same populations (M. Sammul, pers. inform.). The success tions, Kikaste and A¨ dise, were left out because of the very of these actions has been inconsistent, however, in order to small size of these populations (consisting of only some evaluate the potential of a success of conservation actions, remaining vegetative individuals). The sizes of these pop- it is essential to first understand the intrinsic genetic and ulations were estimated through counting the number of demographic hindrances to population growth. tussocks, which in most cases corresponds to the number of The main objective of this study was to find out if the genets. For AFLP analysis, we collected 20 healthy leaves observed decline in population size has induced a decrease in per population from randomly selected generative plants in the amount of genetic variation as well as reproductive plant summer 2009 (except in Anne where we found only seven performance in the populations of L. sibirica. The amplified plants) and stored in silica-gel bags. In 2007 and 2008, we fragment length polymorphism (AFLP) technique was collected the seeds from 15 to 25 randomly selected plants selected for its ability to create a large amount of markers in a in all populations except in Anne and Jo˜hvi, where fewer short time without any previous knowledge of genome of generative plants were available. In 2008, Pressi population plant species (Meudt and Clarke 2007). Specifically, we could not be sampled. We regarded seeds that were com- asked the following questions: (i) what is the overall level of pletely filled and fully developed as viable in contrast to genetic diversity and how is it distributed within and among empty, non-developed seeds and those damaged by insect populations of L. sibirica? (ii) is there an association between larvae or other parasites. The number of viable seeds per within-population genetic diversity and population size? (iii) flower stem was used as one measure for plant fitness. is the plant fitness affected by population size and genetic The germination rates were estimated using seeds col- variation? On the basis of our results, we proposed the lected in 2007. For most populations, we used the healthy- guidelines for reinforcement of populations of L. sibirica. looking seeds from five plants (except for Jo˜hvi and Anne 123

88 Conserv Genet (2013) 14:661–669 663

replicates originating from independent DNA extractions which yielded the average reproducibility of 97.9 %. Neg- ative controls were run at each step of the process to check for contaminations.

Data analysis

AFLP electropherograms were scored as present (1) or absent (0) using software GeneMapperÒ 4.0 (Applied Biosystems, USA). All profiles were checked visually and only these AFLP fragments that could be scored unam- biguously were included in the analysis. The amount of polymorphic fragments and these present only in one population (private bands) was determined. We excluded fragments that were monomorphic across all individuals, Fig. 1 Location of seven studied L. sibirica populations in Estonia. samples with odd profiles and AFLP fragments present or Filled symbols: sampled populations. A Anne, J Jo˜hvi, V Vaä¨gvere, S Sootaga, T Tagula, P Pressi, O˜ O˜ isu. Empty symbols: remnant non- absent only in one individual to reduce genotypic errors sampled populations. A¨ A¨ dise, K Kikaste (Bonin et al. 2004). Because of the slight shift between peaks among the small size AFLP fragments and the assumption that smaller fragments are more likely to be where we used seeds from four and three plants, respec- homoplasious (Paris et al. 2010; Vekemans et al. 2002), we tively), the total number of seeds per population ranged used the size range of 90–500 base pairs. between 64 and 960. Germination was carried out in Petri To visualize the general patterns among populations, the dishes on wet filter paper in the laboratory at fluctuating presence/absence matrix was subjected to a principal day/night length of 14/10 h and temperature of 27/17 °C. coordinates analysis (PCoA) based on Jaccard’s coefficient of similarity using the software PCO (Anderson 2003). AFLP analysis Jaccard’s coefficient does not take into account shared band absences and is therefore adequate for analyzing DNA was extracted from approximately 100 mg of dried dominant markers such as AFLP (Duarte et al. 1999). leaf material using CTAB procedure according to Doyle and Genetic diversity within populations was assessed as Doyle (1987). We checked the quality of obtained DNA on a (i) the percentage of polymorphic loci in a given population 1.5 % agarose gel and excluded samples with poor DNA and (ii) average gene diversity over loci as defined by the quality from further analysis. AFLP procedures were per- software Arlequin 3.5.1.2 (Excoffier and Lischer 2010). To formed as described by Vos et al. (1995) with slight modi- measure the degree of genetic differentiation among popu- fications. Instead of 500 ng of genomic DNA, 300 ng was lations, we used a Bayesian approach presented by Holsinger (II) restricted with 1 UTru1I (isoschizomer of MseI) at 65 °C for et al. (2002) to calculate FST analogue h by the software 2 h, followed by 5 U EcoRI digestion (both restrictases from Hickory 1.1 (Holsinger and Wallace 2004). This method is Fermentas, Lithuania) and ligation of double-stranded oli- appropriate for analysing data derived from dominant gonucleotide adapters with 1 U of T4 DNA-Ligase (Naxo, markers because neither are assumptions about Hardy– Estonia) in 29 TangoTM buffer (also provided by Fermen- Weinberg equilibrium made nor is previous knowledge tas) at 20 °C overnight. Preselective amplification was per- about the level of inbreeding in populations needed. Alter- formed using primer pairs with a single selective nucleotide, native models implemented in this program correspond to EcoRI-A and MseI-C. For selective amplification, we tested different hypotheses: there is no inbreeding (f = 0 model); 16 primer combinations with three selective nucleotides there is no genetic structure (h(II) = 0); neither h(II) nor f are on five individuals. We selected three primer pairs that known (full model). In addition to these three models, f-free amplified most readable AFLP profiles: EcoRI-ACT(FAM)- model provides h(II) values without attempting to estimate MseI-CAT; EcoRI-AGC(NED)-MseI-CAG; EcoRI-AGG f. To test the consistency of results, all models were run three (JOE)-MseI-CAC. Separation of amplification products was times with default sampling parameters (burn-in = 50,000, performed by capillary electrophoresis on an ABI3130 sample = 250,000, thin = 50). Another analogue of FST, Genetic Analyzer (Applied Biosystems, USA). GeneScan namely UST was calculated using analysis of molecular 500 RoxTMSize Standard (Applied Biosystems, USA) was variance (AMOVA) by the software Arlequin 3.5.1.2 (Ex- loaded in each lane. To estimate the reproducibility of the coffier and Lischer 2010). AMOVA is based on squared genotyping process, we compared the profiles from five Euclidean distances among molecular phenotypes and 123

89 664 Conserv Genet (2013) 14:661–669

enables to partition genetic variation within and among correlation coefficient rS = 0.96, p \ 0.001). Both esti- populations (Excoffier et al. 1992). In addition to overall UST, mates of within-population genetic diversity correlated sig- genetic differentiation between pairs of populations (pairwise nificantly with approximate population size (rS = 0.78, UST) was estimated. Using the zt 1.1 software (Bonnet and p \ 0.05 for average gene diversity over loci and log pop- Van de Peer 2002), a Mantel test was performed to find out ulation size and rS = 0.88, p \ 0.05 for proportion of whether the matrix of genetic distances based on linearised polymorphic loci and log population size). pairwise UST values was correlated with the matrix of log- transformed geographical distances. In both AMOVA anal- Genetic structure ysis and Mantel test, the significance levels were calculated by conducting permutation procedures (1,000 permutations). In PCoA, individuals within populations were mainly We tested the differences in the (log-transformed) amount grouped together, although there was some overlap in the of viable seeds per flower stem among populations and years ordination space: plants from Jo˜hvi and Vaä¨gvere showed a (2007 and 2008) by two-way ANOVA without replicates. To strong genetic similarity to each other, also individuals from investigate the relationship between population size, genetic Sootaga and Tagula were clustered together (Fig. 2). The diversity and plant fitness characters, we calculated non- three geographically closest populations—Sootaga, Anne parametric Spearman correlation coefficients rS using R and Vaä¨gvere—did not group together in PCoA; although version 2.11.1 (R Development Core Team 2007). Anne and Vaä¨gvere were genetically quite similar, Sootaga took place at the opposite end along PCoA 2-axis. Plants in O˜ isu population formed a distinct cluster which was com- Results pletely separate from other populations. This general pattern in PCoA was supported by the Mantel test, which revealed no AFLP pattern and genetic diversity significant correlations between genetic and geographic

distances (rm = 0.19, p = 0.28, 1,000 permutations; Fig. 3). AFLP analysis was successfully performed on 85 plants Bayesian analysis of the population structure resulted in from 7 populations, resulting in a unique AFLP banding similar h(II) values for different models. According to the pattern for every individual. Three primer pairs generated in deviance information criterion (DIC), the full model proved total 258 scorable fragments, 199 (77 %) of which were to be the most suitable for the data, having the lowest DIC polymorphic. The mean number of fragments per AFLP value (3,292.05) and yielding estimates of 0.398 (SD 0.032) profile were (±SD): 28(±2.7) for EcoRI-ACT(FAM)-MseI- and 0.704 (SD 0.031) for h(II) and f, respectively. Consid- CAT, 45(±4.4) for EcoRI-AGC(NED)-MseI-CAG and erably worse fitted the data in the f = 0 model (no 40(±6.1) for EcoRI-AGG(JOE)-MseI-CAC. The proportion inbreeding) which had a difference of more than 100 DIC of polymorphic loci varied between 24 and 67 %, being units from the full model (DIC 3,397.37), providing strong lowest in Anne where no private alleles were found and evidence for the occurrence of inbreeding in L. sibirica highest in Tagula where the number of private bands was also populations. The comparison of posterior distributions highest (Table 1). Mean gene diversity over loci of all pop- showed no significant difference between h(II) estimates in ulations was 0.167, varying between 0.109 and 0.230. There these models (0.0060 with 95 % credible intervals -0.0765; was a strong correlation between proportion of polymorphic 0.0637), indicating that the obtained h(II) was not affected by loci and average gene diversity over loci (Spearman rank the evaluating inbreeding coefficient in the full model.

Table 1 Population size, sample size, genetic diversity estimates, proportion of germinated seeds and mean amount of viable seeds per ﬂower stem in the studied populations of L. sibirica Site Population No. of Average Proportion No. of Germination % Mean amount of viable seeds (SD)a size samples gene diversity of polymorphic private (SD)a over loci loci (%) alleles 2007 2008

Anne 7 7 0.109 24 0 41 (21) 24 (17) 9 (7) Jo˜hvi 70 9 0.138 31 1 45 (17) 21(6) 57 (75) Pressi 400 11 0.181 51 4 59 (30) 109 (98) – Va¨a¨gvere 400 13 0.181 52 3 52 (9) 195 (145) 161 (83) O˜ isu 600 17 0.115 38 2 65 (22) 22 (18) 52 (32) Sootaga 1,000 13 0.217 63 3 73 (14) 93 (72) 149 (81) Tagula 1,500 15 0.230 67 5 68 (22) 40 (35) 196 (99) a Sample size for germination % and mean amount of viable seeds is as described in ‘‘Materials and methods’’

123

90 Conserv Genet (2013) 14:661–669 665

largely damaged and eaten in Tagula, Sootaga and O˜ isu which resulted in lower seed counts compared to the amount of viable seeds produced in 2008 in these populations (Table 1). Lowest seed numbers were measured in Anne for both years. Mean viable seed production in 2008 was positively correlated with both average gene diversity over loci and the proportion of polymorphic loci

(rS = 0.94, p \ 0.05 and rS = 0.89, p \ 0.05, respectively), there was no correlation present in 2007. Correla- tions with population size were not signiﬁcant in either year. The germination % ranged between 41 and 73 % (Table 1), the average for all populations was 58 % (SD ± 22 %). Germination % was positively and signiﬁ- cantly correlated with the proportion of polymorphic loci

(rS = 0.82, p \ 0.05) and population size (rS = 0.95, Fig. 2 Principal Coordinate Analysis (PCoA) of the 85 L. sibirica p \ 0.001). The relationship between average gene diver- individuals based on Jaccard coefﬁcient of similarity. X- and Y-axis explain 18.5 and 12.5 % of the total variation, respectively sity over loci and germination % was very close to being signiﬁcant (rS = 0.74, p = 0.058).

Discussion

Genetic diversity and population differentiation

Among several life-history traits, the breeding system is considered to be among the most important to determine the level of genetic variability and its distribution in populations of plant species (Duminil et al. 2007; Nybom and Bartish 2000). Plants with mixed-mating breeding system exhibit much lower levels of within-population gene diversity compared to outcrossing plants (Hamrick and Godt 1989; Nybom 2004). Genetic diversity in examined seven L. sibirica populations was relatively low (mean values 0.167 and 46.5 % for gene diversity and proportion

Fig. 3 The relationship between pairwise population FST values and of polymorphic loci, respectively) and slightly lower than logarithm of geographical distance (km). The relationship was not the average for mixed-mating species measured by RAPD = = signiﬁcant (Mantel test rm 0.19, p 0.28, 1,000 permutations) (0.18) presented by Nybom (2004). This implies that self- pollination or pollination between close relatives can play AMOVA analysis revealed similar levels of genetic quite a substantial part in reproduction of this species. differentiation among populations of L. sibirica: UST value Intrapopulation diversity indices of the same range for was 0.41; 59 % of total molecular variation was attribut- endangered plant species with similar breeding system are able to the variation within populations and 41 % was due obtained also by AFLP techniques, e.g., 0.134–0.234 for to the differences among populations. All values obtained Eryngium alpinum (Gaudeul et al. 2000); 0.172–0.229 for in AMOVA were highly signiﬁcant (p \ 0.001, 1,000 Silene chlorantha (Lauterbach et al. 2011). permutations). Pairwise UST values varied between 0.57 For outcrossing plants, most of total genetic variability is and 0.25 (Table 2). distributed among the individuals within a population, a smaller proportion of it is attributable to the variation Seed production and germination between populations (Hamrick and Godt 1989). On average

27 % (UST 0.27) of RAPD diversity partitions between There were significant differences in viable seed produc- populations when the species is predominantly outbreeding, tion between populations (F6, 227 = 23.8, p \ 0.001) and the mean value of UST is significantly higher (UST 0.4) for years (F1, 227 = 8.7, p \ 0.001). In 2007, seeds were partially selfing plants (Nybom 2004). Thus, UST of 0.41 123

91 666 Conserv Genet (2013) 14:661–669

Table 2 Geographical distances (km; above the diagonal) and pairwise estimated UST values (below the diagonal) between seven populations of L. sibirica Anne Jo˜hvi Pressi Va¨a¨gvere O˜ isu Sootaga Tagula

Anne – 115.7 78.5 16.8 73.3 13.2 61.3 Jo˜hvi 0.506*** – 185.2 99.5 167.6 104.2 177.0 Pressi 0.460*** 0.385*** – 90.0 109.8 91.5 48.0 Vaä¨gvere 0.398*** 0.346*** 0.373*** – 86.4 11.1 77.7 O˜ isu 0.571*** 0.539*** 0.476*** 0.468*** – 75.9 62.9 Sootaga 0.416*** 0.426*** 0.330*** 0.343*** 0.448*** – 73.1 Tagula 0.445*** 0.369*** 0.338*** 0.325*** 0.474*** 0.245*** – *** Level of significance p \ 0.001, based on 1,023 permutations obtained in AMOVA in this study is in accordance with and the small fraction which is carried high up and far by average for mixed-mating plants and together with the the strong wind is very likely to finally set down on an results of the Bayesian analysis (h(II) 0.398) indicates quite a unsuitable environment where establishment is unlikely high level of differentiation between L. sibirica populations. (Sheldon and Burrows 1973). As no correlation was found Much lower between-population diversity indices together between genetic and geographical distances, it is possible with high variability within populations of L. sibirica was that natural barriers like forests have been efficiently pre- recently found in the Czech and Slovak Republic by using venting seed migration even between geographically closer allozymes (Sˇmı´dova´ et al. 2010). They estimated the obtained populations located on the same river system, Vaä¨gvere,

FST value (0.179) as typical for predominantly outcrossing Sootaga and Anne. Similarly, the high pairwise UST values species as presented by Hamrick and Godt (1989). However, (Table 2) imply limited gene exchange between L. sibirica comparing results from studies using different marker sys- populations. The highest UST values were obtained tems like allozymes and genome-wide selectively neutral between Anne and O˜ isu (0.571) and Jo˜hvi and O˜ isu DNA markers (AFLPs and RAPDs) should be avoided. It has (0.539). Also closely situated populations were genetically been suggested that differences in between-population esti- quite differentiated (UST values between Anne, Sootaga mates obtained both from allozyme and RAPD analysis of and Vaä¨gvere ranging between 0.343 and 0.416). Compa- the same populations most likely show differences in rates of rable pattern was obtained in PCoA: plants from geo- mutations at RAPD and allozyme loci rather than reflect graphically closer populations were genetically not so different migration rates (Holsinger and Wallace 2004). similar and O˜ isu was placed completely separately from Allozymes can also be subject to directional selection pres- other populations (Fig. 2). The O˜ isu population is also sure which may enhance the action of genetic drift (Baruffi conspicuous by its very low genetic diversity indices et al. 1995) or be influenced by balancing selection which despite of its considerable size (Table 1). Although the first maintains allozyme polymorphisms (Altukhov 1991). records about single L. sibirica plants in the O˜ isu area are One explanation to high levels of differentiation among from 1860, the main population was discovered in 2003 populations can be restricted or absent gene flow by pollen and its previous history is unknown. Therefore, it is not and seeds (Fischer and Matthies 1998; Hensen et al. 2010; impossible that this population was founded by only a few Hogbin and Peakall 1999; Lauterbach et al. 2011; Schmidt individuals and a low level of genetic diversity has been and Jensen 2000). Pollen flow mediated by insects is maintained through insufficient gene exchange with other generally limited since most pollinators can travel less than populations. In addition to that, almost all populations have 1 km (Kwak et al. 1998) and they tend to visit neigh- maintained private alleles (Table 1), confirming the idea bouring plants (Rasmussen and Brødsgaard 1992). Popu- that populations of L. sibirica might experience very little lations of L. sibirica in Estonia are most likely too far from migration between them and have been under the influence each other to be connected by pollen exchange between of random genetic drift. populations. Gene flow by seeds is more realistic as seeds of this species are supplied with pappus and are adapted to Genetic diversity, population size and plant fitness dispersal by wind, and given the location of several populations on alluvial riverbanks they could also be distrib- In L. sibirica populations, both average gene diversity over uted by flowing water. However, most of the seeds of this loci and proportion of polymorphic loci were significantly species are not considered to travel long distances, landing correlated with population size. This is in accordance with close to the parent plant (Kobiv 2005,Sˇmı´dova´ et al. 2010) the theory that especially small isolated populations are

123

92 Conserv Genet (2013) 14:661–669 667 susceptible to processes such as drift and inbreeding and unsuccessful reproduction in a bad year by producing more also in agreement with high levels of differentiation found seeds in a good year, small populations with reduced in L. sibirica populations. Lowered heterozygosity and genetic diversity might have lost this property. inbreeding are shown to have a negative impact to plant Environmental stress and poor habitat quality can be one fitness (Ellstrand and Elam 1993). Unfortunately, direct possible factor contributing to the reduced plant perfor- estimation of inbreeding is not possible from AFLP data. mance in some populations of L. sibirica. Sub-optimal The f value derived from the Bayesian analysis is hardly growth conditions are shown to lead to diminished mater- reliable due to the dominant nature of the marker system nal investment into the offspring, affecting reproductive and should be regarded with caution (Holsinger and Lewis success (Oostermeijer et al. 1994). In addition to direct 2003–2007). Nevertheless, since the Bayesian model with effect of limited resources availability, it has been proved no inbreeding did not fit the data as well as the full model, that the impact of inbreeding is more severe under stressful the existence of inbreeding (and inbreeding depression) conditions (Armbruster and Reed 2005; Reed et al. 2002). cannot be ruled out in smaller populations of L. sibirica. Seed production can also be largely dependent on pollen Therefore, lowered levels of genetic diversity and availability (Byers 1995; Fischer and Matthies 1997). inbreeding depression can be considered as one factor In L. sibirica populations, the influence of environment is influencing the seed production and germination ability of most obvious in Anne where plants suffer from overly dry this species. However, since molecular markers such as soil conditions, herbivores and competition with high- AFLP are selectively neutral and may lose genetic varia- growing plant species (e.g., Filipendula ulmaria) which tion at different speed compared to loci connected with suppress the growth of Ligularia (K. Lanno, pers. obs.). fitness, the correlation between genetic diversity and fitness Thus both environmental and genetic factors can explain can be rather weak and explain only a small proportion of the low reproductive success in this population. variation in fitness (Reed and Frankham 2003). Therefore, Since it is not possible to disentangle the complex effect the loss of fitness should not be automatically inferred from of genetic and various environmental factors, we cannot the loss of genetic variation measured with molecular claim that genetic erosion is playing a major part in markers. Nevertheless, reductions in plant seed production reproductive success in L. sibirica populations. However, in combination with AFLP or RAPD variation are descri- the results of this study show that it could be one factor bed in several studies (Dittbrenner et al. 2005; Fischer and affecting plant fitness in smaller populations of this species. Matthies 1998; Hensen and Oberprieler 2005; Schmidt and Jensen 2000). In this study, germination percentage was in Conservation implications positive relationship with one estimate of genetic variability (proportion of polymorphic loci). Unfortunately, the Populations of L. sibirica have suffered because of human- sample size used in germination experiment was relatively induced changes in their habitats. Therefore, there is a small which could have influenced the relationships moral obligation to undo the harm. Reinforcement of between germination percentage and genetic diversity populations by introducing new individuals could be one of estimates. We also found a strong positive correlation the measures to support population recovery after the between the amount of viable seeds in 2008 and both habitat has been restored. However, there are two different genetic diversity estimates. In 2007, the correlation was approaches to be considered. Firstly, as the isolated pop- missing: the inflorescences of most plants were seriously ulations at the distribution margins are often genetically damaged by insect larvae in several populations (mostly in separated, one can argue that such distribution pattern Tagula and Sootaga) which resulted in low levels of seed should be preserved and considered as a natural value counts. Unfortunately, it was not possible to estimate the (Ehrlich 1988; Lesica and Allendorf 1995). In this case, amount of developed seeds consumed by parasites. plant material used for reinforcement must originate from Therefore, the seed counts in these populations do not truly the same population. On the other hand, if the population is represent the plants potential to develop healthy, filled genetically impoverished and inbreeding is one of the seeds and it is probable that the missing correlation in 2007 causes of lowered fitness, it may not be sufficient to was conditioned by the observed damage reducing seed overcome limitations caused by lack of genetic variability counts drastically in these populations. The levels of seed by using local plant material (Edmands 2007). In such production in small populations with the lowest genetic cases reinforcement must be concentrated on introducing diversity indices (Anne, Jo˜hvi and O˜ isu) were quite low new genotypes (see Sackville Hamilton 2001 and Vander even when the year was good (Table 1), showing the Mijnsbrugge et al. 2010 for further discussion). possibility that there can be a link between genetic factors Given that there were clear signs of isolation of popula- and reproductive fitness in small populations of this spe- tions of L. sibirica in this study we would argue for pre- cies. While plants in large populations can buffer serving the diversity of genetically different populations and 123

93 668 Conserv Genet (2013) 14:661–669 avoiding the mixing of genotypes from different popula- Dosta´lek T, Mu¨nzbergova Z, Placˇkova´ I (2010) Genetic diversity and tions. However, introducing new genotypes into the smaller its effect on fitness in an endangered plant species, Dracoceph- alum austriacum L. Conserv Genet 11:773–783 populations may be unavoidable because of the possible link Doyle JJ, Doyle JL (1987) A rapid DNA isolation procedure for small between small population size, small genetic diversity and quantities of fresh leaf tissue. Phytochem Bull 19:11–15 lowered fitness revealed in this study. Based on our data, we Duarte JM, dos Santos JB, Melo LC (1999) Comparison of similarity suggest the introduction of new genotypes into L. sibirica coefficients based on RAPD markers in the common bean. Genet Mol Biol 22:427–432 populations in Estonia when population size is less than 100 Duminil J, Fineschi S, Hampe A, Jordano P, Salvini D, Vendramin genets and average gene diversity over loci is lower than GG, Petit RJ (2007) Can population genetic structure be 0.15. If one of these conditions is not met, only local plant predicted from life-history traits? Am Nat 169:662–672 material should be used in the population reinforcement. Edmands S (2007) Between rock and a hard place: evaluating the relative risks of inbreeding and outbreeding for conservation and Considerable isolation of populations of L. sibirica also management. Mol Ecol 16:463–475 imposes the need to conservationally treat them as isolated Ehrlich P (1988) The loss of diversity: causes and consequences. In: populations. This implies that we cannot handle the system Wilson EO (ed) Biodiversity. National Academy Press, Wash- as a metapopulation and in order to persist, each population ington, pp 21–27 Ellstrand NC, Elam DR (1993) Population genetic consequences of individually must exceed the limit of a viable population small population size: Implications for plant conservation. Annu size. We have detected both reduced genetic diversity as Rev Eco Syst 24:217–242 well as reduced plant fitness in smaller populations of this Excoffier L, Lischer HEL (2010) Arlequin suite ver 3.5: a new series species. Thus, our data suggest that populations with less of programs to perform population genetics analyses under Linux and Windows. Mol Ecol Resour 10:564–567 than 200 individuals need special attention as they may be Excoffier L, Smouse PE, Quattro JM (1992) Analysis of molecular prone to enter the extinction vortex. variance inferred from metric distances among DNA haplotypes: application to human mitochondrial DNA restriction data. Acknowledgments This study was funded by Grants No. G7631 Genetics 131:479–491 from Estonian Science Foundation and No. P6062PKPK06 from the Fischer M, Matthies D (1997) Mating structure and inbreeding and Estonian University of Life Sciences. We appreciate the help from the outbreeding depression in the rare plant Gentianella germanica herbarium of the Botany Department of the Estonian University of (Gentianaceae). Am J Bot 84:1685–1692 Life Sciences (TAA). We thank all who helped with field- and lab- Fischer M, Matthies D (1998) RAPD variation in relation to work, especially Thea Kull, Katre Hein and Erkki Sild. population size and plant fitness in the rare Gentianella germanica (Gentianaceae). Am J Bot 85:811–819 Frankham R, Ballou JD, Briscoe DA (2010) Introduction to conserv genet, 2nd edn. Cambridge University Press, Cambridge References Gaudeul M, Taberlet P, Till-Bottraud I (2000) Genetic diversity in an endangered alpine plant, Eryngium alpinum L. (Apiaceae), Altukhov YuP (1991) The role of balancing selection and overdom- inferred from amplified fragment length polymorphism markers. inance in maintaining allozyme polymorphisms. Genetica Mol Ecol 9:1625–1637 85:79–90 Gilpin ME, Soule´ ME (1986) Minimum viable populations: processes Anderson M (2003) PCO: A Fortran computer program for principal of species extinction. In: Soule´ ME (ed) Conserv Biol: the coordinate analysis. University of Auckland, New Zealand, science of scarcity and diversity. Sinauer associates Inc., Distributed by the Author Sunderland, pp 19–34 Armbruster P, Reed DH (2005) Inbreeding depression in benign and Hamrick JL, Godt JW (1989) Allozyme diversity in plant species. In: stressful environments. Heredity 95:235–242 Brown AHD, Clegg MT, Kahler AL, Weir BS (eds) Plant Bachmann U, Hensen I (2007) Is declining Campanula glomerata population genetics breeding and genetic resources. Sinauer threatened by genetic factors? Plant Species Biol 22:1–10 associates Inc., Sunderland, pp 43–66 Baruffi L, Damian G, Guglielmino CR, Bandis C, Malacrida AR, Hegi G (1987) Illustrierte Flora von Mitteleuropa, teil 4. Verlag Paul Gasperi G (1995) Polymorphism within and between populations Parey, Berlin und Hamburg, pp 796–799 of Ceratitis capitata, comparison between RAPD and multilocus Hensen I, Oberpieler C (2005) Effects of population size on genetic enzyme electrophoresis data. Heredity 74:425–437 diversity and seed production in the rare Dictamnus albus Bonin A, Bellemain E, Bronken Eidesen P, Pompanon F, Brochmann (Rutaceae) in central Germany. Conserv Genet 6:63–73 C, Taberlet P (2004) How to track and assess genotypic errors in Hensen I, Oberprieler C, Wesche K (2005) Genetic structure, population genetics studies. Mol Ecol 13:3261–3273 population size, and seed production of Pulsatilla vulgaris Mill. Bonnet E, Van de Peer Y (2002) zt: a software tool for simple and (Ranunculaceae) in central Germany. Flora 200:3–14 partial Mantel tests. J Stat Software 10:1–12 Hensen I, Kilian C, Wagner V, Durka W, Pusch J, Wesche K (2010) Brook BW, Sodhi NS, Bradshaw CJA (2008) Synergies among extinction Low genetic variability and strong differentiation among isolated drivers under global change. Trends Ecol Evol 23:453–460 populations of the rare steppe grass Stipa capillata L. in central Byers DL (1995) Pollen quantity and quality as explanations for low Europe. Plant Biol 12:526–563 seed set in small populations exemplified by Eupatorium Hobbs RJ, Yates CJ (2003) Impacts of ecosystem fragmentation on plant (Asteraceae). Am J Bot 82:1000–1006 populations: generalising the idiosyncratic. Aust J Bot 51:471–488 Dittbrenner A, Hensen I, Wesche K (2005) Genetic structure and Hogbin PM, Peakall R (1999) Evaluation of the contribution of random amplified polymorphic DNA diversity of the rapidly genetic research to the management of the endangered plant declining Angelica palustris (Apiaceae) in Eastern Germany in Zieria prostrata. Conserv Biol 13:514–522 relation to population size and seed production. Plant Species Holsinger KE, Lewis PO (2003–2007) Hickory: a package for analysis Biol 20:191–200 of population genetics data v. 1.1. University of Conneticut, Storrs

123

94 Conserv Genet (2013) 14:661–669 669

Holsinger KE, Wallace LE (2004) Bayesian approaches for the Paris M, Bonnes B, Ficetola GF, Poncet BN, Despre´s L (2010) analysis of population genetic structure: an example from Amplified fragment length homoplasy: in silico analysis for Platanthera leucophaea (Orchidaceae). Mol Ecol 13:887–894 model and non-model species. BMC Genomics. doi:10.1186/ Holsinger KE, Lewis PO, Dipak KD (2002) A Bayesian approach to 1471-2164-11-287 inferring population structure from dominant markers. Mol Ecol Pluess AR, Sto¨cklin J (2004) Genetic diversity and fitness in Scabiosa 11:1157–1164 columbaria in the Swiss Jura in relation to population size. Honnay O, Jacquemyn H (2007) Susceptibility of common and rare Conserv Genet 5:145–156 plant species to the genetic consequences of habitat fragmenta- Rasmussen IR, Brødsgaard B (1992) Gene flow inferred from seed tion. Conserv Biol 21:823–831 dispersal and pollinator behaviour compared to DNA analysis of Hulteń E, Fries M (1986) Atlas of North European vascular plants. restriction site variation in a patchy population of Lotus North of the Tropic of cancer. II. Koeltz Scientific Books, corniculatus L. Oecologia 89:277–283 Ko¨ningstein, p 925 Reed DH, Frankham R (2003) Correlation between fitness and genetic Kahmen S, Poschlod P (2000) Population size, plant performance, diversity. Conserv Biol 17:230–237 and genetic variation in the rare plant Arnica montana L. in the Reed DH, Briscoe DA, Frankham R (2002) Inbreeding and extinction: Rho¨n. Germany. Basic Appl Ecol 1:43–51 The effect of environmental stress and lineage. Conserv Genet Kobiv Yu (2005) Ligularia sibirica (L.) Cass. (Asteraceae) in the 3:301–307 Chornohora mountains (Ukrainian Carpathians): population- Sackville Hamilton NR (2001) Is local provenance important in ontogenetic parameters, morphology, taxonomy and conserva- habitat creation? A reply. J Appl Ecol 38:1374–1376 tion. Ukrainian Bot J 62:383–395 Schleuning M, Matthies D (2009) Habitat change and plant demog- Kukk U¨ (2003) The distribution of Ligularia sibirica (L.) Cass. in raphy: assessing the extinction risk of a formerly common Estonia and changes in its population. Biuletyn Ogrodo´w grassland perennial. Conserv Biol 1:174–183 Botanicznych 12:11–22 Schmidt K, Jensen K (2000) Genetic structure and AFLP variation of Kukk T, Kull T (eds) (2005) Atlas of the Estonian flora. Institute of remnant populations in the rare plant Pedicularis palustris Agricultural and Environmental Sciences of the Estonian (Scrophulariaceae) and its relation to population size and University of Life Sciences, Tartu, p 90 reproductive components. Am J Bot 87:678–689 Kwak MM, Velterop O, Andel J (1998) Pollen and gene fow in Sheldon JC, Burrows FM (1973) The dispersal effectiveness of the fragmented habitats. Appl Veg Sci 1:37–54 achene-pappus units of selected compositae in steady wins with Lauterbach D, Ristow M, Gemeinholzer B (2011) Genetic population convection. New Phytol 72:665–672 structure, fitness variation and the importance of population history Sˇmı´dova´ A, Mu¨nzbergova´ Z, Plakova´ I (2010) Genetic diversity of a in remnant populations of the endangered plant Silene chlorantha relict plant species, Ligularia sibirica (L.) Cass. (Asteraceae). (Willd.) Ehrh. (Caryophyllaceae). Plant Biol 13:667–677 Flora 206:151–157 Leimu R, Mutikainen P, Koricheva J, Fischer M (2006) How general R Development Core Team (2007) R: A language and environment are positive relationships between plant population size, fitness for statistical computing. R Foundation for Statistical Comput- and genetic variation? J Ecol 94:942–952 ing, Vienna, Austria. http://www.R-project.org Lesica P, Allendorf FW (1995) When are peripheral populations Tero N, Aspi J, Siikama¨ki A, Tuomi J (2003) Genetic structure and valuable for conservation? Conserv Biol 9:753–760 gene flow in a metapopulation of an endangered plant species, Loveless MD, Hamrick JL (1984) Ecological determinants of genetic Silene tatarica. Mol Ecol 12:2073–2085 structure in plant populations. Annu Rev Eco Syst 15:65–95 van Treuren R, Bijlsma R, van Delden W, Ouborg NJ (1991) The Menges ES (1991) Seed germination percentage increases with significance of genetic erosion in the process of extinction. population size in a fragmented prairie species. Conserv Biol I. Genetic differentiation in Salvia pratensis and Scabiosa 5:158–164 columbaria in relation to population size. Heredity 66:181–189 Meudt HM, Clarke AC (2007) Almost forgotten or latest practice? Vander Minsbrugge K, Bischoff A, Smith B (2010) A question of AFLP applications, analyses and advances. Trends Plant Sci origin: where and how to collect seed for ecological restoration. 12:106–117 Basic Appl Ecol 11:300–311 Nybom H (2004) Comparison of different nuclear DNA markers for Vekemans X, Beauwens T, Lemaire M, Roldań-Ruiz I (2002) Data estimating intraspecific genetic diversity in plants. Mol Ecol from amplified fragment length polymorphism (AFLP) markers 13:1143–1155 show indication of size homoplasy and of a relationship between Nybom H, Bartish IV (2000) Effects of life history traits and sampling degree of homoplasy and fragment size. Mol Ecol 11:139–151 strategies on genetic diversity estimates obtained with RAPD Vergeer P, Rengelink R, Copal A, Ouborg NJ (2003) The interacting markers in plants. Perspect Plant Ecol 3:93–114 effects of genetic variation, habitat quality and population size Oostermeijer JGB, van Eijck MW, den Nijs JCM (1994) Offspring on performance of Succisa pratensis. J Ecol 91:18–26 fitness in relation to population size and genetic variation in the Vos P, Hogers R, Bleeker M, Reijans M, van de Lee T, Hornes M, Frijters J, rare perennial plant species Gentiana pneumonanthe (Gentian- Pot J, Peleman J, Kuiper M, Zabeau M (1995) AFLP: a new technique aceae). Oecologia 97:289–296 for DNA fingerprinting. Nucleic Acids Res 23:4407–4414 Ouborg NJ, Vergeer P, Mix C (2006) The rough edges of the conservation genetics paradigm for plants. J Ecol 94:1233–1248

123

IV Lanno, K., Sammul, M. 2014. The survival of transplants of rare Ligularia sibirica is enhanced by neighbouring plants. Folia Geobotanica 49: 163-173. Folia Geobot (2014) 49:163–173 DOI 10.1007/s12224-013-9163-3

The Survival of Transplants of Rare Ligularia sibirica is Enhanced by Neighbouring Plants

Kaire Lanno & Marek Sammul

# Institute of Botany, Academy of Sciences of the Czech Republic 2013

Abstract Reintroduction programs are often initiated to restore the viability of endangered plant populations, whose decline is usually caused by loss of suitable habitats. Ligularia sibirica is a species associated with wetlands. It is endangered in Europe and has declined considerably in Estonia since the cessation of traditional management and, in addition, drainage leading to the overgrowth of habitats. The purpose of this work was to estimate the extent of the impact of competition from neighbouring plants and habitat change caused by overgrowing on the survival of transplants of L. sibirica and thereby assess the effectiveness of reinforcement. Laboratory-grown transplants were planted back into their original populations in plots arranged in a two × two (vegetation intact or removed × open or overgrown habitat) factorial experimental design, and their survival was followed for three years. The survival differed notably among populations, but the percentage of surviving plants per plot was on average higher in plots with intact vegetation and in open habitats. The latter indicates that overgrowing indeed decreases habitat quality for this species, despite the fact that plants of L. sibirica can often be found in forested habitats. The lower survival in plots where vegetation had been removed can largely be explained by increased damage caused by animals. In intact plots, by contrast, neighbouring plants provide shelter and protection. Our results stress the importance of restoration and preservation of habitat quality for the protection of this rare species, for which reinforcement may be effective.

Keywords Competition . Habitat change . Ligularia sibirica . Rare plants . Reinforcement . Transplant survival

Introduction

Loss of habitats and the decline in their quality are the main threats to plant species populations (e.g. Vitousek et al. 1997; Hobbs and Yates 2003; Aguilar

K. Lanno (*) : M. Sammul Institute of Agricultural and Environmental Sciences, Estonian University of Life Sciences, Kreutzwaldi 5, Tartu 51014, Estonia e-mail: [email protected]

99 164 K. Lanno, M. Sammul et al. 2006; Fischer and Lindenmayer 2007; Brook et al. 2008). Changes in human-influenced habitats are often caused, on the one hand, by intensification of agriculture and, on the other, by cessation of traditional management. These changes affect plant populations in different ways, including altered species interactions (e.g., increased competition), change of environmental conditions, genetic deterioration, etc. The need to start a restoration program is often realized only after populations of a rare species have declined dramatically or have even gone extinct. One method to restore populations of endangered plant species is reinforcement (enhancement, augmentation). This involves the addition of individuals to existing populations with the goal of increasing their viability by means of increasing population size or genetic variability or both (Falk et al. 1996). Besides reinforcement, the other restoration types are reintroduction (transplanting individuals to a place where the species went extinct) and introduction (establishing a population in a site not previously inhabited by the species) (Bottin et al. 2007;Hölzeletal.2012). Restoration can be undertaken in many different ways depending on species, habitat type, research questions and other circumstances by sowing or planting and possibly applying different habitat preparation methods (see e.g. Milton et al. 1999; Drayton and Primack 2000; Bottin et al. 2007;GuerrantandKaye2007; Menges 2008; Godefroid et al. 2011). One such habitat preparation method is the removal of neighbours. It has been shown in several studies that survival of transplants is often greatly reduced by competition from neighbouring vegetation (e.g. van der Wal et al. 2000; Midoko-Iponga et al. 2005). In restoration projects, a combination of different methods is usually used; however, their effects should be separated when conducting monitoring studies in order to understand which factors affect population dynamics and restoration success the most and to gather information for further restorations. Ligularia sibirica (L.) Cass. (siberian groundsel) is an endangered species in Europe (included in Annexes II and IV of the EU Habitats Directive) growing mainly in alluvial and paludified grasslands. It has always been rare in Estonia like in most other European countries. During the past 40 years, however, its distribution in Estonia has decreased considerably, both in number and size of populations (Kukk 2003). The main reason for this decline are changes in habitat quality, which are primarily caused by cessation of mowing and grazing in former semi-natural grasslands and also by drainage. This has lead to shrub and tree encroachment, bringing about changed light and soil conditions and also increased competition with other species. Given that L. sibirica also often grows under a pretty dense shrub and tree canopy (personal observation), it is not clear whether expensive and disruptive shrub and tree removal is necessary to sustain populations of this species. The main aim of this work is to evaluate the effect of habitat change on the reinforcement of populations of L. sibirica. We concentrate on the impact of two factors on the survival of transplants: (1) reduced light availability caused by overgrowing and (2) competition from neighbouring plants. Our results also allow us to evaluate the effectiveness of population reinforcement of L. sibirica and to suggest methods for this.

100 The Survival of Transplants of Rare Ligularia sibirica 165

Material and Methods

Species

Ligularia sibirica is a 100–130 cm tall perennial plant from the family Asteraceae. It has erect stems with numerous yellow flower heads arranged in clusters, long-stalked kidney-shaped lower leaves and smaller, stalkless upper leaves. The species repro- duces both sexually (seeds are dispersed mainly by wind) and vegetatively (forming short, thick rhizomes). Ligularia sibirica grows in open habitats and semi-open spots near trees and shrubs across a range of different wetland habitats including alluvial grasslands, paludified grasslands and mires (Kukk 2003). Its main distribution area is the European part of Russia and the Siberian taiga zone (Hultén and Fries 1986). In Europe, it occurs in Estonia, Latvia, Poland, Ukraine, Hungary, the Czech Republic, the Slovak Republic, Bulgaria, Romania, Croatia, Austria and France, being very rare in those countries (Hultén and Fries 1986; Hegi 1987). In Estonia, the species is at the north-western border of its continuous distribution area and has been legally protected since 1936.

Study Sites

The study was conducted at four population sites: Sootaga, Väägvere, Tagula and Õisu (Fig. 1). These sites were selected on the basis that the populations were large enough to provide sufficient amounts of seeds for growing transplants and that they offered both open and already overgrown parts of the same habitat alongside each other. Open habitats had few and occasional trees and shrubs and were representative

Fig. 1 Locations of study sites (S – Sootaga, V – Väägvere, T – Tagula, Õ – Õisu)

101 166 K. Lanno, M. Sammul of good and typical conditions of the habitats studied. The overgrown areas (usually also harbouring infrequent individuals of L. sibirica) were situated immediately adjacent to the open areas and, with the exception of the presence of trees and shrubs, were selected to be as similar to the still open habitats with respect to site conditions as possible. These areas originally belonged to the same habitat type as the remaining open habitats with L. sibirica. Special attention was paid to the similarity in moisture conditions. All populations were situated on the flat landscape without slopes. The characteristics of these populations are given in Table 1. In 2007, soil samples were taken from the top 20 cm of soil at every study site randomly from four spots and then mixed, so approximately 400 g of soil was taken from each population. Samples were air-dried and then analyzed for their pH and content of the plant available P, K, N, Mg and organic C in the laboratory of the Department of Soil Science and Agrochemistry in Estonian University of Life Sciences. For estimating canopy openness, hemispherical photographs were taken in August 2009 using a fish-eye lens and analyzed using WinSCANOPY software (Regent Instruments). Photographs were taken from three randomly chosen places in each of the open and overgrown habitats. Leaf area index in the understorey vegetation was also measured in August 2009 with PAR/LAI ceptometer AccuPAR model LP-80 (Decagon Devices, Inc.). In every population, measurements were made from three randomly chosen places in each of the open and overgrown habitats. In each of those places, each reading was the average of five measurements spread evenly across a1m2 area.

Study Design

In autumn 2007, ripe seeds of L. sibirica were collected from 20 different plants in four populations and stored in paper bags. The following February, seeds of five plants from every population were germinated in Petri dishes on moist filter paper (at fluctuating day/night length of 14/10 hours and temperature of 27/17°C,

Table 1 Geographical coordinates, population size (nr. of tufts), population area, soil characteristics (pH, organic C content indicating level of turf accumulation), habitat type, canopy openness (180° projection area devoid of the shade of tree crowns) and leaf area index (LAI) of understorey vegetation of the four populations. The canopy openness and LAI are given separately for open and overgrown habitats

Population Lat. Long. Pop. Pop. area Soil Soil Org. Habitat Habitat Canopy LAI (° N) (° E) size (ha) pH C% type openness (%)

Sootaga 58.487 26.735 1000 0.6 6.8 27 paludified Open 65.5 2.04 forest Overgr. 16.4 1.26 Väägvere 58.475 26.837 400 0.2 6.3 31.8 floodplain Open 32.6 4.3 grassland Overgr. 23.7 0.61 Õisu 58.186 25.545 600 0.2 5.8 29.6 swamp Open 42.6 2.8 forest Overgr. 16.7 0.41 Tagula 57.878 26.37 1500 0.3 7 33.4 spring fen Open 68.7 1.74 Overgr. 21.7 1.64

102 The Survival of Transplants of Rare Ligularia sibirica 167 respectively), and the resulting seedlings were planted in pots containing commercial potting soil (consisting of peat and fertilizer). In May, young transplants consisting of three leaves were planted back into their original populations. Every population was divided into open and overgrown habitats, and eight plots 0.25 m2 in size were randomly placed within these habitats. To test the competitive effect of neighbouring vegetation, the vegetation was removed randomly from half of the plots in both habitats. Four L. sibirica individuals were planted in every plot. The treatments therefore represented all combinations of two factors, each with two levels: a) vegetation intact or removed (removal treatment); b) open or overgrown habitat. There were four replicate plots per treatment, 16 plots in total, with four plants per plot in every population – except for Õisu, where 12 plots and three transplants per plot were used because of a lack of viable seedlings. The corners of the plots were marked with plastic sticks pushed into the ground to enable their exact detection later. The plant individuals were planted to different corners of the plot. In plots where vegetation was removed, this was done by weeding just before planting: all visible plants were pulled out by hand with as much roots as possible and without additional digging in the soil. In some cases, when there were dense tussocks (of species like Carex cespitosa), which are difficult to pull out, they were removed with the spade with the upper layer of soil and roots. The edges of the plots were cut through with a spade about to the depth of 20 cm to prevent spreading of plant roots from outside. Plots were hand weeded twice in each season, which was sufficient to keep the regrowth of unwanted plants to a minimum. For three years, 2008–2010, the plots were visited twice annually in June and in August, and the survival of each transplant of L. sibirica was recorded.

Data Analysis

We used a repeated measures analysis of variance (STATISTICA version 9.1; StatSoft, Inc. 2010) to test for the effect of main factors removal of neighbouring vegetation, overgrowing and population, and also the effect of all their interactions (with 0.05 significance level) on the survival of plants. The proportion of surviving plants in each 0.25 m2 plot for every monitoring time was calculated and arcsine square root transformed. We used a Tukey’s HSD post-hoc test to evaluate the significance of differences among individual measurements.

Results

The main factors – population, removal, overgrowing and time – all had significant effects on the percentage of surviving plants per plot, as well as had many interactions between those factors (Table 2; Fig. 2). The overall mortality per plot per one treatment ranged from about 20 to 100 percent at the end of the experiment. The difference among populations (P=0.001) resulted from a 25 % higher average survival rate in Tagula compared to the other three populations. Pooling the data from all populations shows there was up to 20 % higher mortality of transplants in plots where vegetation was previously removed compared to intact plots (P=0.005; Table 2). This effect appeared in the second year. In the third year, this

103 168 K. Lanno, M. Sammul

Table 2 The results of the repeated-measures analysis of variance showing the effect of different variables on the proportion of survived plants per plot

Source d.f. FP

Intercept 1 621.6 <0.0001 Population 3 6.6 0.001 Overgrowing 1 8.9 0.005 Removal 1 8.6 0.005 Population × Overgrowing 3 6.9 0.001 Population × Removal 3 3.9 0.015 Overgrowing × Removal 1 4.5 0.039 Population × Overgrowing × Removal 3 0.2 0.882 Error 44 Time 5 134.7 <0.0001 Time × Population 15 8.4 <0.0001 Time × Overgrowing 5 1.5 0.199 Time × Removal 5 9.3 <0.0001 Time × Population × Overgrowing 15 2.9 0.0004 Time × Population × Removal 15 4.0 <0.0001 Time × Overgrowing × Removal 5 2.8 0.017 Time × Population × Overgrowing × Removal 15 0.8 0.665 Error 220 difference decreased slightly, being only marginally significant on the final monitoring occasion. The effect of the interaction between time, population and removal on transplant survival was significant (P<0.0001), and the survival of transplants was higher in intact plots at most monitoring times in all populations except Väägvere (although according to the post-hoc test, this difference was not significant). Survival of transplants was about 20 % higher in open compared to overgrown habitats (P=0.005). This effect was consistent throughout the course of the experiment (no effect of factor ‘time’)(P=0.199). The interaction between time, population and overgrowing was statistically significant (P=0.0004). Although survival in open and overgrown habitats was separated in each population, the higher survival in open habitats appeared only in Väägvere at the final monitoring occasion. Compared to overgrown habitats, it was more clear in open habitats that plant survival was higher in intact plots than in plots where neighbours were removed (Fig. 2). In overgrown habitats at the end of the third year, the survival of transplants was similarly low in all plots irrespective of removal treatment (except in Tagula, where the survival was higher than in other populations both in intact and removal plots). Some plants started to flower or reproduce clonally during the experiment. In Väägvere in the second year, about one-third of the plants initially planted in the open habitat removal plots had reproduced vegetatively, and two of them were flowering. In Tagula, a quarter of the plants in overgrown habitat removal plots had reproduced clonally. In the third year there were seven flowering plants in Väägvere and two in Tagula, in same treatments as in the previous year. At the other two study sites, plants

104 The Survival of Transplants of Rare Ligularia sibirica 169

Fig. 2 Dynamics of survival of transplants of Ligularia sibirica in four populations in Estonia (Sootaga, Väägvere, Tagula, Õisu) in intact plots (squares) and plots with vegetation removed (rhombs) in open (a) and overgrown habitats (b). In x-axes is time (1 – 2008 Jun, 2 – 2008 Aug, 3 – 2009 Jun, 4 – 2009 Aug, 5 – 2010 Jun, 6 – 2010 Aug). The vertical bars denote 0.95 confidence intervals

did not flower; during the third year, only three plants (considering all treatments) had clonal offspring.

Discussion

When designing this study, we postulated the hypothesis that reinforcement of populations of Ligularia sibirica is enhanced by removal of neighbouring vegetation and that reinforcement is less effective in areas that have overgrown with trees and shrubs. There was no certainty about the latter because we had previously observed Ligularia growing also in forested areas around several open habitats where Ligularia thrived. In other cases, we had seen this species suffer severely under a forest canopy. We could thus also expect a certain degree of site-specificity in our results. The study showed that while there is some site-specificity, the general tendencies regarding the main hypothesis are strong. Indeed, under the forest canopy, Ligularia suffers, but contrary to expectations, neighbouring herbaceous vegetation affects the growth of transplants positively. Despite the obvious competition from neighbouring vegetation in most populations, the survival of transplants was lower in plots where neighbouring plants were removed. This can be largely related to damage done by wild animals, primarily wild boar and roe deer, whose activity leaves frequent signs. Weeded plots (and possibly

105 170 K. Lanno, M. Sammul also plot markings) attract animal attention. In plots with intact vegetation, neighbouring plants provide shelter and protection to transplants. Similar damage to transplants from animals has also been found in other studies, for example in McKendrick (1995), Jusaitis (2005), Sammul et al. (2006). Provision such shelter in these studies has shown to be especially important during the first months after planting. Additionally, in wetter habitats there is increased erosion in weeded plots, which may reduce the viability of transplants. Only in one population (Väägvere), where the influence of animals as well as soil erosion by water was small, the survival was slightly higher and also first plants started to flower in plots with neighbouring plants removed (in this population, the vegetation was also higher and shaded more than in other sites). So, although the role of competition between Ligularia sibirica and other species is certainly not minor, the positive effect of neighbouring vegetation (offering shelter and preventing soil erosion) can outweigh the negative effect deriving from competition. As also found for other species (e.g. Harper 1977; Grace and Tilman 1990), the most vulnerable to competition in Ligularia are seedlings, in which mortality is very high (pers. observation). It has been shown elsewhere that the effect of reducing competition from neighbouring vegetation may vary by species and site (e.g. Kaye and Brandt 2005). The importance of such facilitative effects and the dependence of balance between facilitation and competition on environmental conditions have been discussed in different reviews (e.g., Bertness and Callaway 1994; Callaway and Walker 1997; Brooker et al. 2008). With respect to restoration, transplant survival and facilitation, there have been more examples of direct facilitative influences (like the effect of nurse plants; e.g. Cavieres et al. 2002). In our case, however, the survival benefited from diffuse facilitation (general facilitative effects of neighbours) and provision of shelter. Since at the planning phase of the experiment, there was no indication that wild animals may target our experimental plots, so we could not anticipate the importance of exclosures. We therefore cannot evaluate survival exactly as a balance between facilitation and competition, that is, the rate of mortality for each population without the impact of animals. Presumably, the survival rate would have been bigger in weeded plots without animals, although it would probably not markedly exceed the survival in intact plots because of the strong effects of certain other negative factors in weeded plots such as increased soil erosion. Even though L. sibirica often grows under a relatively dense forest and shrub cover (Kukk 2003) and also flowers and manages to produce viable seeds in such conditions (personal observation), the observed survival of transplants in overgrown habitats was lower than in open habitats in most populations under study (except in Tagula). Furthermore, in overgrown habitats, the survival of transplants was usually low by the end of the third year in both intact and vegetation removal plots, so the presence or absence of neighbours did not make much difference under the tree cover. The results stress the importance of preserving or restoring habitat quality for protection of this species. Besides reduced light availability in overgrown habitats, other environmental conditions may also become unsuitable. The effect of altered soil moisture conditions may be especially important, for example. Overgrowing appears to primarily affect the survival of seedlings and juvenile plants, which also explains why young new plants are rarely found in overgrown habitats (Kukk 2003; pers. observation). We must, however, also emphasize that there are exceptional habitats such as Tagula, where plants can spread into the forest quite successfully. The age of

106 The Survival of Transplants of Rare Ligularia sibirica 171 the forest is presumably an important factor in this process, along with soil moisture. The spring fen in Tagula is surrounded by a mature forest where there are gaps in the canopy, which provide more light, and the spring fen itself provides moisture. Moreover, in Tagula there is a very strong and viable population (with both the highest number of tussocks and genetic diversity; Ilves et al. 2013) acting as a good source of many immigrant plants for the neighbouring forest – something that depleted populations do not possess. Thus, in further studies more emphasis should be put on evaluating moisture conditions (including seasonal variation) but also on viability of source populations, which is linked to genetic diversity within populations (as shown in Ilves et al. 2013). The distance among populations is often also an important factor to consider when determining possible exchange of genetic material among populations and hence viability; however, according to an analysis of genetic variability, the populations studied (even the closer ones) are quite isolated, as there was no correlation between genetic and geographic distances (Ilves et al. 2013). The populations differed with respect to plant survival, but plants tended to die more often between vegetation periods (during winter), and although the mortality continued during all three years, the largest mortality appeared during the first year (like, e.g., in Jusaitis 2005). Although the given time-scale in this study offers useful information about the influence of considered factors with regard to initial reinforcement success and methodological knowledge for use in further restoration projects, three years is of course quite a short time to draw firm conclusions from. It has been reported that short-term outcomes can sometimes be misleading. For instance, the survival rate of introduced plants may be high in the first years only to drop afterwards (e.g., Kaye 2002). Especially for perennial species with a long life-span, such as L. sibirica,longer monitoring of surviving plants would be informative. There are different criteria for measuring success of restoration (Griffith et al. 1989;Pavlik1996;Menges2008). One of them is the start of flowering and sexual reproduction. In our experiment, a few transplants already started to flower in the second year of the experiment in one population, yet after three years of the study, flowering was rare. We have seen clonal propagation of transplants, but for a population to spread and considerably grow in number, plants should produce seeds. Thus, we would argue that the monitoring of the success of the transplantations should last at least until the viability of seeds and emergence of seedlings can be evaluated. As our results indicate, population reinforcement can be used in conservation of L. sibirica if necessary, but its effectiveness is limited. The addition of new individuals to populations can be effective only if habitat quality is preserved or restored. In this case, supplementation of new individuals can be used to overcome the high mortality of seedlings and juvenile plants, which is the most critical stage of life for L. sibirica. On the other hand, additional planting site preparation through weeding to avoid competition is not advisable, since it reduces the survival of transplants, mainly as a result of herbivore damage.

Acknowledgments We thank all the people who helped with the lab- and field work (especially Thea Kull, Karin Kaljund, and Silja Kana) and Roger Evans for doing a language revision. Two anonymous reviewers provided useful comments on this manuscript. We appreciate the help from the herbarium of vascular plants and mosses (TAA) of Estonian University of Life Sciences. This work was funded by a species protection programme of the Estonian Environmental Board and by grant No. 8745 from the Estonian Science Foundation.

107 172 K. Lanno, M. Sammul

References

Aguilar R, Ashworth L, Galetto L, Aizen MA (2006) Plant reproductive susceptibility to habitat fragmentation: review and synthesis through a meta-analysis. Ecol Lett 9:968–980 Bertness M, Callaway RM (1994) Positive interactions in communities. Trends Ecol Evol 9:191–193 Bottin L, Le Cadre S, Quilichini A, Bardin P, Moret J, Machon N (2007) Re-establishment trials in endangered plants: a review and the example of Arenaria grandiflora, a species on the brink of the extinction in the Parisian region (France). Ecoscience 14:410–419 Brook BW, Sodhi NS, Bradshaw CJA (2008) Synergies among extinction drivers under global change. Trends Ecol Evol 23:453–460 Brooker RW, Maestre FT, Callaway RM, Lortie CL, Cavieres LA, Kunstler G, Liancourt P, Tielbörger K, Travis JMJ, Anthelme F, Armas C, Coll L, Corcket E, Delzon S, Forey E, Kikvidze Z, Olofsson J, Pugnaire F, Quiroz CL, Saccone P, Schiffers K, Seifan M, Touzard B, Michalet R (2008) Facilitation in plant communities: the past, the present, and the future. J Ecol 96:18–34 Callaway RM, Walker LR (1997) Competition and facilitation: a synthetic approach to interactions in plant communities. Ecology 78:1958–1965 Cavieres L, Arroyo MTK, Penaloza A, Molina-Montenegro M, Torres C (2002) Nurse effect of Bolax gummifera cushion plants in the alpine vegetation of the Chilean Patagonian Andes. JVegSci13:547–554 Drayton B, Primack RB (2000) Rates of success in the reintroduction by four methods of several perennial plant species in eastern Massachusetts. Rhodora 102:299–331 Falk DA, Millar CI, Olwell M (1996) Restoring diversity. In Falk DA, Millar CI, Olwell M (eds) Strategies for reintroduction of endangered plants. Island Press, Washington DC, pp XIII–XXII Fischer J, Lindenmayer DB (2007) Landscape modification and habitat fragmentation: a synthesis. Global Ecol Biogeogr 16:265–280 Godefroid S, Piazza C, Rossi G, Buord S, Stevens A-D, Aguraiuja R, Cowell C, Weekley CW, Vogg G, Iriondo JM, Johnson I, Dixon B, Gordon D, Magnanon S, Valentin B, Bjureke K, Koopman R, Vicens M, Virevaire M, Vanderborght T (2011) How successful are plant species reintroductions? Biol Conservation 144:672–682 Grace JB, Tilman D (1990) Perspectives on plant competition. Academic Press, San Diego Griffith B, Scott JM, Carpenter JW, Reed C (1989) Translocation as a species conservation tool: status and strategy. Science 245:477–480 Guerrant EO, Kaye TN (2007) Reintroduction of rare and endangered plants: common factors, questions and approaches. Austral J Bot 55:362–370 Harper J (1977) The population biology of plants. Academic Press, London Hegi G (1987) Illustrierte Flora von Mitteleuropa. Teil 4. Verlag Paul Parey, Berlin und Hamburg Hobbs RJ, Yates CJ (2003) Impacts of ecosystem fragmentation on plant populations: generalising the idiosyncratic. Austral J Bot 51:471–488 Hölzel N, Buisson E, Dutoit T (2012) Species introduction – a major topic in vegetation restoration. Appl Veg Sci 15:161–165 Hultén E, Fries M (1986) Atlas of North European vascular plants: North of the Tropic of Cancer. Koeltz Scientific Books, Köningstein Ilves A, Lanno K, Sammul M, Tali, K (2013) Genetic variability, population size and reproduction potential in Ligularia sibirica (L.) populations in Estonia. Conservation Genet 14:661–669 Jusaitis M (2005) Translocation trials confirm specific factors affecting the establishment of three endangered plant species. Ecol Managem Restoration 60:61–67 Kaye TN (2002) Reintroducing the endangered Pink Sand-verbena to Pacific Coast beaches: direct seeding and out-planting. In Brooks MB, Carothers SK, LaBanca T (eds) The ecology and management of rare plants of northwestern California. Proceedings from Symposium North Coast Chapter California Native Plant Society, California Native Plant Society, Sacramento, pp 131–140 Kaye TN, Brandt A (2005) Seeding and transplanting rare Willamette Valley prairie plants for population restoration. Technical report prepared for the Eugene District, USDI Bureau of Land Management and Institute of Applied Ecology, Corvalis, Oregon Kukk Ü (2003) The distribution of Ligularia sibirica (L.) Cass. in Estonia and changes in its population. Biul Ogrodów Bot 12:11–22 McKendrick SL (1995) The effects of herbivory and vegetation on laboratory-raised Dactylorhiza praetermissa (Orchidaceae) planted into grassland in southern England. Biol Conservation 73:215–220 Menges ES (2008) Restoration demography and genetics of plants: when is a translocation successful? Austral J Bot 56:187–196

108 The Survival of Transplants of Rare Ligularia sibirica 173

Midoko-Iponga D, Krug C, Milton SJ (2005) Competition and herbivory influence growth and survival of shrubs on old fields: implications for restoration of renosterveld shrubland. J Veg Sci 16:685–692 Milton SJ, Bond WJ, du Plessis MA, Gibbs D, Hilton-Taylor C, Linder HP, Raitt L, Wood J, Donaldson JS (1999) A protocol for plant conservation by translocation in threatened lowland Fynbos. Conservation Biol 13:735–743 Pavlik BM (1996) Defining and measuring success. In Falk DA, Millar CI, Olwell M (eds) Strategies for reintroduction of endangered plants. Island Press, Washington DC, pp 127–155 Sammul M, Oksanen L, Mägi M (2006) Regional effects on competition-productivity relationship: a set of field experiments in two distant regions. Oikos 112:138–148 StatSoft, Inc. (2010) STATISTICA (data analysis software system), version 9.1. Available at: http:// www.statsoft.com Van der Wal R, Egas M, van der Veen A, Bakker J (2000) Effects of resource competition and herbivory on plant performance along a natural productivity gradient. J Ecol 88:317–330 Vitousek PM, Mooney HA, Lubchenko J, Melillo JM (1997) Human domination of Earth’s ecosystems. Science 277:494–499

Received: 28 March 2012 /Revised: 5 February 2013 /Accepted: 7 February 2013 / Published online: 19 June 2013

109 CURRICULUM VITAE

First name: Kaire Last name: Lanno Date of Birth: 01.10.1978 Institution: Estonian University of Life Sciences, Institute of Agricultural and Environmental Sciences, Department of Botany, Kreutzwaldi 5, Tartu 51014, Estonia E-mail: [email protected]

Education: 2008-2014 PhD studies in Ecology, Estonian University of Life Sciences 2002-2006 MSc studies in Plant Ecology and Ecophysiology, University of Tartu 2004-2005 one-year course, Biology and Health Education Teacher of Secondary School, University of Tartu 1997-2001 BSc studies in Biology, University of Tartu 1994-1997 Pärnu Raeküla Gymnasium 1985-1994 Surju Primary School

Languages spoken: Estonian, English, Russian

Professional employment: Since 2013 Estonian University of Life Sciences, Institute of Agricultural and Environmental Sciences, Department of Botany, researcher 2006-2013 Estonian University of Life Sciences, Institute of Agricultural and Environmental Sciences, Department of Botany, specialist

Membership in societies: Since 2014 Member of Estonian Naturalist`s Society Since 2001 Member of Estonian Seminatural Communities Conservation Association

110 Research interest: Ecology, population biology, conservation biology, botany

Professional training: 2011 graduate course “7th International School of Conservation Biology”, Rovinj, Croatia

111 ELULOOKIRJELDUS

Eesnimi: Kaire Perekonnanimi: Lanno Sünniaeg: 01.10.1978 Töökoht: Eesti Maaülikool, Põllumajandus- ja keskkonnainstituut, botaanika osakond, Kreutzwaldi 5, Tartu 51014 E-post: [email protected]

Haridus: 2008-2014 doktoriõpe ökoloogia erialal, Eesti Maaülikool 2002-2006 magistriõpe taimeökoloogia ja ökofüsioloogia erialal, Tartu Ülikool 2004-2005 õpetajakoolituse kutseaasta, põhikooli ja gümnaasiumi bioloogia ning terviseõpetuse õpetaja, Tartu Ülikool 1997-2001 bakalaureuseõpe bioloogia erialal, Tartu Ülikool 1994-1997 Pärnu Raeküla Gümnaasium 1985-1994 Surju Põhikool

Keelteoskus: eesti, inglise, vene

Teenistuskäik: Alates 2013 Eesti Maaülikool, Põllumajandus- ja keskkonnainstituut, teadur 2006-2013 Eest i Maaü l ikool, Põl lumajandus- ja keskkonnainstituut, spetsialist

Teadusorganisatsiooniline tegevus: Alates 2014 Eesti Looduseuurijate Seltsi liige Alates 2001 Pärandkoosluste Kaitse Ühingu liige

Teadustöö põhisuunad: Ökoloogia, populatsioonibioloogia, looduskaitsebioloogia, botaanika

112 Erialane enesetäiendamine: 2011 kraadiõppurite kursus “7th International School of Conservation Biology”, Rovinj, Horvaatia

113 LIST OF PUBLICATIONS

Publications indexed in the ISI web of Science database:

Lanno, K., Sammul, M. 2014. The survival of transplants of rare Ligularia sibirica is enhanced by neighbouring plants. Folia Geobotanica 49: 163-173. Ilves, A., Lanno, K., Sammul, M., Tali, K. 2013. Genetic variability, population size and reproduction potential in Ligularia sibirica (L.) populations in Estonia. Conservation Genetics 14: 661 – 669. Kull, T., Sammul, M., Kull, K., Lanno, K., Tali, K., Gruber, B., Schmeller, D., Henle, K. 2008. Necessity and reality of monitoring threatened European vascular plants. Biodiversity and Conservation 17: 3383 – 3402. Sammul, M., Kull, T., Lanno, K., Otsus, M., Mägi, M., Kana, S. 2008. Habitat preferences and distribution characteristics are indicative of species long-term persistence in the Estonian ﬂ ora. Biodiversity and Conservation, 17; 3531 – 3550. Schmeller, D.S., Bauch, B., Gruber, B., Juškaitis, R., Budrys, E., Babij, V., Lanno, K., Sammul, M., Varga, Z., Henle, K. 2008. Determination of conservation priorities in regions with multiple political jurisdictions. Biodiversity and Conservation, 17: 3623 – 3630. Schmeller, D.S., Gruber, B., Bauch, B., Lanno, K., Budrys, E., Babij, V., Juškaitis, R., Sammul, M., Varga, Z., Henle, K. 2008. Determination of national conservation responsibilities for species conservation in regions with multiple political jurisdictions. Biodiversity and Conservation, 17: 3607 – 3622. Sammul, M., Kattai, K., Lanno, K., Meltsov, V., Otsus, M., Nõuakas, L., Kukk, D., Mesipuu, M., Kana, S., Kukk, T. 2008. Wooded meadows of Estonia: conservation efforts for a traditional habitat. Agricultural and Food Science 17: 413 – 429.

114 Conference abstracts:

Lanno, K., Sammul, M. 2012. The need for more sufﬁ cient restoration effort: slow vegetation change in restored coastal meadows. The 8th European Conference on Ecological restoration, 9-14 September, České Budějovice, Czech Republic.

Lanno, K., Sammul, M. 2010. The survival of transplants of rare Ligularia sibirica is enhanced by neighbouring plants. Plant Population Biology: Crossing Borders, 13-15 May, Nijmegen, Netherlands.

Kattai, K., Lanno, K., Sammul, M., Otsus, M. 2009. The importance of functional traits of plants during the restoration of a wooded meadow. 52nd International Symposium of the International Association for Vegetation Science, 30 May – 4 June, Chania, Crete.

Sammul, M., Kattai, K., Lanno, K. 2006. Dynamics of clonal growth types of meadow plants in long-term fertilization experiment. 8th Clonal Plant Workshop, 27-30 June, Pärnu, Estonia.

115

VIIS VIIMAST KAITSMIST

ANTS TAMMEPUU EMERGENCY RISK ASSESSMENT IN ESTONIA HÄDAOLUKORRA RISKIANALÜÜS EESTIS Professor Kalev Sepp 26. mai 2014

ANTS-HANNES VIIRA STRUCTURAL ADJUSTMENT OF ESTONIAN AGRICULTURE – THE ROLE OF INSTITUTIONAL CHANGES AND SOCIOECONOMIC FACTORS OF FARM GROWTH, DECLINE AND EXIT EESTI PÕLLUMAJANDUSE STRUKTURAALNE KOHANEMINE – INSTITUTSIONAALSETE MUUTUSTE JA PÕLLUMAJANDUSETTEVÕTETE KASVU, KAHANEMIST NING TEGEVUSE LÕPETAMIST MÕJUTAVATE SOTSIAALMAJANDUSLIKE TEGURITE OSA Professor Rando Värnik, dotsent Reet Põldaru 5. juuni 2014

MONIKA SUŠKEVIČS KNOWLEDGE, LEARNING AND LEGITIMACY IN MULTI-LEVEL PARTICIPATORY ECOLOGICAL NETWORK GOVERNANCE KAASAMISPROTSESSIDE TÕHUSUS JA TULEMUSLIKKUS ÖKOLOOGILISE VÕRGUSTIKU PLANEERIMISEL NING RAKENDAMISEL MITMETASANDILISES VALITSEMISMUDELIS Professor Mart Külvik, vanemteadur Aivar Leito 16. juuni 2014

MARKO KASS EFFECT OF CRUDE GLYCEROL FEEDING ON FEED INTAKE, LACTATIONAL PERFORMANCE AND METABOLIC STATUS OF DAIRY COWS TOORGLÜTSEROOLI SÖÖTMISE MÕJU LÜPSILEHMADE SÖÖMUSELE, PIIMAJÕUDLUSELE JA AINEVAHETUSLIKULE SEISUNDILE Prof. Emer. Olav Kärt, Dr. David Richard Arney 26. juuni 2014

KATRIT KARUS EFFECTS OF MACROPHYTES, FISH AND METAZOOPLANKTON ON A MICROBIAL FOOD WEB VEETAIMESTIKU, KALADE JA METAZOOPLANKTONI MÕJU MIKROOBSELE TOIDUAHELALE Vanemteadur Priit Zingel 28. august 2014

ISSN 2382-7076 ISBN 978-9949-536-44-3 (trükis) ISBN 978-9949-536-45-0 (pdf)