Merkflo 4255 maxsimmelink bromus racemosus by Merkske

Habitat preference and phytosociological position of

Bromus racemosus L. in the Netherlands and surrounding countries

MSc thesis Max Simmelink July 2014

Habitat preference and phytosociological position of Bromus racemosus L. in the Netherlands and surrounding countries MSc thesis Nature Conservation and Plant Ecology (NCP-80436)

M.R. (Max) Simmelink Master Forest and Nature Conservation, Wageningen University Registration number: 900808762100

July 2014

Supervisors: Prof. Dr. J.H.J. (Joop) SchaminĂŠe, Nature Conservation and Plant Ecology Group & Alterra, Wageningen UR Dr. J.A.M. (John) Janssen, Alterra, Wageningen UR

The MSc report may not be copied in whole or in parts without the written permission of the author and the chair group.

Acknowledgements I have enjoyed working on this research. It gave me the chance to develop me research and field work skills, brought me in contact with inspiring and helpful experts and gave me access to beautiful grasslands. The idea for this research came from Joop Schaminée, which I really appreciate. Before the research I had neither seen Bromus racemosus, nor visited any of the study sites. However I already knew most common grassland plant species, I had visited other interesting grasslands and I was fascinated by grasses. Therefore I was motivated to study a grass species and its habitat. It was convenient that the flowering of B. racemosus period ideally matched my study planning, and the study sites were accessible by train and bike from Wageningen (still I camped often for a few days near the study sites). During my research I became even more convinced that it is a fascinating species, and that the study would have interesting results. I am grateful for all advice and help from my supervisors Joop Schaminée and John Janssen and I really value the good contact with them. Additionally I got advice and help from several other experts. Rein de Waal helped me very much with the humus and soil description. Eddy Weeda and Dick Kerkhof gave many valuable tips and ideas about study sites, the ecology of B. racemosus and existing datasets. The interviews with Albert Corporaal and Piet Schippers gave also many valuable insights. It was very nice and instructive that many enthusiastic experts joined my field work for one or more days (all persons mentioned above, Ronald Buskens, Theo Muusse and an excursion of the Plantensociologische Kring Nederland). I am very thankful for the permissions to visit the grassland reserves of Staatsbosbeheer, Zuid-Hollands Landschap, Utrechts Landschap, Geldersch Landschap and Natuurmonumenten. The contact with all involved employees of these organisations was positive. Especially with René Garskamp, who manages probably the largest total area of grasslands with B. racemosus in the Netherlands. He explained and showed me much about the management of these areas in the Vijfheerenlanden. Further I got valuable data from Maud Raman (relevés from Belgium), Adrie van Heerden (PQ’s of South Holland), Rob Geerts (the CLM dataset), André Aptroot (records in areas of Natuurmonumenten) and Gert Rosenthal, who sent me the valuable work of Susanne Lutz (Lutz 1996). Jan van Walsem and his colleagues of the soil lab helped me a lot with the chemical soil analysis. Finally I appreciate that Leni Duistermaat helped me to study the differences between both subspecies in the Nationaal Herbarium Nederland.

Table of contents Summary .................................................................................................................................... 8 1. Introduction ............................................................................................................................ 9 1.1 The species Bromus racemosus........................................................................................ 9 1.2 Ecology........................................................................................................................... 11 1.3 Sociology........................................................................................................................ 12 1.4 Distribution..................................................................................................................... 12 1.5 Red list status ................................................................................................................. 13 1.6 Research aims................................................................................................................. 14 1.7 Nomenclature ................................................................................................................. 15 1.8 Reading guide................................................................................................................. 15 2. Study area ............................................................................................................................. 16 2.1 Geomorphology and geology ......................................................................................... 16 2.2 Nutrient availability........................................................................................................ 18 2.3 Climate ........................................................................................................................... 18 2.4 Sites in surrounding countries ........................................................................................ 18 3. Material and methods ........................................................................................................... 19 3.1 Fieldwork ....................................................................................................................... 19 3.1.1 Vegetation description............................................................................................. 19 3.1.2 Soil and humus profile description.......................................................................... 19 3.1.3 Soil sampling and analysis ...................................................................................... 20 3.2 Other datasets ................................................................................................................. 20 3.3 Data analysis .................................................................................................................. 20 3.4 Additional environmental data ....................................................................................... 22 4. Results .................................................................................................................................. 23 4.1 Abiotic conditions .......................................................................................................... 23 4.1.1 Hydrology................................................................................................................ 23 4.1.2 Nutrient availability................................................................................................. 25 4.1.3 Other soil characteristics ......................................................................................... 28 4.2 Grassland management .................................................................................................. 28 4.3 Vegetation types in the Netherlands............................................................................... 30 4.3.1 Fieldwork ................................................................................................................ 30 4.3.2 Dutch National Vegetation Database ...................................................................... 41 4.4 Vegetation types in surrounding countries..................................................................... 41 5 Discussion ............................................................................................................................. 47 5.1 Abiotic conditions .......................................................................................................... 47 5.1.1 Hydrology................................................................................................................ 47 5.1.2 Nutrient availability................................................................................................. 48 5.1.3 Other factors ............................................................................................................ 49 5.1.4 Differences between subspecies.............................................................................. 49 5.2 Grassland management .................................................................................................. 50 5.3 Sociology........................................................................................................................ 51 5.4 Historical ecology .......................................................................................................... 53 5.5 Comparison with other grassland annuals...................................................................... 55 5.6 Nature conservation........................................................................................................ 55 5.7 Ideas for future research ................................................................................................. 57 6. Conclusion............................................................................................................................ 58

Verkorte Nederlandse versie .................................................................................................... 59 Inleiding ............................................................................................................................... 59 Onderzoeksgebied ................................................................................................................ 61 Methoden: vegetatieopnamen en bodemmonsters ............................................................... 62 Hydrologie............................................................................................................................ 62 Bodem .................................................................................................................................. 63 Graslandbeheer..................................................................................................................... 64 Vegetatietypen...................................................................................................................... 65 Conclusie.............................................................................................................................. 69 Erweiterte deutsche Zusammenfassung ................................................................................... 70 References ................................................................................................................................ 73 Appendices ............................................................................................................................... 81 Appendix 1: RelevĂŠ table, header data and soil profile description..................................... 81 Appendix 2: Study sites........................................................................................................ 81 Appendix 3: Synoptic table .................................................................................................. 84 Appendix 4: Soil profile diagrams and photos..................................................................... 90 Appendix 5: Groundwater table graphs ............................................................................... 98 Appendix 6: Ordination results .......................................................................................... 105 Appendix 7: Vegetation photos.......................................................................................... 111

Summary Expanded summaries in Dutch and German can be found at the end of the report. Bromus racemosus is a rare grass species of moist meadows. It is placed on the Red Lists of several European countries because it has decreased due to intensification of agricultural grassland management during the last decades. Its winter annual life-cycle is remarkable for a species of permanent grasslands. In the Netherlands, two subspecies occur: subsp. racemosus and subsp. commutatus. These are often regarded as different species in other countries. The aim of this study is to increase the knowledge about the habitat preference of Bromus racemosus L. in the Netherlands. The influence of abiotic conditions and management on its abundance and its syntaxonomic position are studied. Vegetation, soil and humus profile and soil chemistry have been studied in 28 sites in the Netherlands. These sites are mainly situated in floodplains and polders in the river landscape and on the Wadden island Texel. For the data analysis, classification with Twinspan, ordination and univariate statistics were used. RelevĂŠs from the Dutch National Vegetation Database and from nearby countries were compared with relevĂŠs from fieldwork.

The results indicate that B. racemosus is characteristic for moderately moist, moderately nutrient rich Molinio-Arrhenatheretea meadows with a good mineralisation. In the Netherlands and nearby regions in France, England, Belgium and Germany, B. racemosus reaches its optimum in moderately nutrient rich basal communities, Alopecurion associations, Calthion vegetations, moist Arrhenatherion sites and moist, mown Cynosurion sites. The absence of a seedbank and a low dispersal capability make the species vulnerable. Seed ripening and seedling establishment should be successful every year to maintain a population. A management of mowing after 15 June and aftermath grazing is most suitable, since it allows seed ripening and maintains a sufficiently open sward, needed for establishment. B. racemosus requires a high groundwater table and/or river flooding during winter for a limited period every one to few years, as this creates open places in the vegetation. Under these conditions the species co-occurs successfully with perennial grasses. Its cover is highest under somewhat drier, moderately nutrient rich conditions. The species can recolonise former agricultural grassland after a few years of mowing without fertilisation, provided seeds can disperse from adjacent parcels or are reintroduced.

1. Introduction 1.1 The species Bromus racemosus Bromus racemosus L. is a rare grass species of moist meadows that is placed on the Red Lists of several European countries (see chapter 1.5). Its winter annual life-cycle is remarkable for a species of permanent grasslands, which are dominated by perennial grasses (WEEDA 1994). The species looks quite similar to the much more common Bromus hordeaceus (fig. 1), and it is therefore often overlooked or misidentified. According to the Dutch Heukels’ Flora, the species can be subdivided into two subspecies (table 1, fig. 2): Bromus racemosus subsp. racemosus and Bromus racemosus subsp. commutatus (VAN DER MEIJDEN 2005). In the Netherlands, the former is much more common than the latter (WEEDA 1994). SPALTON (2002) studied many floras from different countries and found that most authors regard these two taxa as separate species (B. racemosus and B. commutatus). He inspected 281 specimens of B. racemosus subsp. racemosus and 392 of B. racemosus subsp. commutatus from the British isles and made a key to determinate these taxa. He found no proof for hybridisation. SPALTON (2002) writes that “P.M. Smith (pers. comm., 2001) tried very hard to hybridise B. racemosus and B. commutatus and failed to do so, though he considered that this may have been a technical problem.” This P.M. Smith describes in SMITH (1973) that WILSON (1956) “managed to produce completely fertile F 1 and F 2 hybrids of the two taxa, which were morphologically intermediate between the parents”. In this article we regard the two taxa as subspecies, although we have no data or observations that indicate hybridisation. The names in several languages are given in table 1.

Fig. 1: Detail of the leaf sheaths and nodes of Bromus hordeaceus (above) and Bromus racemosus subsp. racemosus (below). The latter has stiffer hairs on the leaf sheaths and further no hairs on the spikelets (subsp. commutatus has sometimes hairy spikelets). The former has softer hairs and always hairy spikelets (photo: Max Simmelink).

Fig. 2: Spikelets of Bromus racemosus subsp. racemosus (above) and subsp. commutatus (below). The latter has on average longer spikelets and lemmaâ&#x20AC;&#x2122;s. Of both subspecies, individuals with only one or a few spikelets can occur (photos: Theo Muusse).

Table 1: The names of the two subspecies in several languages. Synonyms for the scientific name Bromus racemosus are Serrafalcus racemosus L Parl. and Bromus pratensis Ehrh. Scientific name Bromus racemosus L. subsp. Bromus racemosus subsp. subspecies racemosus commutatus (Schrad.) Syme Scientific name Bromus racemosus L . Bromus commutatus Schrad. species English (including Smooth Brome, Bald Brome, Meadow Brome, Hairy Chess, Hairy American names)* (Smooth) Chess, Spiked Brome(Brome, European Brome Grass) German Trauben-Trespe, Traubige Trespe Verwechselte Trespe, Wiesen-Trespe French Brome en grappe(s), Brome Ă Brome variable, Brome confondu grappe(s) Dutch Trosdravik, Velddravik Grote trosdravik, Velddravik, Verwisselde dravik * Smooth Brome is also used for B. inermis and B. ramosus, Hairy Brome for B. ramosus and Smooth Chess for B. secalinus.

1.2 Ecology B. racemosus is one of the few annuals in moist grasslands. In the Netherlands it is most frequent in meadows on river clay, and further in some areas with marine clay, clayey peat, fine silty sand and other stream valley deposits. The species is absent from highly productive agricultural grasslands. B. racemosus flowers in May and early June and sets seeds in June, before the meadows are mown traditionally. It is monocarpic; after ripening of the seeds the plants die. Superficially drying of the soil in summer is therefore tolerated (WEEDA 1994). The seeds are relatively heavy, with a mean weight of 3.75 mg, s.d. 0.90 according to JENSEN (2004) and 1.4-4.8 mg according to LUTZ (1996). They germinate directly after ripening as soon as moist conditions occur (LUTZ 1996). JENSEN (2004) observed germination in late summer. The seeds do not show primary dormancy (JENSEN 2004). Light is not needed for germination. Therefore they also germinate when buried, provided they become moist. This prevents the formation of a seedbank (LUTZ 1996). The seeds have a good germination ability: 86% respectively 93% germinated in experiments described by LUTZ (1996) and JENSEN (2004). Further LUTZ (1996) did not find a lower germination rate or seedling survival rate in smaller populations. Because B. racemosus is a short-lived species, it is likely to responds fast when environmental conditions change. It was observed that the species disappeared from sites where it was abundant one or a few years before, and that it remarkably increased its population on a parcel within a few years (LUTZ 1996). The species is reported to decline in several European countries because of agricultural improvement of meadows, for example through drainage, high levels of fertilisation, a high mowing frequency and ploughing of grasslands (PRESTON et al. 2002, ARNOLDS et al. 1999, WEEDA 1994). Further the abandonment of mowing of wet meadows contributes to the decline (BĂ&#x2013;HLING et al. 1998, ROSENTHAL 2003, LUTZ 1996).

LUTZ (1996) did research about the sociological, ecological and population biological behaviour of the species in the region of Bremen (Northwest Germany). Still there is limited knowledge about which site conditions and management practices are optimal for the species.

1.3 Sociology In literature different syntaxonomic positions are mentioned for B. racemosus, but it is clear that the species is mainly found in the class Molinio-Arrhenatheretea Tüxen 1937. In many parts of Germany B. racemosus is considered a character species of the Calthion palustris Tüxen 1937, an alliance of wet moderately nutrient rich meadows (BÖHLING 1998, PÄZOLT & JANSEN 2004, LUTZ 1996, BUCKART et al. 2004). In the Netherlands it is considered a weak character species of the Alopecurion pratensis Passarge 1964, occurring in basal communities and associations. This alliance is recognised in the Netherlands since the 1990’s (DROK 1992, SCHAMINÉE et al. 1996). Before these vegetations were assigned to the Arrhenatherion elatioris Koch 1926 or the Calthion palustris (SCHAMINÉE et al. 1996). The alliance is classified in the Arrhenatheretalia Tüxen 1931, but there are also arguments to place it in the Molinietalia Koch 1926 (to which the Calthion palustris belongs). Regarding hydrology it is intermediate between the drier Arrhenatheretalia and the wetter Molinietalia. The alliance is found on places that are flooded in winter by a rising groundwater table and/or river water. In spring the groundwater table sinks and in summer superficial drying occurs. The Alopecurion pratensis is mainly differentiated negatively to other alliances, which makes it difficult to classify basal communities. Differential species to the Arrhenatherion elatioris and the Cynosurion cristati Tüxen 1947 are Symphytum officinale, Phalaris arundinacea and Carex acuta (SCHAMINÉE et al. 1996).

1.4 Distribution According to HULTEN & FRIES (1986) the species is native in most parts of Europe. The distribution range of subsp. commutatus reaches further eastward (fig. 3). According to USDA (2013) the species is native to Europe (except northern Scandinavia), Morocco, Tunisia, the Caucasus and temperate parts of Asia. Further the species is considered naturalised in Canada, Mexico, the United States, Australia, Chile and South Africa.

Fig. 3: Distribution of Bromus racemosus subsp. racemosus (left) and subsp. commutatus (right), illustrations from HULTEN & FRIES (1986).

BĂ&#x2013;HLING et al. (1998) hypothesise that subsp. commutatus was introduced into Central Europe in Roman times, due to transport of crops from Southern Europe. Also for subsp. racemosus no records from before the Roman times are known. HULTEN & FRIES (1986) state that both taxa are native in Europe, but spread towards Northern Europe by humans. In GreatBritain, B. racemosus mainly grows in the southern half of England. It can spread through transport of impure grass seeds (PRESTON et al. 2002). Within the Netherlands, the species was present in most low-lying areas and in some brook valleys in higher parts of the country, but it was relatively rare in the north. During the second half of the 20th century, B. racemosus faced a strong decline (>50%) of its distribution in the Netherlands (ODĂ&#x2030; et al. 2006). Nowadays it is still abundant in many nature reserves in the central river area, but elsewhere it is mostly restricted to a few small areas (fig. 4).

Fig. 4. The distribution of B. racemosus in the Netherlands (FLORON 2014). The blue squares indicate the distribution before 1975 (left) respectively 2005 (right) and the red dots afterwards (both in grid cells of 5*5km). This illustrates that the strong decrease of the species continued also after 1975.

1.5 Red list status On the Dutch red list B. racemosus is assessed as vulnerable (LNV 2004), because of its rarity and especially its strong decline (ODĂ&#x2030; et al. 2006). The species is therefore also mentioned on a list of target species, and additionally because of the international importance of the Netherlands for this species (BAL et al. 2001). On the proposal for a new version of the Dutch red list for vascular plants the species is again mentioned as vulnerable (SPARRIUS et al. 2013). In the German red list for plants, B. racemosus subsp. commutatus is assessed not endangered, but B. racemosus subsp. racemosus as category 3 endangered (LUDWIG & SCHNITTLER 1996). However, B. racemosus subsp. racemosus has a wider distribution in Germany than B. racemosus subsp. commutatus (BFN 2013). Germany has a high international responsibility for the conservation of B. racemosus subsp. racemosus, since the country contains a high

proportion of the world population (ca. 10-33%), and is situated in the centre of its distribution range (LUDWIG et al. 2007). The red list for vascular plants of Flanders mentions B. racemosus subsp. commutatus as a rare species, but states that B. racemosus subsp. racemosus is currently not endangered (VAN LANDUYT et al. 2006). However, in another chapter of the same book ZWAENEPOEL (2006) writes that this species is rare in Flanders and has probably faced a very strong decline. In Wallonia, B. racemosus subsp. racemosus is mentioned on the red list as endangered, and B. racemosus subsp. commutatus as vulnerable (OFFH 2013). On the British red list, both species are placed in the category least concern (CHEFFINGS & FARRELL 2005), although both species have declined (PRESTON et al. 2002).

1.6 Research aims The aim of this study is to increase the knowledge about the habitat preference of B. racemosus in the Netherlands. Particularly the relation of abiotic conditions and management with its abundance and its syntaxonomic position have been studied. Further, the impact of landscape and land use history has been considered. This should lead to a better understanding of which factors have contributed to the alarming decrease of the species, and which management measures could contribute to its preservation. Further it can generate knowledge about the ecological and syntaxonomic indicator value of the species. The study is not designed to explore the ecological difference between both subspecies of B. racemosus, since too few locations with subsp. commutatus were known. The Main research question is: Under which site conditions and in which vegetation types does Bromus racemosus occur in the Netherlands? There are five sub-questions: 1) Under which abiotic conditions does B. racemosus occur in the Netherlands? 2) What is the influence of grassland management on the abundance of B. racemosus? 3) In which vegetation types does B. racemosus occur in the Netherlands, and which site conditions are typical for these communities? 4) In which vegetation types does B. racemosus occur in nearby European countries? 5) How could landscape and land use history have affected the occurrence of B. racemosus in the Netherlands? Two answer these questions, fieldwork, data analysis and literature analysis have been done. The research can be characterised as descriptive, correlative and explorative, therefore causal relationships cannot be proven.

1.7 Nomenclature Nomenclature of communities is according to SCHAMINﾃ右 et al. (1996). Nomenclature of vascular plant is according to Flora Europaea (TUTIN et al. 2001); a deviation is the name of B. racemosus, since its subspecies are considered species in Flora Europaea (B. racemosus and B. commutatus). Nomenclature of mosses follows SIEBEL et al. (2006).

1.8 Reading guide In chapter 2 the study area is described, focusing mainly on the geology and geomorphology. This serves as a framework for the sub-questions 1, 3 and 5. Chapter 3 covers the materials, methods and datasets used; besides data from own field work several existing datasets were used. The results are presented in chapter 4, starting with the abiotic conditions (sub-question 1), particularly hydrology and nutrient availability. Next the grassland management (sub-question 2) is addressed. To answer sub-question 3, the vegetation types in which B. racemosus grows are described, with attention to species composition, syntaxonomy, abiotic conditions and management. Subsequently these vegetations are compared to vegetation types in surrounding countries (sub-question 4). Sub-question 5 is not addressed in this chapter; since this is answered mainly based on literature research, it is part of the discussion. In the discussion (chapter 5) all results are interpreted and compared to findings from literature, starting with about the same paragraph structure as in chapter 4. Additionally several new aspects are covered. This includes chapter 5.1.4, focusing on the differences between both subspecies, and the paragraphs 5.4-5.7. Paragraph 5.4 on historical ecology deals with sub-question 5. To place the findings of the research in a wider context, paragraph 5.5 compares B. racemosus with other grassland annuals. Paragraph 5.6 covers the perspectives for the preservation of B. racemosus and the vegetation in which it occurs. The last paragraph presents some ideas for future research. After the conclusion (chapter 6) several appendices follow, containing the raw data, data about the study sites, photos and diagrams. The soil data and diagrams are mainly in Dutch, since the Dutch soil and humus terminology often does not have equivalent English terms.

2. Study area The study considered 28 nature reserves in the Netherlands where B. racemosus had been observed recently (fig. 5 and appendix 2). All sites are grasslands owned by nature conservation organisations and in most sites the management is outsourced to farmers. Most sites are situated in the central part of the Dutch river landscape, in floodplains along the rivers Nederrijn/Lek, Waal (both Rhine branches) and Meuse, and in polder areas near these rivers (former floodplains). A concentration of locations is situated in the polder area Vijfheerenlanden. Further one site is located along the IJssel (another Rhine branch), and three along the Zwarte Water. One site is situated near the border with Belgium, close to the stream Het Merkske. Three sites are in the polders of the Wadden island Texel. Most locations have been grassland since at least the mid of the 19th century, but some parcels have been forest or arable land until the first half of the 20th century (Projectteam WatWasWaar.nl 2013).

Fig. 5: Left: The locations of the studied sites in the Netherlands. See also appendix 2. Right: The locations of relevĂŠs from surrounding countries (including the Dutch relevĂŠs), see chapter 2.4 & 4.4 (locations described in table 5).

2.1 Geomorphology and geology Most sites are close to the sea level, between -0.8 m and 8 m above sea level (with NAP as vertical datum). Het Merkske is with 14.5 m the highest site (HET WATERSCHAPSHUIS 2013). All site have relatively flat terrain. Some floodplains have steep slopes and maximal height differences of 4 m. 16

The soil consists of Holocene deposits, only Het Merkske has Pleistocene deposits covered by a 30 cm thick layer of recently formed peat. The soil is dominated by mineral material deposited by rivers, tidal rivers or the sea. Some sites have peaty soils with maximal 40% organic matter, mixed with mineral soil deposited by rivers. B. racemosus mainly occurs in the central and western part of the Dutch river area (FLORON 2014), where the amplitude of water levels is smaller and more influenced by the sea than in the eastern part of the country (AGGENBACH et al. 2007). Within the river landscape, the sites are located in floodplains of meandering rivers between the smaller summer dike and the higher winter dike, and in polders inside the dikes. The soils of the polder sites are overbank deposits in flood basins with transitions to natural levees (KOOMEN & MAAS 2004). The soil has a loamy or clayey top layer, sometimes with a layer of peat below. Towards the west the natural levees are lower and the texture is finer. The flood basins are larger and wetter, because they are flatter and because of tidal influence; therefore peat was formed in the past (JONGMANS et al. 2013). On some locations, the soil was levelled or excavated during the last centuries (ALTERRA 2013), or soil was added from ditches that were dug. Ditches were created in all of the polder grasslands, to enhance drainage in winter and to prevent desiccation in summer, by letting in river water. In many reserves a natural water regime has been set, with a higher water level in winter and a lower level in summer (D. Kerkhof pers. com.). Because of drainage, the peaty layer oxidised and subsidence occurred. The sides of the lots along the ditches remained wetter and therefore subsided less than the middle of the parcels, resulting in a bathtub shape. The water in these polders is often a mixture of river water, rain water and ground water. It is well buffered, but not very calcium rich (D. Kerkhof pers. com.). Most sites in the floodplains still have their original relief, only few are levelled or excavated. The sites lie on meandering ridges and channels (in one case forming a pointbar) and on flat plains (KOOMEN & MAAS 2004). A few sites lie on short steep slopes and dikes. Some sites have flat or undulating deposits from tidal rivers (KOOMEN & MAAS 2004), that are nowadays separated from the river by summer dikes. The sites on the Wadden sea island of Texel are from marine origin and still have brackish groundwater (VAN GOETHEM & VAN ROOIJEN 2011). De Bol is situated in a former sea bay that was closed in 1875 (KLOOSTERHUIS et al. 1986). Dijkmanshuizen was diked already at the end of the 14th century. Around polder Waal en Burg first dikes were created in 1488, but they could not withstand the sea; only since 1612 proper dikes protect the polder (WESTHOFF & VAN OOSTEN 1991). In De Bol and Waal en Burg B. racemosus grows on flat plains and near former tidal creek beds. Dijkmanshuizen is situated on a flat area that was eroded by the sea (KOOMEN & MAAS 2004). In the 20th century the drainage systems in the polders of Texel were improved, leading to lower groundwater tables and a taller vegetation (WESTHOFF & VAN OOSTEN 1991).

2.2 Nutrient availability Large parts of the Netherlands have been enriched with nutrients during the last fifty years. Most agricultural land has been heavily fertilised. A high atmospheric nitrogen deposition increases the nutrient availability everywhere (BOBBINK & HETTELINGH 2011). The river water quality improved since the 1970â&#x20AC;&#x2122;, but is still polluted with sulphate (LOEB et al. 2007) and nutrient rich. In floodplains, nutrient limitation still occurs in some nature reserves, but in many sites fertilisation and/or flooding with eutrophicated water terminated nutrient limitation (LOEB et al. 2009).

2.3 Climate The Netherlands has a maritime temperate climate. In the study sites, the average temperature ranges from around 3.0 degrees in January to ca. 17.5 degrees in July; the mean annual temperature is 10.0 degrees. The precipitation is on average 800-850 mm per year, with a precipitation surplus of 200-300 mm (KNMI 2011). The spring of 2013 was colder than normal, and vegetation developed relatively late. Most plants started to flower two weeks later than on average in the past twelve years, and a few days later than in the mid of the 20th century (VAN VLIET 2014).

2.4 Sites in surrounding countries Data from surrounding countries were analysed and compared to the data from the study sites in the Netherlands. Only sites from lowland areas were selected. All sites are situated in the vicinity of a river or stream. These sites include a region near Bremen in Germany, four nature areas in Belgium, four in northern France and two in southern England (fig. 5 presents a map; the locations and references are listed in table 5 in chapter 4.4).

3. Material and methods An overview of the methods and data used per sub-question can be found in table 2, at the end of this chapter.

3.1 Fieldwork Fieldwork has been done in the period 19 May - 5 July 2013 in 28 nature reserves in the Netherlands where B. racemosus had been observed in recent years; 144 relevés have been made (107 solely by Max Simmelink and 37 together with experienced vegetation researchers). Further 18 relevés from the Dutch National Vegetation Database (DNVD; SCHAMINÉE et al. 2006, DENGLER et al. 2011) from the period 2007-2010 were added. Their locations were visited to obtain additional environmental and soil data. In total 162 relevés were included in the research. Only 6% of all relevés (originating from three sites) contained the rare B. racemosus subsp. commutatus. The managers of the visited sites were asked to describe the management of the parcels were relevés have been made. Mostly they could indicate how often and when the specific parcels were mown, and if there was aftermath grazing. Sometimes this information was only available as an average for the whole grassland area.

3.1.1 Vegetation description The Braun-Blanquet method was used (BRAUN-BLANQUET 1964, SCHAMINÉE et al. 1995), with the Braun-Blanquet scale for cover-abundance adjusted by BARKMAN et al. (1964). Relevés of 9 m2 have been made, if possible in a square of 3*3 m, if necessary (because of ditches or gradients) in a rectangular form of 2*4,5 m. Relevés were placed on locations with B. racemosus on representative, homogenous sites, preferably in the middle of patches where the species was present. Locations with a relatively high cover of B. racemosus were chosen more often. If the species occurred within a gradient, additional relevés just outside the zone with B. racemosus have been made in a few cases (nine relevés in total).

3.1.2 Soil and humus profile description For half of the relevés (81), an earth auger was used to study humus and soil profile until a depth of around 95 cm. Additionally for 152 relevés a bread knife was used to inspect the upper 20 cm of the humus profile without destroying the structure. The humus and soil characteristics were described according to the method of the Field guide Humus Forms (VAN DELFT et al. 2006). The humus forms were classified with help of Rein de Waal, using descriptions and photos from the field. Hydromorphic properties of the soil were used as an indication for groundwater table fluctuations. The texture of the soil was estimated by hand. The texture of samples from the above 20 cm was controlled afterwards by Rein de Waal.

3.1.3 Soil sampling and analysis For every relevé one composite soil sample has been taken. With a gouge ten subsamples of 20 cm depth were taken in a systematic pattern within the relevé, and mixed in a bag. The soil samples were taken to the laboratory within 1-10 days (within that period they were stored in a refrigerator if possible). There they were stored in a freezer for around 3 months. A soil chemical analysis of the composite samples was executed in the soil laboratory of the chair group Nature Conservation and Plant ecology with help of Jan van Walsem. The soil analysis was done according to the guidelines as described in HOUBA et al. (1995). The following measurements were carried out: Organic Matter percentage (OM%), with the loss of ignition method; pH-H2O and pH-KCl; N-total and P-total (both in mmol/kg, by destruction); Pw (mg/kg, by water extraction); K and Ca (both in mg/kg). K and Ca were measured by water extraction, using the Pw extraction. First 2mL water were added to the soil (1,2mL), and 22 hours later the rest of the water (70mL) was added. This is not standard, but deviates little from other water extraction methods. N-total, P-total and Pw were measured in the auto analyser “Skalar San Plus System auto analyser”. Ca and K were measured in the spectrometer “Fast sequential Atomic Absorption Spectrometer, Varian AA240FS”. NO3 and NH4 were not recorded, because the soil became to warm during fieldwork to measure reliable values. The CN-ratio was estimated by the formula (OM%/2)/(N-total* 14.0067/10000).

3.2 Other datasets Vegetation samples made by the Centrum voor Landbouwpublicaties en Landbouwdocumentatie (CLM) in the period 1934-1958 (with a peak in the period 19461953) and published in KRUIJNE et al. (1967) were analysed. The CLM used a sampling method with around 100 subsamples of 25 cm2 per sampled parcel, and environmental and management data were collected. For most analyses only the 414 samples from May and June were used (from a total of 1695 samples), of which 87 contained B. racemosus. Outside that period B. racemosus was seldom observed (in 28 of 1281 samples), and probably often overlooked. Further PQ’s made by for the province of South Holland in the period 1991-2013 were analysed. These PQ’s are also included in the DNVD. The 206 PQ’s in which B. racemosus had been observed at least once were used; 790 relevés made between 1 May and 10 July were included, of which 420 relevés contained B. racemosus. Additionally the total set of 958 relevés containing B. racemosus from the DNVD were analysed. Finally relevés from surrounding countries were acquired, to compare them with the Dutch relevés (see chapter 2.4).

3.3 Data analysis All data from fieldwork were entered in Turboveg (HENNEKENS & SCHAMINÉE 2001). Mean Ellenberg indicator values (ELLENBERG et al. 2001) and Wamelink indicator values (WAMELINK et al. 2007) were calculated for all relevés, based on ordinal cover.

The relevés from other countries were made with several cover scales. For the analysis these were transformed to the ordinal scale (VAN DER MAAREL 1979), which corresponds to the Braun-Blanquet scale for cover-abundance adjusted by BARKMAN et al. (1964). For the international comparison Ellenberg indicator values were based on absence/presence, since the originally different cover scales could make the ordinal cover method less precise. The vegetation data from fieldwork were imported into Juice (TICHÝ 2002). The relevés were classified in a hierarchical divisive way with help of a modified version (which respects the cluster heterogeneity) of the programme Twinspan (ROLEČEK et al. 2009). The following settings were used: 2 cut levels: 0-25; number of clusters: 19; measure of cluster heterogeneity: total inertia; min. group size: 2. The resulting classification was improved by hand, by dividing, merging and translocating some clusters, and moving some relevés. Species were assigned to clusters according to fidelity, measured by the phi coefficient (CHYTRÝ et al. 2002). The resulting communities were assigned to (sub)associations by comparison with tables from SCHAMINÉE et al. (1996) and from SynBioSys (SCHAMINÉE et al. 2007). The relevés from the DNVD and the relevés from nearby countries were also classified with Twinspan in Juice. Canoco 5 (TER BRAAK & ŠMILAUER 2012, TER BRAAK 1988) was used for multivariate analyses. Indirect and direct ordination were used to analyse and depict the relation between plant communities and environmental gradients. Because of the short length of gradient of the field work data (2.9), the linear methods Principal Component Analysis (PCA) and Redundancy analysis (RDA) were used. Some environmental variables were log-transformed by Canoco using the formula Log(ax+b): Moss cover: a10, b1; K: a10, b0; Ca: a1, b1; Pw: a10, b1; P-total: a1, b0. For the RDA, interactive-forward-selection with the Monte Carlo permutations test was used. Layer covers, soil variables and management variables were used as environmental variables. Further the Ellenberg values for moisture and nitrogen were included, since moisture and available nitrogen are important factors that had not been measured. Only variables with P adj. (false discovery rate) < 0.05 were included. Data and results from the ordinations that are not covered in chapter 4 are presented in appendix 6. All relevés from the DNVD containing B. racemosus were compared to the relevés from fieldwork using Twinspan. The relevés from other European countries and from one Twinspan cluster from the DNDV were ordinated together with the relevés from fieldwork. A representative subset of 237 of the 412 available relevés was used, since a very crowded ordination space is difficult to interpret. The unimodal method Detrended Correspondence Analysis (DCA) was used, since the length of gradient was 3.24. The option detrending by 26 segments was chosen. IBM SPSS Statistics 22.0 (IBM CORP. 2013) was used for univariate statistics. Because many variables could not be transformed to a normal distribution, non-parametric tests were used. The cover of B. racemosus was correlated to environmental variables and to covers of some frequent species with Spearman’s rank correlation coefficient test. To test the differences of environmental values between plant communities, the Kruskal Wallis test with stepwise stepdown multiple comparisons was used. The Mann-Withney U test was used when two groups of relevés (often groups with and without B. racemosus) were compared.

3.4 Additional environmental data Measurements of the groundwater table of some sites were available on www.dinoloket.nl (GDN 2013), indications for the annual groundwater table fluctuations (grondwatertrappen) on www.bodemdata.nl (ALTERRA 2013). Inundation frequencies of some floodplains could be estimated with help of AGGENBACH et al. (2007) and the websites www.live.waterbase.nl (RIJKSWATERSTAAT 2013) and www.ahn.nl (HET WATERSCHAPSHUIS 2013). The latter contains altitudinal data. Together with information from site managers and hydromorphic properties in the soil profiles, these data were used to describe the hydrology. The geomorphologic map of the Netherlands was used to determine the geomorphology of the sites (KOOMEN & MAAS, 2004). Historical maps were inspected on www.watwaswaar.nl to check for major changes in landscape or land use since around 1850 (PROJECTTEAM WATWASWAAR.NL 2013). The data from these website were not used in statistical analyses. Table 2: Overview of data and methods used per sub-question Sub-question (see below table) 1 2 Fieldwork: Relevés + + Fieldwork: soil & humus profile + description, soil analysis CLM ‘relevés’ + + DNVD relevés Relevés other countries PQ’s South-Holland + Additional environmental data + Management data + Literature + + Historical maps Twinspan Ordination + Univariate statistics + +

3 + +

+ +

+ + + + + + + +

+ +

Sub-questions: 1) Under which abiotic conditions does B. racemosus occur in the Netherlands? 2) What is the influence of grassland management on the abundance of B. racemosus? 3) In which vegetation types does B. racemosus occur in the Netherlands, and which site conditions are typical for these communities? 4) In which vegetation types does B. racemosus occur in nearby European countries? 5) How could landscape and land use history have affected the occurrence of B. racemosus in the Netherlands?

4. Results 4.1 Abiotic conditions 4.1.1 Hydrology In appendix 5, groundwater table graphs for a few relevés are presented. The Ellenberg value for moisture of the relevés varied in the range 5-8. B. racemosus cover was negatively correlated to the Ellenberg value for moisture (r=-0.206, P=0.009, n=158). However, an optimum curve is a better approximation (fig. 6), since sites with a low moisture value had a lower average B. racemosus cover. Some relevés were the species was absent might be too dry. The optimum is situated around the Ellenberg value 6.2. In the dataset with PQ’s in South Holland the optimum is comparable, around 6.0.

Fig. 6: The cover of B. racemosus related to the average Ellenberg value for moisture per relevé. R2 linear = 0.018, R2 quadratic = 0.199.

B. racemosus was negatively correlated to some species indicating moist circumstances. Two of these might compete with B. racemosus sometimes, since they can form math-like structures that could prevent its establishment; the moss Calliergonella cuspidata (r=-0.365, P<0.001) and the grass Agrostis canina (r=-0.320, P<0.001).

Inside the dikes On sites where inundation by river water does not occur, the species was only found on sites with a mean highest groundwater table* higher than 40 cm below the soil surface, mostly higher than 25 cm. The species is most abundant on sites that never or rarely inundate by rainor groundwater, and lacks on sites that are inundated for a long period each winter. In areas where the soil consists of clay on top of peat, B. racemosus has typically a higher cover on the

higher border zones of parcels with a bathtub shape. In the middle, where rain water stagnates during winter, the species has a lower cover or is absent. The mean lowest groundwater table* is usually between 60 and 110 cm. The species is absent or rare in areas where this is higher than 50 cm. Therefore the species occurs mainly on sites with Grondwatertrap II and III* (ALTERRA 2013, GDN 2013). Most sites in polders have a higher water table in winter than in summer. In a few parcels (often community 5), the water table was regulated in such a way that the neighbouring agricultural parcels would be productive, with lower water table in winter and higher in summer. Many sites had such a water table for a period in the past, when they were agricultural land. * The mean highest groundwater table (GHG) is calculated over a period of eight years, using the three highest groundwater measurements of each year (the groundwater is measured 24 times per year). For the mean lowest groundwater table (GLG) the three lowest measurements are used. The Grondwatertrap is a classification based on these two values (II means GHG<40, GLG 50-80, III means GHG<40, GLG 80-120).

Outside the dikes In areas that are flooded by river water less then about once every ten years (like along parts of the Lek), the species was only found in sites with roughly the same groundwater table as inside the dikes (RIJKSWATERSTAAT 2013, GDN 2013). On locations that are flooded more often, the inundation frequency and time probably determine the limits of B. racemosus. The species occurs here at relatively dry places with mean highest groundwater table lower than 80 cm below soil surface (Grondwatertrap VII). Its upper boundary could often be related to the height that is inundated every few years. On its lower boundaries B. racemosus is again limited by long inundations, but it seems to tolerate inundation with river water for a longer period (probably almost two months) than inundation with rain and groundwater inside the dikes. Two examples of sites where B. racemosus occurs relatively high in the gradient over larger areas are Cortenoever and the Amerongse Bovenpolder. In Cortenoever the species occurs between a height of around 6m (almost the lowest parts of the floodplain) and 7.6 m. This upper border seems to correspond roughly to the altitude which is inundated most years (AGGENBACH 2007, RIJKSWATERSTAAT 2013, HET WATERSCHAPSHUIS 2013). In the Amerongse Bovenpolder the species occurs between around 5m and 7.5 m (the highest relevĂŠ is 7.2 m). It also occurs at the highest parts of the natural levees, and on small dikes. The higher loam contact compared to Cortenoever may enable the species to grow on higher places. The Amerongse Bovenpolder will probably inundate completely when the water height reaches a value of maximal 7.2 m above NAP. During the last twenty years this happened seven times, but during the last ten years only twice. The lowest relevĂŠs lie around 70cm below the average river level of 6.0 m. They probably become inundated more often by rising groundwater (Rijkswaterstaat 2013, Het Waterschapshuis 2013). The Hengstpolder is inundated with river water every year relatively long, for around fifty days. Some other sites (Willige Langerakse Waard, Genninger Buitenlanden and De Brommert) also seem to become inundated regularly for moderate long periods; most years with river water but every year with ground water. In the lowest parts of these floodplains, were water stagnates for long periods, the species lacks. Inundation in the flowering period is very rare, but became more frequent during the 20th century (AGGENBACH et al. 2007). It was observed in Cortenoever in the begin of June 2013

for a few days (according to the manager, this was the second time in 30 years; fig. 7). B. racemosus died and was not able to produce seeds in zones were the water height was more than ca. 80 cm. The lower the water depth had been, the more individuals were still able to survive and produce seeds.

Fig. 7: The flooding of Cortenoever in the begin of June 2013. Only the higher ridges were not flooded. After the flooding the vegetation of the lower parts was partly dead, B. racemosus did not survive (photos: Max Simmelink).

4.1.2 Nutrient availability The data from fieldwork (n=158) indicate that B. racemosus is more abundant on relative nutrient rich, productive sites; its cover is correlated negatively to the Ellenberg value for light (r=-0.359, P<0.001) and positively to the Ellenberg value for nitrogen (r=0.252, P=0.001), N25

total (r=0.211, P=0.008), P-total (r=0.289, P<0.001), Pw (r=0.250, P=0.002) and K (r=0.162, P=0.041). Further it may be correlated negatively to the CN-ratio (r=-0.154, P=0.054); this ratio varied most often between 7-20, rarely between 20-30. The Wamelink indicator value for the CN-ratio is correlated stronger (r=-0.324, P<0.001). Further there could be a weak positive correlation to OM% (r=0.148, P=0.064), which varied in the range 4-40%. On Texel, the species seemed to be absent in sites without a humus containing top layer. B. racemosus had the highest positive correlation with the following, mostly nitrophilous species (n=162): Lolium perenne (r=0.503, P<0.001), Trifolium dubium (r=0.290, P<0.001), Ranunculus repens (r=0.285, P<0.001), Taraxacum officinale s.l. (r=0.253, P=0.001), Poa trivialis (r=0.247, P=0.002), Brachythecium rutabulum (r=0.244, P=0.002), Holcus lanatus (r=0.220, P=0.005), Cerastium fontanum subsp. vulgare (r=0.215, P=0.006) and Bromus hordeaceus (r=0.211, P=0.007). The correlation with Alopecurus pratensis is also quite strong (r=0.251, P=0.003, n=139) if the relevés from Texel (where A. pratensis was rare) are excluded. The species Alopecurus pratensis, Poa trivialis, Lolium perenne and Ranunculus repens were correlated stronger to most indicator values and measured values related to nutrients than B. racemosus. B. racemosus cover was positively correlated (n=162) to herb cover (r=0.358, P<0.001) and negatively to moss cover (r=-0.183, P=0.021). B. racemosus cover was also negatively correlated to species richness (r=-0.288, P=0.001). These correlations can be related to the fact that B. racemosus cover is higher under nutrient richer conditions. B. racemosus was not observed in extremely nutrient rich sites. Hardly any of the visited sites had been fertilised within the last five years. The Ellenberg value for nitrogen of relevés with B. racemosus varied between 4.36-6.42. In the DNVD only 2,5% of the 958 relevés with B. racemosus have higher Ellenberg values, but 6% of the relevés have lower values (mostly 3.24.3). In the data from the CLM B. racemosus is negatively correlated to measures for the availability of K (K-getal; r=-0.258, P<0.001, n=395) and P (P-getal; r=-0.121, P=0.015, n=406); it was not found on sites with very high values. However, from the remarks about fertilisation (only checked for relevés with B. racemosus) it appears that its cover is somewhat higher in fertilised than in unfertilised sites. This might indicate that a low dose of fertiliser has a positive effect, but heavy fertilisation a negative effect on the species. Further B. racemosus was positively correlated to OM% (r=0.150, P=0.002, n=407), although it was almost absent from sites with more than 50% OM. From the dataset with PQ’s in South Holland the following correlation of B. racemosus cover with indicator values was found (only relevés with B. racemosus included, n=420): Ellenberg value for nitrogen (r=0.109, P=0.026) and Wamelink value for CN ratio (r=-0.287, P<0.001). Further it was found that several of Wamelink indicator values for nutrient richness were decreasing until the establishment of B. racemosus in the PQ’s, and the species richness of the PQ’s became higher (r=0.199, P<0.001, n=435). This could indicate a decrease in nutrient richness before B. racemosus established. Most indicator values did not change significantly between the period with presence of B. racemosus and the period after disappearance. However, the CN-ratio became significantly higher after its disappearance than during the presence of B. racemosus (table 3). From the last observation of B. racemosus onwards, the litter cover increased (r=0.319, P<0.001,

n=151). Further there were significant differences between the period after disappearance of B. racemosus compared to before and during its presence for the following parameters (table 3): the herb cover decreased, the bare soil cover increased and the Ellenberg value for light increased. These changes indicate reduced productivity, slower mineralisation and lower nutrient availability. B. racemosus might disappear because germination is more difficult when the litter cover increases and/or because of low nutrient availability. The changes in cover of other grass species before, during and after the presence of B. racemosus were investigated (table 3). The cover of Poa trivialis decreased every phase. Lolium perenne, Holcus lanatus and Agrostis stolonifera had the same cover before as during the presence of B. racemosus, but a lower cover after B. racemosus disappeared. Alopecurus pratensis cover was highest during the presence of B. racemosus and lowest after. Agrostis capillaris increased every phase, but was still quite rare after the disappearance of B. racemosus (present in 22% of the relevĂŠs). The cover of B. racemosus was positively correlated to (n=790) the cover of Alopecurus pratensis (r=0.184, P<0.001), Lolium perenne (r=0.163, P<0.001) and Poa trivialis (r=0.095, P=0.007). These species were stronger correlated to the Ellenberg value for nitrogen (r=0.418/0.291/0.456, P<0.001) than B. racemosus (if all 790 relevĂŠs are included: r=0.062, P=0.081). Table 3: Results of the analysis of the PQ dataset of South Holland to find differences between the periods before the first observation of B. racemosus, during its presence and after its (temporal or definite) disappearance. For each period the median value is given, numbers between brackets are mean values (only given if the median provides insufficient information about the differences; this is the case for the cover related variables, since these have been measured using scales with few levels). The letters indicate significant differences. Test: Kruskal Wallis test with stepwise stepdown multiple comparisons. The P-values represent the asymptotic significance (2-sided test). E is an abbreviation for Ellenberg value and W for Wamelink value. B. racemosus was excluded for the calculation of the indicator values. The cover of species is according to the ordinal scale. Variables that did not show a significant difference are not presented (including moss cover, the Ellenberg values for nitrogen, reaction and moisture and the Wamelink values for NH4, PO4 and Ca). N=790, d.f.=2. Period Before During After Test stat. P-value Species richness 32 a 36 b 36 b 46.991 <0.001 E_light 6.94 a 6.96 a 7.01 b 9.586 0.008 W_CN ratio 13.0 b 12.8 a 13.2 b 11.790 0.003 W_K 191 b 181 a 181 a 26.068 <0.001 W_P-total 548 b 524 a 546 ab 7.620 0.022 W_N-total 4687 b 4371 a 4592 ab 7.598 0.022 W_NO3 7.20 b 6.93 a 7.02 ab 6.711 0.035 Average height low herb layer 27 b 30 b 25 a 15.126 0.001 (cm) Cover herb layer (%) 94 (92.6) b 94 (92.9) b 94 (89.9) a 12.684 0.002 Cover litter layer(%) 2 (4.8) a 2 (6.1) a 2 (9.9) b 16.975 <0.001 Cover bare soil (%) 2 (2.3) a 2 (2.6) b 2 (3.35) c 19.948 <0.001 Poa trivialis cover 5 (5.2) c 5 (5.0) b 4 (3.7) a 47.654 <0.001 Lolium perenne cover 3 (2.7) b 3 (2.9) b 2 (2.0) a 13.973 0.001 Holcus lanatus cover 4 (3.4) b 4 (3.2) b 3 (2.7) a 10.018 0.007 Agrostis stolonifera cover 4 (3.9) b 4 (3.8) b 4 (2.9) a 19.569 <0.001 Alopecurus pratensis cover 3 (2.7) b 3 (3.0) c 2 (2.0) a 23.607 <0.001 Bromus hordeaceus cover 2 (1.8) b 0 (1.1) a 0 (1.1) a 24.309 <0.001 Agrostis capillaris cover 0 (0.2) a 0 (0.5) b 0 (0.8) c 25.086 <0.001

4.1.3 Other soil characteristics The acidity varied in the following ranges: pH-H2O 5.1-7.9, pH-KCl 4.0-7.5, Ellenberg value for reaction 4.95-7.13. The pH and the Calcium concentration are not significantly correlated to the cover of B. racemosus. B. racemosus occurs normally on soils with a fine texture. In the data from field work, loam percentage was 20-90% and clay 4-26%, but B. racemosus cover is not significantly correlated to those parameters. If only the driest communities (3-5) are selected, there is a positive correlation to loam percentage (r=0.223, P=0.040, n=85). In the data from the CLM the species is more often present on sites with a finer texture (Mann-Withney U: 16402.5, P=0.011, n=407), and it was also found on sites with a clay percentage much higher than 26%. However, it was also found a few times on sites with a coarse texture. Almost all relevĂŠs had a type of mull as a humus form; these are humus forms with a thin or absent ectorganic horizon and a good mineralisation. No soils with a podzol were found.

4.2 Grassland management According to the data from the CLM, the species was more frequent in meadows with aftermath grazing than in grasslands that are only mown or grazed (Test stat.=77.741, P<0.001, d.f.=5, n=380). In the months May and June, B. racemosus was found in 12,5% of the 48 meadows that were only mown, in 48,3% of the 116 meadows with aftermath grazing and in 4,3% of 92 pastures that were sampled. This makes B. racemosus the species with the strongest preference for meadows with aftermath grazing of around 200 common grassland species studied by the CLM (KRUIJNE et al. 1967). Besides the management, also abiotic conditions might cause this strong preference. Meadows that were not grazed were often wetter than optimal for B. racemosus, and pastures were often too dry or more nutrient rich than optimal. With the data from fieldwork, it was tested whether B. racemosus reaches a higher cover on meadows with aftermath grazing, compared to meadows that are only mown. This was indeed the case (Mann-Withney U: 1292.0, P=0.007, N=151, fig. 8). Within the 120 relevĂŠs with aftermath grazing, it was tested whether the yearly number of mowings was related to the cover of B. racemosus (92 relevĂŠs were mown once, 28 twice). The cover was higher when mown twice (Mann-Withney U: 1689.5, P=0.010, N=120, fig. 9).

Fig. 8: Cover-abundance of B. racemosus on the ordinal scale in meadows with and without aftermath grazing.

Fig. 9: Cover-abundance of B. racemosus on the ordinal scale in meadows with aftermath grazing that are mown once or twice a year.

Most sites visited during fieldwork are usually mown in the second half of June or later. Only two productive, species poor sites were mown earlier, around the first of June; here B. racemosus was locally abundant. In pastures, the cattle density should be low enough to allow seed production. During fieldwork the species was found quite abundant in two extensively grazed pastures, with mosaics of long and short vegetation. It grew in intermediate grazed sites, but was not found in shortly grazed vegetation and very rare in tall vegetation. In the Brakelse Benedenwaard B. racemosus subsp. commutatus grew quite abundant along cattle paths.

4.3 Vegetation types in the Netherlands 4.3.1 Fieldwork The classification resulting from Twinspan was improved by hand, by dividing, merging and translocating some clusters, and moving some relevés. The relevé table is included in appendix 1. The original Twinspan classification is given in the header data. The original clusters 14 and 15 were not considered sufficiently homogenous. Therefore they were further divided with help of Twinspan, and the resulting clusters were partially merged with other clusters. Three relevés were deleted from the table; two of these (WA7 & WA8) formed one cluster without occurrence of B. racemosus, the other relevé (LA2) was considered insufficiently homogeneous and contained few B. racemosus. The classification has resulted in eleven clusters of relevés (fig. 10 and appendix 1 & 3). These are separated also in the PCA if the first three axes are considered (fig. 11 and 12). The first axis can be interpreted as a moisture gradient, the second as a pH gradient and the third as a nutrient richness gradient. B. racemosus cover is mainly correlated to axis 1 (r=-0.3114) and 3 (r=-0.3262); in fig. 13 the cover of B. racemosus is plotted together with some correlated environmental variables. In the RDA, axis 1 stands for moisture, axis 2 for nutrient richness and axis 3 for pH. The Ellenberg value for moisture explained 10% of the variation, the Ellenberg value for nitrogen 6.5% and the pH-H2O 4.5%, all with P(adj.)=0.002. In appendix 6 the graphs and the detailed results from the RDA are given. All relevés from fieldwork can be placed within the class Molinio-Arrhenatheretea, based on the large number of character species of this class in all relevés. The character species with the highest constancy values in the relevés are Cerastium fontanum subsp. vulgare, Ranunculus acris, Rumex acetosa, Trifolium pratense, Cardamine pratensis and Holcus lanatus. Within this class, B. racemosus has been found in all Dutch alliances (Calthion, Alopecurion, Arrhenatherion, Cynosurion) except the Junco-Molinion. Below for each community some characterising species and environmental factors are mentioned. In table 4 the values and significant differences of environmental variables per community can be found. In the relevé table (appendix 1) and the synoptic table (appendix 3) all species characterising the communities are listed. In appendix 4 some typical soil profiles are shown per community. Photos of all communities are presented in appendix 7.

Fig. 10: A Twinspan tree with the numbers of the communities. See also chapter 3.3 and appendix 1 (relevĂŠ table). Table 4: The median values of several environmental variables per community. The communities 9 and 10 are ignored, since they were only observed on one site and few relevĂŠs have been made. The similar communities 1 and 2 are combined to increase the sample size. Total n=153, d.f.=7. The letters indicate significant differences. Test: Kruskal Wallis test with stepwise stepdown multiple comparisons. The asymptotic significance (2 sided test) is <0.001 for all shown variables. E is an abbreviation for Ellenberg value. W_GVG is the Wamelink indicator value for the mean annual spring groundwater table in cm below the soil surface. Community

3 (n=15)

4 (n=28)

5 (n=42)

6 (n=10)

7 (n=10)

8 (n=8)

E_moisture

1+2 (n=8) 6.37 c

6.99 d

11 (n=33) 6.96 d

Test Stat. 100.274

5.93 b

5.48 a

6.33 c

6.83 cd

7.39 d

W_GVG (cm)

33.8 abc

40.8 cd

49.6 e

41.9 d

38.6 cd

31.9 a

37.0 bc

32.4 ab

101.607

E_nitrogen

4.40 a

5.04 b

5.78 c

5.73 c

5.54 c

5.31 b

5.63 c

5.03 b

101.210

Pw (mg/kg)

0.8 a

2.7 b

3.8 b

7.4 d

4.3 bc

3.4 b

6.2 cd

5.8 cd

69.434

P-tot (mmol/kg)

6.2 a

10.0 b

20.1 c

27.4 e

22.0 cd

18.7 c

36.65 e

24.9 d

82.800

K (mg/kg)

15 a

46 b

131 c

253 d

183 cd

107 bc

178 cd

278 d

59.305

pH-KCl

5.83 bc

5.81 c

7.05 d

5.04 b

4.87 ab

7.11 d

7.17 d

4.64 a

72.309

pH-H2O

6.35 ab

6.77 b

7.59 c

6.02 a

5.93 a

7.66 c

7.70 c

5.78 a

76.867

E_pH

5.48 a

5.57 ab

6.44 d

5.96 c

6.21 c

6.47 d

6.44 d

5.85 bc

83.232

Ca (mg/kg)

282 ab

142 a

654 c

330 b

194 a

609 cd

856 d

291 b

66.289

OM %

9.0 b

4.9 a

10.1 b

14.6 cd

11.8 bc

8.5 b

12.7 cd

15.4 d

79.119

Clay %

13.5 ab

10 a

12 ab

18 c

17.5 bc

10 ab

18.5 c

20 c

59.700

Loam %

63 ab

50 a

54 a

73 bc

67.5 abc

55 a

76 bc

80 c

54.155

E_light

7.14 c

7.02 c

6.86 b

6.76 a

6.78 ab

6.88 b

6.74 a

6.89 b

54.382

Herb cover (%)

88.5 a

95 c

92.5 bc

90 ab

95 bc

94 abc

90 ab

26.454

Moss cover (%)

50 c

2.5 b

10 c

43.691

# of species

32 bc

29 b

32 bc

25 a

37.5 d

41 e

31.5 bc

34 cd

65.706

Fig. 11: Principal Component Analysis (PCA) of 158 relevĂŠs from fieldwork, that have been assigned to eleven communities with help of Twinspan, as shown in the relevĂŠ table (appendix 1) and fig. 10. The communities are depicted by symbols, as explained in the legend. Axis 1 and 2 are plotted, eigenvalues are 0.124 and 0.092 respectively. Only continuous environmental variables with a correlation of more than 0.30 with one of the plotted axes are shown. Abbreviations: E: Ellenberg value, W: Wamelink indicator value. GVG: mean spring groundwater level, GLG: mean lowest groundwater level (both in cm below soil surface).

Fig. 12: see fig. 11 for explanation. Here the first and the third axis are depicted, eigenvalues are 0.124 and 0.077 respectively.

Fig. 13: The first and the third PCA axes, with symbols indicating the cover of B. racemosus in the relevĂŠs. Continuous environmental variables that are correlated to B. racemosus cover with significance P< 0.065 (according to Spearmanâ&#x20AC;&#x2122;s rank correlation coefficient test) are shown.

1&2: Fragment of Rhinantho-Orchietum morionis & Rhinantho-Orchietum morionis [Calthion palustris] These communities were only found on the island of Texel. The Rhinantho-Orchietum morionis Bruin et Weeda 1996 belongs to the Calthion palustris and is related to the Cynosurion cristati Tüxen 1947. This rare community is only known from the Netherlands in polders near the coast. The only character species of this association is Orchis morio (=Anacamptis morio; SCHAMINÉE et al. 1996). Community 1 (all relevés from Dijkmanshuizen) is considered as a fragment of the association, because Orchis morio is absent. Further these relevés were separated by Twinspan from community 2 (relevés from De Bol and Waal en Burg) already in the first division. Several differentiating species mentioned by SCHAMINÉE et al. (1996) occur in both communities; Ophioglossum vulgatum, Rhinanthus minor, Luzula campestris, Hypochoeris radicata and Leontodon taraxacoides subsp. taraxacoides (= Leontodon saxatilis). It is a low productive vegetation, with a relatively low cover of the herb layer and a high cover of the moss layer. This can be related to low nutrient concentrations (table 4). B. racemosus has a low cover in most relevés, and is completely absent in two relevés of community 2. These were placed only a few meters form relevés assigned to community 3, where B. racemosus was abundant. The meadows are mown late, after 15 July, and afterwards grazed by sheep until 1 February. The phenological development of B. racemosus is a few weeks later than on most other location, except for community 3 (also on Texel) and 9 (also mown late). The abiotic conditions are similar in both communities, but community 1 has a higher OM % and higher Ellenberg values for moisture than community 2. The soil consists of marine deposits from former salt marshes. During winter there seems to be a high groundwater table, but inundation occurs probably rarely. (SCHAMINÉE et al. (1996) mention that Orchis morio cannot withstand longer inundation). The humus forms belong to several hydromull forms, with hydromorphic properties within 25 cm. The loam content is relatively low, on average 56% and at least 24%. Calcareous material is sometimes present within 20 cm. The measured pH is neutral, but the Ellenberg values indicate a somewhat lower pH. The texture often changes between 10-15 cm depth from loamy to more sandy, but often deeper layers are loamy and/or humus rich to peaty. Some layers also contain shells.

3: Lolio-Cynosuretum [Cynosurion cristati] The Lolio-Cynosuretum Braun-Blanquet et De Leeuw ex Tüxen 1937 was also found exclusively on Texel. It shares many species with the communities 1&2, but Orchis morio is absent and several species differentiating the Rhinantho-Orchietum morionis are rare. The average number of species is lower and the grass species Cynosurus cristatus, Lolium perenne, Poa trivialis and B. racemosus have a higher cover. While the cover of the herb layer was high, a moss layer was almost absent. These differences compared to the former communities can be partly explained by higher nutrient concentrations (table 4). Most sites have a quite low to moderately high production. They are mown in July and afterwards grazed by sheep until 1 February. This management deviates from the standard management of sites with this association, which is only grazing (SCHAMINÉE et al. 1996). This community was found on young marine deposits. During winter groundwater table is high, but there were no signs for longer inundation. The Ellenberg values for moisture are 35

lower than for communities 1&2 (table 4). The humus form is mostly hydromull, sometimes wormmull (if the hydromorphic properties are below 25 cm). Calcareous material is sometimes present within 20 cm. In some of the relevés where this is not the case, a thin AMh-horizon (a weakly developed root math) occurs. The OM% of community 3 is lower than in other communities, 3.5-10%. The loam content is relatively low, on average 48%, and at least 19%. The measured pH is neutral, but the Ellenberg values indicate a somewhat lower pH. The texture often changes quite abrupt; the upper layer is mostly more loamy, but also deep layers can be loamy and/or humus rich to peaty.

4: Arrhenatheretum elatioris typicum [Arrhenatherion elatioris] The Arrhenatheretum elatioris Braun 1915 is a community typical for somewhat drier meadows in floodplains. Several species that are characteristic according to SCHAMINÉE et al. (1996) are present: Arrhenatherum elatius, Tragopogon pratensis, Crepis biennis, Galium mollugo and Peucedanum carvifolia. The relevés can be assigned to the subassociation typicum, that occurs on moister sites than the other subassociations (SCHAMINÉE et al. 1996). B. racemosus reaches moderately high covers in several relevés. However, it lacks in some relevés that have a relatively low Ellenberg value for moisture and a low loam content, like in the highest zones in Cortenoever. On two locations B. racemosus subsp. commutatus was found. One of these locations, the Brakelse Benedenwaarden, was not mown but grazed with a low cattle density. Most relevés are mown in the second half of June or the beginning of July, and grazed afterwards. The Ellenberg values for moisture are lower than in all other communities (table 4). In most locations, the ground water table is lower than 120 cm below soil surface during summer, and lower than 40 cm in winter. AGGENBACH et al. (2007) states that the groundwater table is normally deeper than 100 cm in the growing season. The community occurs mainly in floodplains on higher ridges (“arcuate terrain elevations”; KOOMEN & MAAS, 2004). A few relevés lie on lower parts of dikes and on flat terrain. River inundation frequency ranges from every year for a short period to every 10-20 years, a few relevés inside the dikes are never inundated. SCHAMINÉE et al. (1996) states that the association only occurs on sites that are inundated less than 20 days per year. The humus form is most often a terrestric mull (without hydromorphic properties) with calcareous material within 20cm; often Kalkwormmull, sometimes Nesvaagmull (when the Ahorizon is developed weakly). Deep homogenisation, fast mineralisation and a relatively low OM % are characteristic for such soils (JONGMANS et al. 2013). The pH is higher than in some other communities and nutrient concentrations are moderately high (table 4). The loam content is relatively low, 56% on average. The texture often changes gradual within the profile, regularly deeper layers are more sandy (fining upwards), but fining downwards also occurs.

5: BC Bromus racemosus-Alopecurus pratensis-[Molinio-Arrhenatheretea] This basal community is the poorest in species of all communities (table 4). Most species are characteristic for the Molinio-Arrhenatheretea. No species have a high fidelity to this community, but Ranunculus repens, Alopecurus pratensis, Alopecurus geniculatus and Veronica arvensis occur more often than in most other communities. This basal community could alternatively be assigned to the Alopecurion pratensis, since Alopecurus pratensis is present with often moderately high covers. Further the vegetation is intermediate between the alliances Molinietalia and Arrhenatheretalia, which is typical for the Alopecurion pratensis according to SCHAMINÉE et al. (2006).

Because B. racemosus reaches its highest cover (often 2a-3) in this community, the basal community can be named after the species. The species that often reach higher covers (Poa trivialis, Ranunculus repens, Lolium perenne and Holcus lanatus) have a much broader ecological amplitude. The most similar basal community described in SCHAMINÉE et al. (1996) is Alopecurus pratensis-Lychis flos-cuculi-[Alopecurion/Molinietalia]. In that community Alopecurus pratensis is more often dominant than in the basal community described here. Further Lychis flos-cuculi is very rare in these relevés, Symphytum officinale is rare and Carex acuta is absent. There are relevés within this community that are atypical and have a lower cover of B. racemosus; since these do not form a clear group, they were not excluded. The most deviating relevés are SB2 and KB4; these are close to the Ranunculo-Alopecuretum geniculati Tüxen 1937, which belongs to the Plantaginetea majoris Tüxen et Preising in Tüxen 1950. That community is found on sites that are inundated longer. B. racemosus only occurs in the transitions of that community to drier grasslands, like this basal community. Several parcels belonging to this community were owned by farmers until recently, and had been fertilised heavily. According to a manager, they had been very species poor and dominated by Lolium perenne, Poa trivialis and sometimes Holcus lanatus or Alopecurus pratensis. They were bought by nature organisations, which have stopped fertilisation and often mow the grasslands twice a year, to decrease the nutrient availability. Some parcels were bought around five years ago and already had a locally high cover of B. racemosus. Other parcels were bought 10-40 years ago. 55% of the relevés is mown twice a year, and 79% is grazed after the last mowing (mostly by cows, some by sheep). One location with this community, a part of Polder Achthoven (relevés PA7&8), was not mown but grazed with a low cattle density. This community occurs on fluvial deposits with a relatively high OM % (5-40%) and a high loam content (mostly 50-90%). It is not bound to a certain geomorphology, soil type or humus form within the order mull. Some of the relevés outside the dikes are often or regularly inundated by river water. Most relevés inside the dikes are never or rarely inundated. On location that are inundated every winter, B. racemosus is rare to absent. The most characteristic abiotic factors are a high nutrient availability (table 4) and a moderately high to high groundwater table during winter.

6: Fritillario-Alopecuretum pratensis [Alopecurion pratensis] This quite species rich community was assigned to the Fritillario-Alopecuretum pratensis Westhoff et Den Held ex Corporaal, Horsthuis et Schaminée 1993. The relevés range from moist to quite dry. The wettest relevés have some similarity to the Calthion, the driest to the Arrhenatherion. Fritillaria meleagris is a local character species. Further the combination of Alopecurus pratensis, Ranuculus ficaria (=Ficaria verna), Carex acuta and Sanguisorba officinalis indicates that the association belongs to the Alopecurion pratensis (SCHAMINÉE et al. 1996). The moss layer has a higher cover than in most other communities and consists mainly of Brachythecium rutabulum and Calliergonella cuspidata. This community was found on two sites, both mown twice a year and not grazed. They are situated in floodplains along the Zwarte Water (near the IJsselmeer), on flat to slightly undulating terrain. In most winters these sites are inundated by river water for a limited period. On one locations, low dikes have been build in the past to stimulate sedimentation.

According to SCHAMINÉE et al. (1996) the community flourishes mainly near estuaries, where inundation dynamics are damped. The soil is moderately nutrient rich. Calcareous material is not present in the top layer of the soil, and the measured pH values are lower than in other floodplain communities (table 4). The Zwarte Water transports less calcareous material than the Rhine and Meuse (AGGENBACH et al. 2007). The humus forms are often beekhydromull or zure wormmull (depending on the depth of hydromorphic properties), and often a thin AMh-horizon is present. In many profiles the texture changes abruptly within 50 cm from clayey to more sandy material. Sometimes peat is present beneath the mineral top layer.

7: Sanguisorbo-Silaetum [Alopecurion pratensis] This species rich community belongs to the Sanguisorbo-Silaetum Klapp ex Hundt 1964, which is very rare in the Netherlands (WEEDA et al. 2002; WEEDA 1991). Character species are Sanguisorba officinalis and Silaum silaus. Further the high frequency of Alopecurus pratensis (although with low cover), Symphytum officinale, Lathyrus pratensis and Centaurea jacea is typical. B. racemosus had quite a low cover. The relevés HP1&2 belong to a drier variant with a higher B. racemosus cover than the other relevés. This variant is differentiated by species that also occur in the Arrhenatherion; Galium mollugo, Galium verum, Crepis biennis, Trisetum flavescens and Pimpinella major. The moister relevés are closer to the Calthion and are differentiated by Mentha aquatica, Stellaria palustris, Eleocharis palustris and Achillea ptarmica. The vegetation is mown in the end of June or the begin of July and grazed afterwards. The community was found on two sites in the western river area. In the Hengstpolder it occurs on slightly undulating tidal river deposits (KOOMEN & MAAS, 2004), nowadays separated by summer dikes from the river Nieuwe Merwede (a follow-up of the Waal). In most winters it is inundated by river water for around 50 days; it is usually inundated in February, and the water is kept within the dikes till March / April. Except for subsp. racemosus, also subsp. commutatus has been recorded (WEEDA 1991). During the fieldwork, subsp. commutatus was not observed in the relevés, but it was seen on higher zones at the base of some dikes. In the Gansooiense Uiterwaard it grows on a flat floodplain along the Bergse Maas, which was dug in 1904 (before there was only a small stream; JONGMANS et al. 2013). The river only inundates this floodplain every 5-10 years (according to the manager), but large parts of the floodplain are wet in winter because of seepage. B. racemosus grows here only on a somewhat higher border zone. This community is one of the wettest (table 4), partly because most relevés belong to the wetter variant. SCHAMINÉE et al. (1996) stated that inundation can occur in winter, but that the soil desiccates superficially in summer. The loam content is relatively low, 56% on average. Nutrient concentrations are moderately high. Calcareous material is present in the upper 20 cm of the soil, and the pH is higher than in most other communities (table 4). Therefore the humus form is often kalkwormmull, or kleidyromull (if hydromorphic properties occur within 25 cm). The profile consist of quite homogeneous loamy clay, except for a thin sandy layer at around 20 cm depth in the Gansooiense Uiterwaard.

8: Fragment of Sanguisorbo-Silaetum & Calthion palustris This moderately species rich community has no characteristic species. The species composition is most similar to the Sanguisorbo-Silaetum, because of the co-occurrence of Alopecurus pratensis, Lathyrus pratensis, Symphytum officinale and Crepis biennis. Further it has some similarities to the Calthion palustris, expressed by the presence of Calliergonella 38

cuspidata, Lychnis flos-cuculi and Caltha palustris. Frequent species that occur in both mentioned communities are Filipendula ulmaria, Vicia cracca, Galium palustre, Carex disticha and Equisetum palustre (SCHAMINÉE et al. 2007). Carex acuta and Poa trivialis reach a high cover. The sites are mown in the end of June and grazed afterwards. All relevés are situated in flat to slightly undulating terrain in floodplains along the Lek/Nederrijn. Here the groundwater table is relatively high during summer, because of sea influence respectively the presence of a weir. Most relevés are from the Willige Langerakse Waard, a flat low-lying floodplain that was sedimented by the Lek under tidal fluctuations, and afterwards protected by a summer dike (KOOMEN & MAAS, 2004). It is regularly inundated by river water during winter. The loam content is relatively high, on average 75%, and the OM % is also quite high. The soil is nutrient rich. Calcareous material is present in the top layer, and the pH is relatively high. The hums form is a kleihydromull.

9: BC Rhinanthus angustifolius-Lysimachia vulgaris-[Calthion palustris] This species rich community was only found in one location in the Dertienmorgenwaard along the Lek, in a clay pit that was dug at least 100 years ago (PROJECTTEAM WATWASWAAR.NL 2013). The pit has zones with mainly Molinio-Arrhenatheretea / Calthion species and zones were Valeriano-Filipenduletum species dominate. Further it has some similarity to the Sanguisorbo-Silaetum. The three relevés differ substantially. KD1 is least productive and most species rich, and contains Dactylorhiza incarnata, Ophioglossum vulgatum and Climacium dendroides. The most remarkable similarity between the relevés was a (mostly) high cover of Rhinanthus angustifolius and Lysimachia vulgaris. Only B. racemosus subsp. commutatus occurred here. The phenological development of B. racemosus was a few weeks later than on most other location, except for Texel. It is mown very late, in august, and never grazed. The location is rarely inundated by river water, but the lower parts are flooded in winter by ground and rain water. Calcareous material was present in the top layer, the pH was quite high (pH-H2O=7.7) and nutrient concentrations were moderate. Two humus forms were found: kleihydromull and nesvaagmull (with an A horizon of <2cm).

10: Angelico-Cirsietum oleracei [Calthion palustris] The Angelico-Cirsietum oleracei Tüxen 1937 em. Raabe 1946 has been found on one location, a very gentle slope of the stream valley of Het Merkske. It is a very species rich meadow, called Waterbeemd. The species Cirsium oleraceum and Crepis paludosa are character species for this Calthion association. Carex acutiformis is a differentiating species, further Primula elatior, Filipendula ulmaria and Juncus acutiflorus are often found in this community (SCHAMINÉE et al. 1996). There was a very high moss cover, dominated by Calliergonella cuspidata. Most sites with this vegetation have a lower moss cover and are more rough (SCHAMINÉE et al. 2007). B. racemosus was very rare in this meadow. The Waterbeemd is mown late, in the end of July or in August, and never grazed. Seepage of calcareous groundwater from a deep soil layer causes a neutral pH (pH-H2O=6.5) and a constant high groundwater table, but flooding does not occur. There is a dominant peaty OAh horizon of around 30 cm, the humus form is therefore moereerdmoder. Below this layer clay is present until around 80cm, beneath there is sand. In comparison with other communities, Ellenberg values for moisture and the OM % are high; nutrient concentrations are moderate. 39

11: Ranunculo-Senecionetum aquatici juncetosum articulati [Calthion palustris] The community Ranunculo-Senecionetum aquatici Van Schaik ex Schaminée et Weeda 1996 corresponds to the communities Senecioni-Brometum racemosi Tüxen et Preising 1951 nom.amb. and Brometum-Senecionetum (Tüxen 1951) Oberdorfer 1957. Because Senecio aquaticus and B. racemosus rarely occur together, and B. racemosus occurs often in other alliances, SCHAMINÉE et al. (1996) deviated from these name. Also MEISEL (1969 in LUTZ 1996) regarded the name as inappropriate, because the name giving species don’t have a high frequency in and fidelity to this community. LUTZ (1996) states that both species are most frequent in the Senecioni-Brometum racemosi, but their optima are different; Senecio aquaticus might prefer more constant moist sites. The character species Senecio aquaticus was not found in this community. The species Juncus effusus and Ranunculus flammula differentiate the association against other Calthion associations. The presence of the species Glyceria fluitans, G. maxima, Rumex crispus, Lolium perenne and Agrostis stolonifera is typical for the subassociation juncetosum articulati. This subassociation is characterised by high nutrient concentrations, a high clay content and high groundwater fluctuations (SCHAMINÉE et al. 1996). Sometimes Carex acuta, C. disticha and/or Poa trivialis reach a high cover. Some relevés are also similar to the Lolio-Cynosuretum loletosum uliginosi. This Cynosurion subassociation usually originates from former Calthion vegetation that are fertilised, grazed and often drained (SCHAMINÉE et al. 1996); these factors might explain this similarity. The moss cover is relatively high, Calliergonella cuspidata and Brachythecium rutabulum are frequent. Most site are mown relatively late, during the last days of June or the first of July. 43% of these relevés are mown twice a year and 86% is grazed after the last mowing. Most parcels were bought by nature organisations longer ago than those of community 5, around 25-50 years ago. According to a manager of many parcels with community 5 and 11 (R. Garskamp pers. com.), those with community 11 were never heavily fertilised and already moderately to very species rich when acquired. Those with community 5 were moderately to very species poor when bought. Almost all relevés are situated in polders on overbank deposits in flood basins and on transitions to natural levees (KOOMEN & MAAS, 2004). The association can also occur along streams and in floodplains (like along the Linge near Asperen), but only where the amplitude is less than 3m (AGGENBACH et al. 2007). The communities has high Ellenberg values for moisture (table 4). In the winter the ground water table is high, and in some sites inundation by a mixture of rain and river water occurs for a short period. B. racemosus is usually rare or absent in lower lying, longer inundated parts of these meadows. These parts are often the central parts of the parcel, which often have a bathtub shape. The sides of the parcels along the ditches are higher and have more B. racemosus The lower parts are often more typical for the Ranunculo-Senecionetum aquatici, while the zones with most B. racemosus have some similarity to community 5. The top soil consists out of clay, beneath often a thick layer of peat can be found, starting between 25 and 95 cm. On some of the locations, the soil was levelled or excavated during the last centuries (ALTERRA 2013). The OM % is higher than in most other communities, often 12-20%. Also the loam content is relatively high, on average 77%. The P and K concentrations are higher than in most other communities, but the Ellenberg values for nitrogen are moderate (table 4). Most often calcareous material is not present in the upper 20 cm; the average pH is the lowest of all communities (table 4). The humus form is typically 40

beekhydromull, or in case of a lower groundwater table beekvaagmull (A horizon <5cm) or zure wormmull (A>5cm). A few relevés have a thin AMh-horizon.

4.3.2 Dutch National Vegetation Database The 958 relevés from the Dutch National Vegetation Database (DNVD; SCHAMINÉE et al. 2006) were classified by Twinspan, to discover clusters that deviate from the communities found during fieldwork. The three most deviating clusters are mentioned here. One cluster with 20 relevés is characterised by species indicating salty or desalinating conditions: Alopecurus bulbosus, Juncus gerardii and sometimes Glaux maritima, Aster tripolium, Plantago maritima and Artemisia maritima. This community could be assigned to the Trifolio frageri-Agrostietum stoloniferae (Lolio-Potentillion anserinae, Plantaginatea majoris). It shares also species wit the Armerion maritimae. Most of the relevés are from around 1940 along the former coast of the Zuiderzee (a bay of the North Sea), which had been dammed in 1932. There is only one such a location with a recent relevé (from 2010), the small reserve Merrevliet on the former island of Voorne. B. racemosus has a low cover in this association, typically r-1. A cluster with 17 relevés indicates very wet, nutrient poor conditions and is mainly situated in peat areas. It is a Parvocaricetea community (Scorpidio-Caricetum diandrae and/or Carici curtae - Agrostietum caricetosum diandrae) with Pedicularis palustris, Carex diandra, Lathyrus palustris, Eriophorum angustifolium, Hierochloe odorata and Carex elata. The cover of B. racemosus is typically low, r-1. Finally, a cluster of 51 relevés is assigned to the Ranunculo-Senecionetum caricetosum paniceae (SCHAMINÉE et al. 1996). Species with a high fidelity to this cluster are Galium uliginosum, Dactylorhiza majalis, Carex panicea, Potentilla palustris, Senecio aquaticus, Deschampsia cespitosa and Luzula multiflora. Ellenberg values indicate conditions that are wetter, more nutrient poor and more acid than in community 11, and somewhat drier and richer then in the cluster mentioned above. All relevés are from the period 1936-1978, mainly in the northeast of the Netherlands. Another cluster that is in between both subassociations of the Ranunculo-Senecionetum contains several relevés with Carex panicea from the period 1991-2009 located in South Holland (like in the Smoutjesvlietlanden north of Goudriaan and the Tuinder- of Kogjespolder south of Kaag).

4.4 Vegetation types in surrounding countries Most relevés from the region of Bremen (Germany), Belgium, northern France and southern England are quite similar to the Dutch relevés and are assigned to the same alliances. Table 5 shows which alliances (according to the syntaxonomic system of SCHAMINÉE et al. 1996) can be assigned per country. B. racemosus can reach a high cover in most alliances, but in the Calthion the cover is relatively low for each country. In the DCA (fig. 14, see also appendix 6) the Dutch relevés are distributed over a large part of the ordination space. Most relevés from fieldwork are placed on the nutrient richest half of the ordination space (the lower part), except for the relevés from Texel (all Cynosurion relevés and the uppermost group of Calthion relevés). These latter Rhinantho-Orchietum morionis relevés do not resemble any other relevés. The Dutch relevés in the nutrient poor, wet part (middle left) are mainly from the Ranunculo-Senecionetum caricetosum paniceae (from the DNVD). 41

The available relevés from Bremen (Germany) and Flanders (Belgium) are relatively wet and acid. They belong to the Calthion and basal communities of nutrient rich sites. The relevés from Famenne & Fagne (southern Belgium) and the Bassin de la Sambre (northeastern France) are drier and more basic. This Alopecurion vegetation is characterised by the presence of Colchicum autumnale and a high cover of Alopecurus pratensis and B. racemosus. The French relevés span a range that is comparable to the Dutch. The English relevés overlap only with the drier Dutch Alopecurion relevés. Although their Ellenberg value (table 6) for moisture is almost the same as for the Dutch relevés, their position in the ordination indicates drier conditions. Further, these relevés have remarkably the highest average Ellenberg value for continentality (3.2 instead of 2.7-3.0).

Table 5: The table demonstrates that groups of relevés separated per country can be assigned to the alliances following SCHAMINÉE et al. (1996). The columns indicate countries (not communities or clusters of relevés). Per combination of a country and an alliance, a unique subset of relevés was used, shown in a framed section of the column. The number of relevés per group is given above in italics (in total 412 relevés are included). In the footnotes, given between [ ], the locations (see also fig. 5 in chapter 2), the original names of the syntaxa and references are given. The table contains percentage frequency values and average non-zero cover in % (as superscript). Country D N B F E Calthion palustris Bromus racemosus Caltha palustris Lychnis flos-cuculi Lotus pedunculatus Dactylorhiza majalis Cirsium palustre

22 [1] 100 82 91 41 23 5

12 10 4 8 5 1

53 [3] 100 26 75 58 8 64

Alopecurion pratensis Bromus racemosus Sanguisorba officinalis Fritillaria meleagris Lathyrus pratensis Silaum silaus Alopecurus pratensis

19 [4] 100 63 37 84 5 100

Arrhenatherion elatioris Bromus racemosus Arrhenatherum elatius Dactylis glomerata Trisetum flavescens Tragopogon pratensis Peucedanum carvifolia

25 [5] 100 80 92 68 36 20

Cynosurion cristatus Bromus racemosus Cynosurus cristatus Lolium perenne Bellis perennis Trifolium dubium Agrostis capillaris

14 [6] 100 93 100 57 79 64

Basal communities Molinio-Arrhenatheretea Bromus racemosus Alopecurus pratensis Ranunculus repens Alopecurus geniculatus Lolium perenne Bromus hordeaceus

8 [2] 100 63 75 50 50 88

17 4 15 7 11 8

42 [7] 100 95 98 33 93 62

4 4 3 6 2 3

4 10 3 4 3 5

34 [8] 100 62 91 47 12 15 21 [9] 100 0 0 95 57 67

16 [11] 100 44 63 50 0 25

3 7 3 9 1 3

5 5 13

16 [13] 100 94 94 63 81 13

5 13 8 7 3 5

16 [14] 100 75 88 56 88 38

9 8 11 2 6 20

8 7 10 13 9 4

13 [12] 100 0 0 69 0 100

41 [10] 100 66 98 37 80 49

4 14 30 6 16 5

40 [15] 100 18 98 23 98 15

5 3 4 3

23 [16] 100 70 65 65 74 65

6 23 6 3 2 3

25 13 16 3 14 16

8 5 13 3 12 2

2 10 3 2 3 4

1 2 3 4 5 6 7 8

13 14 15

Senecioni-Brometum racemosi (Calthion palustris), near Bremen, LUTZ 1996 (relevés from several authors, 1951-1996) BC Molinio - Arrhenatheretea, near Bremen, LUTZ 1996 (relevés from several authors, 1951-1996, but most from LUTZ 1996) Communities 1, 2, 10, 11 and eleven additional relevés from the DNVD (SCHAMINÉE et al. 2006) from 1936-1974, assigned to the Ranunculo-Senecionetum caricetosum paniceae Communities 6, 7 Community 4 Community 3 Community 5 Snoekengracht near Leuven, Dijlevallei near Leuven & Bourgoyen near Gent (Flanders), CALLEBAUT et al. (2007) & VLAVEDAT (Flemish Vegetation Databank; VANDENBUSSCHE & HOFFMANN 2001); assigned to the Calthion palustris based on an analysis of the species composition. Colchico-Brometum racemosi (Bromion racemosi, Molinietalia), Femenne & Fagne (southern Wallonia), SOUGNEZ & LIMBOURG 1963; assigned to the Alopecurion pratensis based on an analysis of the species composition. Dijlevallei near Leuven & Bourgoyen near Gent (Flanders), CALLEBAUT et al. (2007) & VLAVEDAT (Flemish Vegetation Databank; VANDENBUSSCHE & HOFFMANN 2001); classified as basal communities based on the absence of character species on the alliance level. Brometo-Senecietum (Bromion racemosi, Molinietalia), Bassin de la Sambre, GEHU 1961 & Cirsio dissecti - Scorzoneretum, Bocage Virois, Basse-Normandie, DE FOUCAULT 1980; assigned to the Calthion palustris based on an analysis of the species composition; in the relevés of DE FOUCAULT 1980 additionally some Junco-Molinion species are present. Arrhenatheretum Colchicetosum (Cynosurion cristati), Bassin de la Sambre, GEHU 1961; assigned to the Alopecurion pratensis based on an analysis of the species composition; species from the Cynosurion were rare and Arrhenatherion species not very frequent; several moisture indicating species were present. Hordeo secalini-Arrhenatheretum elatioris & Galio veri-Trifolietum repentis (both Arrhenatherion), Seine-maritime, FRILEUX et al. 1988 Oenantho peucedanifoliae - Brometum racemosi (Cynosurion cristati), Bocage Virois, Basse-Normandie, DE FOUCAULT 1980 Senecio aquatici-Oenanthetum mediae, Seine-maritime, FRILEUX et al. 1988 & Senecio-Brometum racemosi (Bromion racemosi, Agrostietalia-stoloniferae, Agrostio-Arrhenatheretea elatioris), Boulonnais, DE FOUCAULT 1986; classified as basal communities based on the absence of character species on the alliance level. Unpublished relevés from A. Corporaal, near Cricklade, 1991, all with Fritillaria meleagris & from M. Raman, near Oxford, 2010; assigned to the Alopecurion pratensis based on an analysis of the species composition.

Fig. 14: Detrended Correspondence Analysis (DCA) of a selection of 237 relevés with B. racemosus (74 from fieldwork, 11 from the DNVD and 152 from nearby countries). The communities 8 and 9 are not included. The communities are depicted by the form of the symbols, the countries are marked with a colour (Germany (D): green, Belgium (B): red, France (F): blue, border region of Belgium and France (BF): purple, Netherlands (N): transparent). The numbers between [ ] refer to the footnotes of table 5. Axis 1 and 2 are plotted, eigenvalues are 0.327 and 0.221 respectively. Ellenberg values are plotted as environmental values. In the upper oval relevés from community 1 and 2 are situated. In the lower oval the majority of the relevés from the community Ranunculo-Senecionetum caricetosum paniceae (made in 1936-1974) is placed, no recent Dutch relevés are situated in this part.

Table 6: Average Ellenberg values, species richness and cover of B. racemosus (ordinal scale) of the alliances and basal communities per country (D: Germany, B: Belgium, F: France, BF: Belgium and France, N: Netherlands). Community Moisture Nitrogen pH Continent Light Richness Br_cover Alopecurion_BF 5.2 4.2 4.4 2.9 6.7 40.4 4.5 Alopecurion_E 4.7 4.0 4.6 3.2 7.0 30.7 2.2 Alopecurion_N 6.1 4.5 4.4 3.0 6.7 38.6 3.1 Arrhenatherion_F 4.7 4.5 4.7 3.0 7.0 35.1 4.0 Arrhenatherion_N 4.6 5.0 4.2 2.8 6.8 33.1 4.0 BC_B 5.5 5.0 4.3 2.7 6.8 22.6 2.1 BC_D 5.0 5.0 4.2 2.5 6.8 21.4 4.9 BC_F 5.7 4.6 4.4 2.7 6.9 24.3 4.1 BC_N 5.4 5.0 4.4 2.7 6.7 25.4 4.5 Calthion_B 6.5 4.4 4.3 2.7 6.7 28.6 1.8 Calthion_D 6.1 4.2 4.1 2.7 6.7 29.1 4.0 Calthion_F 6.1 4.0 4.1 2.7 6.8 29.7 3.3 Calthion_N 6.0 4.1 4.3 2.9 6.8 38.4 2.9 Cynosurion_F 5.1 4.2 3.9 2.7 6.8 26.9 5.6 Cynosurion_N 4.6 4.3 4.2 2.7 7.0 29.1 4.9 Average of Total 5.6 4.4 4.3 2.8 6.8 31.2 3.4

5 Discussion 5.1 Abiotic conditions 5.1.1 Hydrology The results indicate that the species has a quite narrow hydrological amplitude. It grows only on sites with moist conditions in winter or limited periods of inundation. The species is rare in permanent wet areas. The species cannot survive long inundation according to GALL (1995 in LUTZ 1996), probably because the wintering plants need oxygen. The species seems to withstand inundation with river water better than inundation with rain and ground water. River water could be favourable because it might contain more oxygen and nutrients. In floodplains, the risk of spring flooding, detrimental for B. racemosus, is higher, as was observed in Cortenoever. GREVILLIOT (1996) observed that long inundations during spring and late mowing for two consecutive years caused a decrease of B. racemosus in a site in the Meuse valley of Northern France. Also flooding in summer could be detrimental according to LUTZ (1996). VAN ECK et al. (2006) found that the lower limit of many plant species in floodplains can be explained better by occasional summer floods rather than by the more frequent winter floods, since plants are generally more tolerant to winter flooding. They found that many species failed to recolonise lower lying areas, even fourteen years after the last summer flood. Probably B. racemosus benefits from summer desiccation, in competition with other species, since it can survive as a seed until moist conditions occur. GOWING et al. (2002) suggest that the winter annual life cycle of B. racemosus could be an adaptation to avoid summer drought and/or to set seed before the first hay-cut. In drier places the species may lack because of sensibility to desiccation in spring, or because of a weaker competitive ability. Outside the dikes the species was found under drier circumstances than inside the dikes, provided the soil was sufficiently loamy and inundation occurred at least every 5-10 years. The hydrologic amplitude is smaller inside the dikes than outside the dikes. But inside the dikes the species had on average a higher cover than outside the dikes. The more constant water regime could enable the species to build up a large population with a high cover. However, if the hydrology changes, the risk of extinction is higher than in floodplains rich in gradients. There the population may shift between higher and lower places within the gradient. The finding that the cover of B. racemosus is higher on somewhat drier places is confirmed by literature. KALUSOVĂ et al. (2009) found that B. racemosus lacks in sites that are waterlogged for a long period each year, and is more abundant on relative dry sites. Also LUTZ (1996) found that the cover of B. racemosus was higher on relative dry meadows than on relative moist. The species was quite abundant on relatively species poor desiccated and nutrient rich grasslands, sometimes even dominating. In some parts of the vegetation however, grasses like Alopecurus pratensis and Bromus hordeaceus were dominant and B. racemosus rare. All these grasslands were currently unfertilised and not mown before the seeds were ripe (LUTZ 1996). If managers respond to desiccation by mowing earlier, this can prevent B. racemosus from seed production.

KLIJN & WITTE (1988) consider B. racemosus as an indicator for lithotrophic seepage in the Pleistocene parts of the Netherlands (areas that are on average higher and drier than the Holocene parts, were most of the fieldwork was done). Lithotrophic seepage was present in 64% of the 76 1-km2 grid cells were B. racemosus occurred in the Pleistocene. In the Holocene part the species did not occur more often on sites with seepage. B. racemosus decreased strongly in the Pleistocene parts of the Netherlands, which may be related to the decline of seepage in many areas. Since moisture could not be measured in the field, mean Ellenberg values were used. This method has been criticised because it is a form of circular reasoning, and since the indicator values are partly based on the subjective field experience of plant ecologist. Further the indicator values are ordinal, so averaging is inappropriate from a strict mathematical perspective. Still (weighted) averaging is often applied by vegetation ecologist, since it works well in practice (DIEKMANN 2003).

5.1.2 Nutrient availability The outcome that B. racemosus prefers relatively nutrient rich sites is partly a consequence of the fact that no relevĂŠs were made in very nutrient rich parcels without B. racemosus. In the CLM research such parcels were included, and B. racemosus was negatively correlated to P and K. Several other grass species have their optimum under richer conditions. Competition with these grasses is not the only explanation for the absence of B. racemosus in heavily fertilised grasslands. Besides, such parcels are mown in April or May, which prevents seed ripening. When nutrient richness of former agricultural grasslands is decreased by mowing without fertilisation, B. racemosus seems to be among the first species that may recolonise the grasslands, provided seeds are present. Its cover is on average higher under moderate nutrient rich conditions, where it can compete with perennial nitrophilous grasses, as indicated by the positive correlation with the cover of these species. VAN DUUREN et al. (1981) studied succession on former cultivated grasslands in the Drentsche Aa. These were bought by a nature organisation and mown without fertilisation, resulting in lower nutrient availability. In two parcels on peaty soils with seepage from deep groundwater B. racemosus was present. During nine years of observation, its frequency and cover within these parcels increased (although its average cover became not much higher than 1%). These parcels developed towards the community Angelico-Cirsietum oleracei. GREVILLIOT et al. (1998) describe that B. racemosus was frequently present in unfertilised Senecioni-Oenanthetum mediae sites (comparable to the Senecioni-Brometum racemosi), but absent from fertilised sites of this association, already under fertilisation levels of 30 to 100 kg N per ha per year. It is however not sure whether the fertilisation is the reason for the absence. The fertilised meadows were mown three to four times per year, during the period May â&#x20AC;&#x201C; September. It is possible that the first cut was to early for B. racemosus. The unfertilised sites were only mown once or twice, and some were grazed afterwards. KALUSOVĂ et al. (2009) found that B. racemosus subsp. racemosus has a unimodal peaking distribution along the available phosphorus gradient. They consider B. racemosus subsp. commutatus as a typical species for less fertile soils.

On the longer term, a low dose of fertiliser every few years is probably most beneficial for B. racemosus; this idea was already stated by WEEDA (1994).

5.1.3 Other factors The species is probably not dependent on the soil texture, but on correlated factors like moisture and nutrient availability. It may be rarer on soils with a coarse texture, because these are more prone to desiccation. In floodplains, rising river water levels have more influence on the groundwater level in sandy soils than in clayey soils, leading to more frequent inundation (AGGENBACH et al. 2007). Heavy clay seems less suitable for B. racemosus, because its low permeability may lead to longer stagnation of water. A good substrate for germination seems to be important for B. racemosus. The PQ dataset of Zuid-Holland indicates that a decreasing CN-ratio and an increasing litter cover could affect B. racemosus negatively. During fieldwork the species was only observed on sites with a good mineralisation, few litter and without a well-developed root math. Further moss cover was negatively correlated to B. racemosus cover. High moss cover may prevent germination. Additionally high moss cover, especially of Calliergonella cuspidata, occurred more often on relative wet sites, that may be suboptimal for B. racemosus.

5.1.4 Differences between subspecies In literature, two different views about the ecological difference between the subspecies of B. racemosus are found. According to the first, there are no or only few differences. In Belgium an analysis of the Flemish vegetation databank revealed no differences between the sites were the two subspecies grow (ZWAENEPOEL 2006). In England both species have almost the same habitat, but subsp. racemosus has a slight preference for wetter places than subsp. commutatus. They are both not strictly bound to meadows, also growing in road verges, waysides and borders of tracks and fields (PRESTON et al. 2002). Also according to WEEDA (1994), the species have comparable site preferences and occur regularly together in the Netherlands. In the second view, the niches only overlap to some extent. BĂ&#x2013;HLING et al. (1998) states that both subspecies hardly occur together in Baden-WĂźrttemberg (Southwest Germany). Subsp. commutatus grows in warm, basiphilous, ruderal sites, cereal fields and grasslands. Subsp. racemosus occurs in wetter, moderately acid grasslands, more often in meadows than in pastures. This ecological difference is in agreement with the Ellenberg values, and with the statement of WEEDA (1994) that subsp. commutatus prefers drier sites and has a stronger preference for calcareous soils than subsp. racemosus. During field work subsp. commutatus was found in only three locations, therefore ecological differences cannot be proven. We got the impression that subsp. racemosus can indeed grow on wetter places, and that subsp. commutatus has its optimum in drier, less acid sites, and can grow in taller vegetation.

5.2 Grassland management B. racemosus has no seedbank, because seeds germinate immediately when they become moist (when stored dry for 5 years, around 1/3 survived, but in nature dry storage is not likely to occur; LUTZ 1996). Appropriate, constant management is necessary to assure that the population survives the two most important bottlenecks of the species life-cycle each year. The most obvious bottleneck is that it should not be mown or grazed before the seeds have ripened. Mowing after 15 June is advisable. LUTZ (1996) also recommended mowing in the end of June or beginning of July. She did not find B. racemosus on meadows mown in May. The species can probably adapt to mowing time, if it is regular and not too early. On some sites that were mown late (Kleiput Dertienmorgenwaard, Texel), the species also flowered later. On a grazed parcel in Polder Achthoven the individual plants differed much more in phenology than in mown sites; some individuals were starting to flower while others had already ripe seeds. Mowing at the right moment can stimulate the seed shedding, especially when the hay lies on the field for some days and is tedded. This could distribute seeds over the parcel. Probably mowing machines also transport seeds between parcels (LUTZ 1996). In the Vijfheerenlanden, species poor parcels are often mown directly after species rich parcels, to stimulate dispersal of plants. Probably because of this management, in this area B. racemosus is often present on former agricultural parcels that were bought around five years ago. Since the seeds are relatively heavy (JENSEN 2004, LUTZ 1996), they probably fail to overcome relatively large distances by air transport. The second bottleneck is the establishment of the seedlings in summer and autumn. The vegetation should not be too dense and long, which can be prevented by aftermath grazing and/or mowing twice a year. Aftermath grazing generates open spots in the sward through trampling, which are a good substrate for germination. MEISEL (1969 in LUTZ 1996) states that B. racemosus has the highest frequency in meadows with aftermath grazing. OBERDORFER (1994 in LUTZ 1996) however writes that B. racemosus is more frequent in meadows that are only mown than in meadows with aftermath grazing. BAKKER et al. (1980) found that germination of seedling of common grassland species in summer and autumn depends much stronger on the management of meadows than germination in spring. When the vegetation is dense in summer hardly any seedlings are found, but immediately after a hay-cut a high number of seedlings emerges. Therefore they conclude that human interference to open the sward is necessary for the late summer germination peak. Mowing and hay removal are important to maintain an open sward. Litter may prevent the establishment of seedlings, since the germinating seeds may desiccate before they reach the soil. Light shortage does not prevent germination, but may prevent the development of seedlings (LUTZ 1996). The advantage of mowing twice is questionable. Since this management practice is mostly applied on nutrient rich parcels, B. racemosus may reach a higher cover partly because of the high nutrient availability. Further mowing twice will decrease nutrient availability faster and may on the longer term lead to a lower B. racemosus cover, but to a higher species richness and herb cover (OOMES 1992). The cover of B. racemosus can change rapidly when the management changes. In the Amerongse Bovenpolder a parcel was described as meadow with much B. racemosus five 50

years ago. But during field work the parcel was grazed and B. racemosus was not found. However, in another less short grazed parcel a few plants were found. GREVILLIOT (1996) states that B. racemosus is normally absent when grassland is being managed as a pasture in the Meuse area of northern France. In the Achterbergse Hooilanden the species was recorded abundant three years ago in some parcels, while it was almost absent during fieldwork. In other parcels is was still extremely abundant. We don’t have an explanation for this. In a parcel near the Linge, D. Kerkhof (personal com.) observed a high cover of B. racemosus of around 25%, but a few years later the species was almost absent. Probably the species suffered because the hay was not removed from the field. In Bremen a parcel had a cover up to 50%, but the next year the species was not found. Three years later a few individuals were found. It was hypothesised that this could be explained by the fact that the parcels was not mown that year (LUTZ 1992). The data about the management may not always be reliable, because not all managers have a good overview about the management of the sites, and farmers deviate sometimes from the planned management.

5.3 Sociology The sociological analysis of the relevés from fieldwork revealed that B. racemosus reaches its optimum in a nutrient rich Molinio-Arrhenatheretea basal community most related to the alliance Alopecurion. Further it prefers Alopecurion associations, drier Calthion associations, wetter Arrhenatherion sites and moist, mown Cynosurion sites. This is a wider range of alliances than suggested in most literature. B. racemosus may be considered as a character species for the Molinio-Arrhenatheretea, with a preference for moderately moist and nutrient rich meadows. SCHAMINÉE et al. (1996) mentioned that B. racemosus is a weak character species of the Alopecurion, that also occurs in the Calthion and some subassociations of the LolioCynosuretum. Its occurrence in the Arrhenatheretum typicum was not mentioned. Several authors that describe floodplain communities (DE FOUCAULT 1986, FRILEUX et al. 1988, SOUGNEZ & LIMBOURG 1963) write that B. racemosus is absent to rare in the wettest flood meadows (like the Ranunculo-Alopecuretum geniculati), and the driest sites (like higher Arrhenatherion sites), and has its optimum in between. These intermediate sites can often be assigned to the Alopecurion, moist variants of the Arrhenatherion or basal communities. The British community MG4 Alopecurus pratensis - Sanguisorba officinalis grassland, as described by RODWELL (1992), is a floodplain meadow comparable to the Dutch Alopecurion pratensis. Although RODWELL (1992) does not mention B. racemosus at all, GOWING et al. (2002) mention that it occurs in this community. BOTTA-DUKÁT et al. (2005) studied lowland wet meadows in Central Europe. They conclude that the Alopecurion pratensis cannot be differentiated from the alliances Agrostion albae, Cnidion venosi, Deschampsion cespitosae and Veronico longifoliae-Lysimachion vulgaris. Therefore they propose to use the name Deschampsion cespitosae Horvatić 1930 for all these traditional alliance; they place this alliance in the Molinietalia. They mention B. racemosus subsp. commutatus as a diagnostic species for mesic, continental Deschampsion meadows in Hungary and southern Slovakia. However, it was only present in three of 26 relevés belonging to this association. Other character species are Medicago lupulina and Daucus carota. These meadows were mainly dominated by Festuca pratensis, and sometimes by Alopecurus 51

pratensis, Deschampsia cespitosa or Poa pratensis. They were situated in between wet alluvial meadows and mesic Arrhenatherion meadows. Floods had a limited to no influence. In Germany, B. racemosus subsp. racemosus is mainly considered as a character species of the Calthion palustris, an alliance which is often present along smaller streams and in polders (LUTZ 1996). BÖHLING (1998) states this for Baden-Württemberg (Southwest Germany). Also OBERDORFER (1957) shares this opinion for southern Germany, and characterises this alliances as wet to alternating wet, fertilised and mown twice a year. He describes an association group of four types of B. racemosus meadows typical for base-rich soils with a low lime content along streams and rivers in the lowlands. Besides typical Calthion species, also species that are considered Alopecurion species in the Netherlands are frequent: Sanguisorba officinalis, Silaum silaus and Alopecurus pratensis. Further Colchicum autumnale is present, like in some French and Wallonian areas (SOUGNEZ & LIMBOURG 1963, GEHU 1961). In Mecklenburg-Vorpommern the species is mentioned as characteristic for the Calthion association Cirsio oleracei - Angelicetum sylvestris (PÄZOLT & JANSEN 2004), an equivalent of the Dutch Angelico-Cirsietum oleracei. This association was also found in the Netherlands (community 10) and in the Flemish area Snoekengracht. Also in the region around Bremen, the species was typical for the Calthion, and more specific the Senecioni-Brometum racemosi (LUTZ 1996). WILLNER et al. (2013) found B. racemosus with a low cover (+) in some Austrian sites of the Cirsietum rivularis Nowiński 1928, belonging to the Calthion. In the Vienna Woods these wet, nutrient rich meadows occur widespread. Characteristic species are Cirsium rivulare, Dactylorhiza majalis and Trollius europaeus. In France and Wallonia vegetations with B. racemosus are often placed in the alliance Bromion racemosi (SOUGNEZ & LIMBOURG 1963, GEHU 1961, DE FOUCAULT 1986), which corresponds to the Calthion palustris. However, in some case the species composition is closer to the Alopecurion pratensis. Besides it is observed in Arrhenatheretalia elatioris vegetations like Arrhenatherion elatioris (FRILEUX ET AL. 1988, DE FOUCAULT 1988) and, mainly in Southwest France and Northwest Spain, in the Lino biennis-Gaudinion fragilis (DE FOUCAULT 1988). This Mediterranean-Atlantic vegetation type is characterised by Linum bienne, Gaudinia fragilis, Crepis vesicaria, Malva moschata, Vulpia bromoides and Oenenathe pimpinelloides. Although B. racemosus occurs regular in the Calthion and can locally be used as a character species, table 5 and fig. 14 demonstrate that its optimum lies in drier communities, where it reaches higher covers. In the end of the 20th century, B. racemosus had disappeared from many Calthion sites around Bremen and was mainly found in species poor basal communities of the MolinioArrhenatheretea (LUTZ 1996), that are similar to the Dutch basal community. It is remarkable that relevés from the Senecioni-Brometum racemosi near Bremen are quite similar to the relevés of the Ranunculo-Senecionetum aquatici caricetosum paniceae from the DNVD. Well developed examples of this subassociation with B. racemosus are all from the period 19361978, mainly in the northeast of the Netherlands; currently B. racemosus seems to be absent from many former sites in this area. B. racemosus sometimes occurs in vegetations on desalinating sites, as found in the DNVD. This vegetation was described by BOER (1955) as the Association of Alopecurus bulbosus and Bromus racemosus (Armerion). This association developed on the former coast of the Zuiderzee in desalination former salt marshes, after this sea had been dammed to create the IJsselmeer in 1932. On some location this association was already present in 1939, elsewhere

it was developed optimally 1945-1950. The species was abundant in a saline-glycophytic phase of the succession, intermediate between Armerieto-Festucetum litoralis and LolioCynosuretum and/or Arrhenatheretum elatioris. Recently (2010) the species was found on a comparable desalinating grassland in Voorne (derived from the DNVD). In Belgium the species is also known from such sites (ZWAENEPOEL 2006). LUTZ (1996) describes that M. Kinder (pers. com.) found a brackish site outside the dikes where B. racemosus occurred with Alopecurus bulbosus, Juncus gerardii, Plantago maritima and Hordeum secalinum; species that also occur in the association described by BOER (1955). In the relevés of Texel some saline species occur with a low frequency and cover; in the past these sites might have had a vegetation more similar to association described by BOER (1955). ZELENÝ & SCHAFFERS (2012) demonstrate that mean Ellenberg values “inherit information about compositional similarity to other vegetation samples”. Therefore analyses that correlate mean Ellenberg values to species composition related aspects should be avoided. For this research however, Ellenberg values were sometimes the only available variables for important environmental factors, like moisture. Therefore Ellenberg values were still used in the Kruskal Wallis test for differences between communities (table 4) and in the RDA. The outcomes of these analyses should thus be interpreted cautious.

5.4 Historical ecology The Dutch landscape has been transformed by humans since centuries, and these changes have probably influenced the abundance of B. racemosus. There is no information about the occurrence of B. racemosus before the 20th century, only about communities in which it occurs. The Dutch rivers were diked between ca. 1000 and 1300 (HAARTSEN & HARTEN 1999, JONGMANS et al. 2013), which resulted in the division of polders and floodplains. POSCHLOD et al. (2009) state that typical Arrhenatheretum stands originate from end 17th / beginning 18th century, when Arrhenatherum elatius arrived in Central Europe (from the south). Before, probably another grassland community occurred at the higher zones of the floodplains. According to POSCHLOD et al. (2009), the Alopecurion pratensis exists since the late Middle Ages in Central Europe, and reached its largest extent in the period 1750-1950. SCHILT & CORPORAAL (2012) describe an optimum for Salland (along the downstream parts of the IJssel) in the 15th - begin 20th century. Water management measures have stimulated the development of the Alopecurion and its association Fritillario-Alopecuretum pratensis. Since the end of the Middle Ages drainage with wind mills decreased the groundwater table and increased mineralisation. To increase fertility, small dikes and hedges were constructed to create watering systems with basins, where sedimentation is stimulated (SCHILT & CORPORAAL 2012). These measures probably led to succession of Junco-Molinion stands to Calthion and Alopecurion, and made the sites more suitable for B. racemosus (A. Corporaal pers. com.). The watering systems existed until the end of 19th century when fertilisers became available. The new management with fertilisation and without watering caused eutrophication, acidification and, since the installation of pumping engines in the begin of the 20th century, desiccation. This resulted in species poor basal communities (SCHILT & CORPORAAL 2012), in which B. racemosus is currently not present.

The Sanguisorbo-Silaetum mainly occurred along the western parts of the river Maas, mostly on places were the river water was let in through openings in the dikes (“rivieroverlaten”) to prevent a high river water level. After closure of these openings the association decreased because of acidification by rain water (SCHAMINÉE et al. 1996). In general the Alopecurion pratensis declined in the 20th century because of high levels of fertilisation, herbicides and deep drainage (SCHAMINÉE et al. 1996). In the begin of the 20th century groynes were constructed along the Dutch rivers (JONGMANS et al. 2013) and the shipping channels were deepened. This made the rivers deeper and flowing faster. This led to the sedimentation of coarser sediments. Further the flood plains grew higher by sedimentation and water tables sunk because of drainage and the incision of the rivers. The amplitudes of water height became larger partly because of these developments (SCHAMINÉE et al. 2001, AGGENBACH et al. 2007). B. racemosus mainly occurs on finer sediments on relatively moist natural levees in less dynamic parts of floodplains, so it could be disadvantaged by these developments. Since the 19th century brickworks use much sediment from the floodplains, creating lakes and wet low-lying parcels. Today still many floodplains are being excavated, also for nature development and to prevent flooding by creating more space for water (HAARTSEN & HARTEN 1999, JONGMANS et al. 2013). This development could further reduce the occurrence of B. racemosus, since the species is mainly restricted to moderately moist floodplains with original relief. An increasing area of Dutch natural grasslands is managed as rough pastures grazed by semiwild cattle breeds, especially in the floodplains. There is a risk that meadows with B. racemosus will be grazed in future. In pastures the species is limited to moderately grazed patches, but most often absent. Along the eastern coast of the former Zuiderzee and downstream parts of the IJssel, B. racemosus was widespread in the past (fig. 4, chapter 1.4). Because of tidal influence and storms, the flow regime of the IJssel and Zwarte Water was more erratic, and more clay was deposited along rivers than nowadays (SCHILT & CORPORAAL 2012). In 1932 the Zuiderzee was dammed to create the IJsselmeer. B. racemosus profited temporarily from the desalination of former salt marshes that became part of the Northeastern Polder, as described by BOER (1955). During the 20th century, B. racemosus almost disappeared from this region. This might be related to some extent to these changing hydrological dynamics. A. Corporaal (pers. com.) suggested that the reduced sedimentation might be disadvantageous for the species. However, the species can be very abundant in polders without sedimentation, and does not tolerate too frequent and erratic flooding. In polders, many flood basins were only subjected to deep drainage since the 1950’, often combined with land consolidation (HAARTSEN & HARTEN 1999). Before drainage, soils were only developed in the uppermost layer, and most parcels could only be used as meadows (JONGMANS et al. 2013). Drainage may be positive for B. racemosus, since it reduces the risk of long water stagnation. But the resulting increase of pasture management was probably negative. Further drained agricultural meadows are often mown before the seeds of B. racemosus are ripe. Drainage also caused subsidence, especially in soils containing peat (JONGMANS et al. 2013); the central parts became lower than the sides and became wetter than optimal for B. racemosus. For nature development in polders, rewetting is often applied. If the water table becomes much higher within a short period, the risk that populations of B. racemosus will disappear is high.

5.5 Comparison with other grassland annuals In the dataset from fieldwork, the cover of B. racemosus was positively correlated to the cover of two other annual species germinating mainly before the winter: Trifolium dubium (r=0.290, P<0.001) and Bromus hordeaceus (r=0.211, P=0.007). KRUIJNE et al. (1967) found that these species have, like B. racemosus, the highest relative abundance in meadows with aftermath grazing. This holds also for Rhinanthus minor and R. angustifolius, but for the latter species the relative abundance in hay meadows without aftermath grazing was almost as high. Like B. racemosus, Rhinanthus species are annual species without seedbank, that need mowing after seed ripening. They germinate in spring, and also prefer a sufficiently open sward. Since they are hemiparasites, they can use resources from perennial species (WEEDA 1988). The presence of Rhinanthus species may lead to an increase of the cover of dicots, because Rhinanthus species mainly parasite on grasses, but also since these species create open spaces in the sward each summer after dying (HELLSTRĂ&#x2013;M et al. 2011). This latter mechanism may to some extent also apply to B. racemosus. Bromus hordeaceus has a comparable life-cycle, but a much broader abiotic range than B. racemosus. It grows also in dry and very nutrient rich grasslands, road verges and ruderal sites (WEEDA 1994). Therefore it is present in most parts of the rural landscape, which makes it less vulnerable to changes in management or abiotic conditions; it can recolonise lost areas from neighbouring road verges or parcels. Vulpia bromoides is also a winter annual grass. It grows together with B. racemosus on the island of Texel, but generally prefers more sunny and sandy sites with an open vegetation (WEEDA 1994). RICE & DYER (2001) did experiments with Bromus tectorum (=Anisantha tectorum), a winter annual that forms a short term persistent seedbank that persists up to five years. They found that seed aging leads to reduced growth and lower competitive strength. Therefore they expect that seed dormancy is a more successful strategy in environments with a higher chance of extreme events. MCDONALD et al. (1996) found that most species typical for flood-meadows of the British Alopecurus pratensis-Sanguisorba officinalis association (including both subspecies of B. racemosus and B. hordeaceus) do not form a persistent seedbank of more than a year. This might indicate that in this habitat growth and competitive strength are more important for population maintenance than the ability to form a seedbank. Still the unpredictability of timing and duration of floods seems to be a risk for species without a seedbank, especially for annuals.

5.6 Nature conservation In none of the visited sites, the management was deliberately optimised for B. racemosus. Many managers encountered during the field research did not even know B. racemosus, and in some cases recognised it wrongly as Bromus hordeaceus.

Since the Netherlands harbour still quit some B. racemosus populations in a wide variety of vegetation types, the country has an international responsibility for the preservation of the species. B. racemosus is an endangered species, but with the right management it will be relatively easy to maintain its population, and to increase its distribution area with former agricultural grasslands. Reintroduction might easily lead to success, if the sward is open enough and seeds or hay containing seeds is spread at the right moment. B. racemosus prefers more nutrient rich conditions than many other rare species, but it is often still present with a low cover in moderate productive, species rich vegetations. Sites that can nowadays be assigned to the BC Bromus racemosus - Alopecurus pratensis could impoverish through regular mowing; on wetter sites succession can lead to the Ranunculo-Senecionetum aquatici, and on drier sites to Alopecurion or Arrhenatherion vegetations (D. Kerkhof pers. com.). Therefore its preservation can be combined with the preservation of other species and vegetation types. Within the Natura 2000 policy, B. racemosus might profit from the protection of the habitat type 6510, Lowland hay meadows (Alopecurus pratensis, Sanguisorba officinalis), that includes Arrhenatherion and Alopecurion meadows. Still it would be worthwhile to optimise the management of some former agricultural areas for B. racemosus, including fertilisation in low doses on the longer term. This could probably be combined with the conservation of certain animal species, such as meadow birds. Productive agricultural grasslands on moist, well buffered clayey soils can be transformed to grasslands of the Alopecurion through mowing management, provided the groundwater table is high in winter, and superficial desiccation occurs in summer (SCHAMINĂ&#x2030;E et al. 2001). Under this management the BC Bromus racemosus - Alopecurus pratensis could develop as a phase in succession. ROSENTHAL (2003) regards B. racemosus as a suitable target species for the restoration of wet grasslands, and states that its most important habitat are extensive, open wet grasslands belonging to the Calthion. He compiled species groups with comparable species, and made a Bromus racemosus group with Veronica longifolia, Rhinanthus angustifolius and Carex cespitosa. In the Netherlands, only Rhinanthus angustifolius often co-occurs with B. racemosus. To test in which period B. racemosus is usually found, the datasets of the Dutch National Vegetation Databank, the CLM and South Holland were inspected. It appeared that B. racemosus is rarely included in relevĂŠs that are made outside the period 1 May â&#x20AC;&#x201C; 10 July. In theory, the species could be discovered also outside the flowering season. In practice it is either to small, or wrongly determined as Bromus hordeaceus; this species has a similar life cycle and phenology, but is found often vegetative. Therefore vegetation inventories in sites with B. racemosus should preferably take place during the period mentioned above. B. racemosus is a good indicator for changing environmental conditions because of its annual life-cycle. It could be used to detect changes in hydrology, nutrient concentrations and the availability of open spots for germination. Further the species may serve to illustrate the importance of knowledge of the specific habitat preference and population ecology of rare plant species, when aiming at the preservation of biodiversity. With stable management B. racemosus maybe preserved without extra effort , but it may disappear within a few years if the grassland management is changed in a way that negatively affects its reproduction.

5.7 Ideas for future research The genetic, ecological and sociological differences between the subspecies (or maybe species) are important questions for future research. Experiments with different management regimes could help to further increase the knowledge about the optimal management, including fertilisation, different species of grazers and grazing intensities, and different numbers of mowings per year. Further the optimal method for reintroduction could be determined by experiments with different seed and/or hay quantities, apply dates and pre-treatments to open the sward. The adaptation of the phenology to management and hydrological regimes is important, and its consequences for reintroduction could be investigated.

6. Conclusion B. racemosus is characteristic for moderately moist, relatively nutrient rich MolinioArrhenatheretea meadows with good mineralisation. In the Netherlands, B. racemosus reaches its optimum in nutrient rich basal communities, Alopecurion associations, drier Calthion vegetations, moist Arrhenatherion sites and moist, mown Cynosurion grasslands. In surrounding countries the species occurs in a similar range of vegetation types. Since it is a winter annual species without a seedbank, seed ripening and seedling establishment should be successful every year to maintain a population. A relatively open sward in summer and autumn is probably essential for establishment. A management with mowing once or twice after 15 June and aftermath grazing is preferred by the species. B. racemosus requires a high groundwater table and/or river flooding during winter for a limited period every one to few years. Under these conditions the species competes successfully with nitrophilous grasses. Its cover is even highest under slightly drier, moderately nutrient rich conditions. The species can colonise former agricultural meadows after a few years of mowing without fertilisation. Since the dispersal abilities of B. racemosus are limited, reintroduction may be considered. On the longer term a low dose of fertiliser every few years is probably needed to maintain a high cover of the species. Outside the dikes, the dynamic environment of the riverine landscape offers mostly suboptimal conditions, but the species can shift along the gradients to persist on the longer time. In homogeneous polders, optimal conditions can lead to a high cover of B. racemosus. However, in case of a change in hydrology or management the population may collapse easier than in floodplains. Threatening developments are the expansion of grazing in nature areas, the increasing amplitude of river discharge, the excavation of flood plains and rewetting projects with a heavy and/or sudden increase of the groundwater table.

Verkorte Nederlandse versie Deze verkorte versie is geschreven voor een artikel in Stratiotes met medewerking van Eddy Weeda, waarvoor dank. De focus ligt op Nederland; er wordt niet ingegaan op de deelvraag over vegetatietypen met Bromus racemosus in omringende landen.

Habitatvoorkeur en plantensociologische positie van Bromus racemosus L. (Trosdravik) in Nederland Inleiding De Trosdravik (Bromus racemosus; Afb. I) is een tegenwoordig zeldzame, eenjarige grassoort van vochtige graslanden. De soort komt in grote delen van Europa voor (HULTÉN & FRIES 1986). In Nederland staat hij te boek als kwetsbaar vanwege de sterke achteruitgang (>50%) in de loop van de 20e eeuw (ODÉ et al. 2006; SPARRIUS et al. 2013; Afb. II). Door zijn gelijkenis met de zeer algemene Bromus hordeaceus (Zachte dravik) wordt B. racemosus vermoedelijk af en toe over het hoofd gezien. Door zijn kale, enigszins glimmende, van groen naar oranje en paars verkleurende aartjes is hij echter goed herkenbaar. De beharing van de bladscheden, die bij Trosdravik stijver is, vormt het meest betrouwbare kenmerk. De Trosdravik is een van de weinige winterannuellen in vochtig grasland (WEEDA 1994). B. racemosus bloeit in mei en begin juni. De vruchten rijpen in juni, vóór de traditionele maaidata, waarna de planten afsterven. De rijpe vruchten ontkiemen zodra ze bevochtigd worden, ongeacht de lichtbeschikbaarheid. Aangezien in het veld doorgaans alle vruchten ontkiemen beschikt de soort niet over een zaadbank (LUTZ 1996; JENSEN 2004). In dit opzicht gedraagt Trosdravik zich als een graan, net als de nauw verwante Dreps (Bromus secalinus). In Nederland kwam B. racemosus vroeger veel voor in hooilanden en hooiweiden in grote delen van het Holoceen, en in sommige beekdalen in het Pleistoceen. De soort groeit op rivierklei, zeeklei, kleiig veen, beekafzettingen en slibrijk zand (WEEDA 1994). Tegenwoordig is B. racemosus grotendeels beperkt tot natuurreservaten, en komt hij alleen in het MiddenNederlandse rivierengebied nog tamelijk algemeen voor. Plantensociologisch hoort B. racemosus thuis in de Klasse der matig voedselrijke graslanden (Molinio-Arrhenatheretea). Tegenwoordig geldt B. racemosus in Nederland als een zwakke kensoort van het Alopecurion pratensis (SCHAMINÉE et al. 1996), maar in Duitsland wordt de Trosdravik doorgaans kenmerkend geacht voor het Calthion palustris (BUCKART et al. 2004). Ook LUTZ (1996), die de sociologie, ecologie en populatiebiologie van B. racemosus in de regio Bremen (Noordwest-Duitsland) onderzocht, beschouwt B. racemosus als een Calthionsoort. Er is nog vrij weinig kennis over de eisen die B. racemosus aan groeiplaats en beheer stelt, en onzekerheid over de syntaxonomische positie van de soort. Daarom heeft Max Simmelink voor zijn MSc thesis onderzoek gedaan naar de Trosdravik, met als hoofdvraag: “Op welke standplaatsen, onder welke beheervormen en in welke vegetatietypen komt Bromus racemosus in Nederland voor?”

Afb. I: Aartjes van Bromus racemosus subsp. racemosus (boven) en subsp. commutatus (onder). De laatste heeft gemiddeld langer aartjes en lemma's. Bij beide ondersoorten komen ook exemplaren met slechts één of enkele aartjes voor (foto's: Theo Muusse).

Volgens Heukels’ Flora kan men twee ondersoorten onderscheiden: B. racemosus subsp. racemosus en subsp. commutatus (VAN DER MEIJDEN 2005; Afb. I). SPALTON (2002) beschouwt deze taxa als aparte soorten (B. racemosus en B. commutatus), net als de meeste buitenlandse flora’s. In dit onderzoek worden de taxa als ondersoorten beschouwd, hoewel wij niet beschikken over gegevens of waarnemingen die een bewijs vormen voor hybridisatie. Subsp. commutatus, die in Nederland zeer zeldzaam voorkomt (WEEDA 1994), is slechts aanwezig in 6% van de vegetatieopnamen uit dit onderzoek. De naamgeving van syntaxa volgt SCHAMINÉE et al. (1996), de nomenclatuur van vaatplanten en mossen respectievelijk VAN DER MEIJDEN (2005) en SIEBEL et al. (2006).

Afb. II: De verspreiding van B. racemosus in Nederland voor en na 1990 (FLORON 2014).

Onderzoeksgebied Voor het onderzoek zijn 28 bekende groeiplaatsen van B. racemosus in Nederland bezocht (Afb. III), alle in beheer bij Staatsbosbeheer, Provinciale Landschappen of de Vereniging Natuurmonumenten. Veel groeiplaatsen zijn gelegen op rivierklei en klei-op-veengronden in het Midden-Nederlandse rivierengebied, in binnendijkse polders tussen de rivieren (zoals de Vijfheerenlanden) maar ook in uiterwaarden. Daarnaast zijn uiterwaarden van de Gelderse IJssel, de Overijsselse Vecht en het Zwarte Water, het beekdal van Het Merkske (in de Baronie van Breda aan de grens met BelgiĂŤ) en polders op Texel bezocht. Op dit Waddeneiland groeit Trosdravik op fijne zand- en zavelbodems van mariene oorsprong, waar het grondwater nog licht brak is (VAN GOETHEM & VAN ROOIJEN 2011).

Afb. III: De onderzoekslocaties.

Methoden: vegetatieopnamen en bodemmonsters In de periode 19 mei tot 5 juli 2013 zijn 144 vegetatieopnamen gemaakt, en zijn 18 opnamen uit de Landelijke Vegetatie Databank (LVD; SCHAMINÉE et al. 2006) uit de periode 20072010 bezocht om extra kopgegevens te verzamelen. Alle opnamen zijn gemaakt volgens de Braun-Blanquet-methode (BRAUN-BLANQUET 1964, SCHAMINÉE et al. 1995), met de door BARKMAN et al. (1964) aangepaste opnameschaal. Steeds zijn opnamen van 9 m2 gemaakt op plekken met B. racemosus, vooral op plaatsen waar de soort een relatief hoge bedekking haalde. Als de soort in een gradiënt voorkwam, zijn vaak ook opnamen gemaakt op plaatsen waar hij niet of nauwelijks aanwezig was. De bodemopbouw en het humusprofiel zijn met een grondboor en broodmes bestudeerd volgens de methode van de Veldgids Humusvormen (VAN DELFT et al. 2006). Per opname is een bodemmonster van de bovenste 20 cm genomen, door met een guts tien steken te nemen en te mengen. De bodemchemische analyses zijn uitgevoerd volgens de richtlijnen van HOUBA et al. (1995). Onderzocht zijn N-totaal en P-totaal (destructie, autoanalyser), Pw (waterextractie, autoanalyser), K en Ca (waterextractie, spectrometer), het aandeel organisch materiaal, pH-H2O en pH-KCl. Grondwaterstanden zijn geschat aan de hand van hydromorfe kenmerken in het bodemprofiel, meetgegevens van www.dinoloket.nl (GDN 2013) en de kaart met grondwatertrappen op www.bodemdata.nl (ALTERRA 2013). Overstromingsfrequenties zijn geschat met behulp van de websites www.live.waterbase.nl (RIJKSWATERSTAAT 2013) en www.ahn.nl (HET WATERSCHAPSHUIS 2013). De vegetatieopnamen zijn ingevoerd in Turboveg (HENNEKENS & SCHAMINÉE 2001), waarbij de gemiddelde Ellenbergwaarden zijn berekend (ELLENBERG et al. 2001), gewogen naar de ordinale bedekking. De opnamen zijn geëxporteerd naar Juice (TICHÝ 2002) waar ze geclassificeerd zijn met behulp van een aangepaste versie van Twinspan (ROLEČEK et al. 2009). Deze classificatie is op een aantal punten handmatig verbeterd. Voor ordinatie en visualisatie is Canoco 5 gebruikt (TER BRAAK & ŠMILAUER 2012). Vanwege de relatief korte ‘length of gradient’ (2.9) is er gekozen voor de Principal Component Analysis (PCA). Voor univariate statistiek is IBM SPSS Statistics 22.0 (IBM CORP. 2013) gebruikt. Ook alle 958 vegetatieopnamen met B. racemosus uit de LVD zijn met behulp van Twinspan geclassificeerd. Daarnaast zijn de vegetatieopnamen die het Centrum voor Landbouwpublicaties en Landbouwdocumentatie (Pudoc, tegenwoordig opgenomen in het CLM) in de periode 1934-1958 heeft gemaakt (KRUIJNE et al. 1967) nader bestudeerd; gegevens over beheer en milieu zijn gerelateerd aan de bedekking en aanwezigheid van B. racemosus. Alleen de 414 opnamen die in mei en juni zijn gemaakt, zijn geanalyseerd; in 87 van deze opnamen is B. racemosus aanwezig. In de opnamen die gemaakt zijn in de overige maanden is B. racemosus slechts aangetroffen in 28 van 1281 opnamen. Omdat het mogelijk is dat de soort vaak over het hoofd is gezien in die periode, zijn deze opnamen niet gebruikt voor de statistische analyses.

Hydrologie De Trosdravik komt voor in vrij vochtige graslanden met gemiddelde Ellenbergwaarden voor vocht in het traject 5-8, met een optimum op relatief droge terreindelen met Ellenbergwaarden rond 6,2. Binnendijks is hij aangetroffen op plaatsen met een GHG van 0-25(-40) cm en een GLG van 60-110 cm onder maaiveld, wat overeenkomt met grondwatertrap II en III (Alterra 2013, GDN 2013). Langdurige overstroming verdraagt B. racemosus niet (GALL 1995, geciteerd naar Lutz 1996). In lagere perceeldelen, waar regenwater maandenlang stagneert, is hij 62

nagenoeg afwezig. Oppervlakkige uitdroging in de zomer kan een concurrentievoordeel betekenen, aangezien de soort als zaad kan overzomeren. In uiterwaarden die zelden overstroomd worden (minder dan eens in de ongeveer tien jaar) lijkt Trosdravik in hetzelfde deel van de vochtgradiënt voor te komen als binnendijks. In uiterwaarden die vaker worden overstroomd, komt hij ook op drogere plaatsen voor met een GHG van 80 cm of lager onder maaiveld (Grondwatertrap VII). Dit is echter alleen waargenomen op plaatsen waar de bodem minimaal ongeveer 40% leem bevat. Inundatie met rivierwater lijkt B. racemosus beter te verdragen dan inundatie met stagnerend regenwater: hij is buitendijks aangetroffen op plaatsen die in de winter enkele weken tot bijna twee maanden worden overstroomd. Inundatie tijdens de bloei leidt echter tot afsterven, ook als deze van korte duur is. Dit is waargenomen in Cortenoever begin juni 2013; op plaatsen waar het water hoger dan zo’n 80 cm boven maaiveld had gestaan, werden geen zaden meer gevormd. De hydrologische amplitude van B. racemosus blijkt dus groter in de uiterwaarden dan in de polders, maar bedekkingen van >12,5% worden vaker binnendijks bereikt. Daar kan de soort onder constant gunstige omstandigheden een grote populatie opbouwen. Daar staat tegenover dat in de meer dynamische uiterwaarden vaak hoogtegradiënten aanwezig zijn, waarlangs de soort kan pendelen. In de vlakke polders is het risico groter dat de Trosdravik geen uitweg vindt bij wijzigingen in de hydrologie.

Bodem De Trosdravik komt vooral voor op zavel en lichte klei; bij het veldwerk is hij aangetroffen op plaatsen met een leempercentage van 20-90% en 4-26% klei. In de dataset van KRUIJNE et al. (1967) uit het midden van de 20e eeuw bevinden zich ook enkele opnamen met Trosdravik op grof zand dan wel zware klei. De zuurgraad van de onderzochte groeiplaatsen bedroeg pH-H2O 5.1-7.9 en pH-KCl 4.0-7.0. Deze waarden blijken niet significant gecorreleerd aan de bedekking van B. racemosus. Bij bijna alle opnamen werd een humustype gevonden dat tot de humusorde mull te rekenen is. Deze humusvorm wordt gekenmerkt door een goede afbraak en een dunne of afwezige ectorganische horizon. Wortelmatten waren soms zwak ontwikkeld, maar doorgaans afwezig. Mull vormt een goed substraat voor de kieming van de vruchten, een sleutelfactor in de overleving van een eenjarige soort, zeker als deze geen zaadbank vormt. Bromus racemosus bereikt zijn hoogste bedekking op relatief voedselrijke standplaatsen. Zijn bedekking is positief gecorreleerd (Spearman’s ρ, n=158) aan de concentraties N-totaal (r=0,211, P=0,008), P-totaal (r=0,289, P<0,001), Pw (r=0,250, P=0,002), K (r=0,162, P=0,041) en de Ellenbergwaarde voor stikstof (r=0,252, P=0,001). Daarnaast zijn de positieve correlatie met de bedekking van de kruidlaag (r=0,358, P<0,001) en de negatieve correlatie met de soortenrijkdom (r=-0,288, P=0,001) te relateren aan zijn voorkeur voor relatief voedselrijke bodems. Bijna geen van de onderzochte percelen was de afgelopen vijf jaar bemest. Mogelijk verliest B. racemosus de concurrentie op recent zwaar bemeste percelen, of is afwezig omdat zulke percelen gemaaid worden voordat B. racemosus zaad zet. In het opnamenbestand van KRUIJNE et al. (1967) dat veel bemeste graslanden omvat, is de bedekking van B. racemosus juist negatief gecorreleerd met het K-getal (r=-0,258, P<0,001, n=395) en het P-getal (r=-0,121, P=0,015, n=406). Uit de opmerkingen over bemesting valt echter op te maken dat zijn bedekking in onbemeste graslanden gemiddeld wat lager was dan in de bemeste graslanden waarin hij voorkwam. Op voormalige landbouwgronden, waaronder ‘raaigrasakkers’, bereikt de Trosdravik soms al na vijf jaar verschralingsbeheer plaatselijk hoge bedekkingen. Het is te verwachten dat verschraling op de lange termijn tot een lagere bedekking van B. racemosus leidt, en lichte bemesting een gunstig effect heeft op de soort.

Graslandbeheer In de opnamen van KRUIJNE et al. (1967) komt B. racemosus vaker voor in hooiweiden (gemaaid en nabeweid) dan zowel in hooilanden als in weilanden (Kruskal Wallis test, Test stat.=77,741, P<0,001, d.f.=5, n=380). De soort werd aangetroffen in 13% van de hooilanden, 48% van de hooiweiden en 4% van de weilanden. Van 200 algemene graslandsoorten was Trosdravik destijds de soort met de sterkste relatieve voorkeur voor hooiweiden (KRUIJNE et al. 1967). Met de opnamen van eigen veldwerk is getest of de soort gemiddeld een hogere bedekking heeft in hooiweiden dan in hooilanden; dat bleek inderdaad het geval te zijn (Mann-Withney U: 1292,0, P=0,007, N=151, Afb. IV). De Trosdravik wordt zelden in weilanden aangetroffen, omdat deze eenjarige soort hier door vraat meestal niet tot bloei kan komen. Bij een lage veebezetting ontstaat er vaak een hoge, ruige vegetatie, waarin B. racemosus vermoedelijk de concurrentie verliest. Tijdens het veldwerk is de soort echter wel aangetroffen in enkele weilanden met een mozaïek van kortgevreten en ruige vegetatie, waarbij de Trosdravik vooral in de randzones voorkomt. Maaibeheer heeft de voorkeur, mits de eerste maaibeurt na de zaadrijping plaatsvindt; de meeste percelen met Trosdravik worden na 15 juni gemaaid, een enkel perceel begin juni. Door het hooien blijft de vegetatie voldoende laag en open, en bij het maaien en afvoeren van het hooi kan zaad verspreid worden. Nabeweiding is gunstig omdat het tot een lagere vegetatie en bodembeschadiging leidt, waardoor B. racemosus gemakkelijker kan kiemen en de concurrentiedruk niet te hoog is. Een constant hooiweidebeheer helpt B. racemosus jaarlijks de kritieke fasen van vruchtzetting en vestiging ieder jaar goed te doorstaan. Aangezien het om een eenjarige soort zonder zaadbank gaat, kan één jaar ongunstig beheer (zoals te vroeg maaien, niet maaien, maaien zonder afvoeren of intensieve begrazing) al funest zijn.

Afb. IV: Bedekking van B. racemosus op de ordinale schaal in hooilanden met en zonder nabeweiding.

Vegetatietypen Classificatie van de tijdens dit onderzoek verzamelde vegetatieopnamen leidt tot onderscheiding van 11 vegetatietypen. Deze zijn ook gescheiden in de PCA als de eerste drie assen worden beschouwd (Afb. V en VI). De eerste as komt overeen met de vochtgradiĂŤnt, de tweede met een gradiĂŤnt in zuurgraad en de derde met voedselrijkdom. De bedekking van B. racemosus is voornamelijk gecorreleerd met as 1 (r=-0,3114) en 3 (r=-0,3262). Alle opnamen kunnen tot de Molinio-Arrhenatheretea worden gerekend. Daarbinnen zijn vier van de vijf verbonden (Calthion palustris, Alopecurion pratensis, Arrhenatherion elatioris, Cynosurion cristati) vertegenwoordigd; alleen in het Junco-Molinion ontbreekt B. racemosus. In de vegetatietabel (Appendix 1) en de synoptische tabel (Appendix 3) staan alle soorten genoemd die de vegetatietypen kenmerken. In Tabel I zijn de medianen van milieuvariabelen per vegetatietype gegeven. Tabel I. De medianen van verschillende milieuvariabelen per vegetatietype (exclusief de typen 9 en 10, waarvan te weinig opnamen waren). De vergelijkbare gemeenschappen 1 en 2 zijn gecombineerd voor een grotere steekproefgrootte. De letters geven significante verschillen weer. Test: Kruskal Wallis test with stepwise stepdown multiple comparisons. N totaal =153, d.f.=7. P (asymptotic significance, 2 sided test) is <0,001 voor alle variabelen. E is een afkorting voor Ellenbergwaarde. Vegetatietype

1+2 (n=8) 3 (n=15)

4 (n=28)

5 (n=42)

6 (n=10)

7 (n=10)

8 (n=8)

11 (n=33) Test Stat.

E_vocht

6,37 c

5,93 b

5,48 a

6,33 c

6,83 cd

7,39 d

6,99 d

6,96 d

100,274

E_stikstof

4,40 a

5,04 b

5,78 c

5,73 c

5,54 c

5,31 b

5,63 c

5,03 b

101,210

Pw (mg/kg)

0,8 a

2,7 b

3,8 b

7,4 d

4,3 bc

3,4 b

6,2 cd

5,8 cd

69,434

P-tot (mmol/kg)

6,2 a

10,0 b

20,1 c

27,4 e

22,0 cd

18,7 c

36,65 e

24,9 d

82,800

K (mg/kg)

15 a

46 b

131 c

253 d

183 cd

107 bc

178 cd

278 d

59,305

pH-KCl

5,83 bc

5,81 c

7,05 d

5,04 b

4,87 ab

7,11 d

7,17 d

4,64 a

72,309

Ca (mg/kg)

282 ab

142 a

654 c

330 b

194 a

609 cd

856 d

291 b

66,289

OM %

9,0 b

4,9 a

10,1 b

14,6 cd

11,8 bc

8,5 b

12,7 cd

15,4 d

79,119

Leem %

63 ab

50 a

54 a

73 bc

67,5 abc

55 a

76 bc

80 c

54,155

Kruidlaag (%)

88,5 a

95 c

92,5 bc

90 ab

95 bc

94 abc

90 ab

26,454

Moslaag (%)

50 c

2,5 b

10 c

43,691

Aantal soorten

32 bc

29 b

32 bc

25 a

37,5 d

41 e

31,5 bc

34 cd

65,706

Afb. V: Principal Component Analysis (PCA) van 158 opnamen gemaakt tijdens het veldwerk. De vegetatietypen zijn afgebeeld als symbolen, zie de legenda. As 1 en 2 zijn weergegeven, de eigenwaarden zijn respectievelijk 0,124 en 0,092. Alleen continue omgevingsvariabelen met een correlatie van meer dan 0,30 met as 1 of 2 zijn afgebeeld. E is een afkorting voor Ellenbergwaarde.

Afb. VI: Zie Afb. V voor de verklaring. As 1 en 3 zijn hier weergegeven, de eigenwaarden zijn respectievelijk 0,124 en 0,077.

• 1 & 2: Fragment van Rhinantho-Orchietum morionis & Rhinantho-Orchietum morionis (Calthion palustris): Deze gemeenschappen zijn enkel aangetroffen op Texel. De associatie is alleen van Nederland bekend, heeft Anacamptis morio als enige kensoort en vertoont verwantschap met het Cynosurion cristati (SCHAMINÉE et al. 1996). Het is een relatief schraal vegetatietype, waarin de Trosdravik slechts lage bedekkingen haalt. • 3: Lolio-Cynosuretum (Cynosurion cristati): Deze gemeenschap is evenals 1 & 2 enkel aangetroffen in hooiweiden op Texel. Ten opzichte van 1 & 2 is deze gemeenschap soortenarmer en productiever. Bromus racemosus bereikt hier relatief hoge bedekkingen. • 4: Arrhenatheretum elatioris typicum (Arrhenatherion elatioris): Trosdravik komt vooral in regelmatig overstroomde uiterwaarden van het Rijnsysteem voor in relatief vochtige vormen van het glanshaverhooiland. Op de hoogste delen ontbreekt hij vaak, behalve als de bodem voldoende lemig is. • 5: RG Bromus racemosus-Alopecurus pratensis-[Molinio-Arrhenatheretea]: Deze hier nieuw beschreven rompgemeenschap is aanwezig in voedselrijke percelen in polders en uiterwaarden, die meestal tot 5-40 jaar geleden landbouwgrond waren. Het zijn soortenarme begroeiingen, waarin Trosdravik plaatselijk een bedekking van zo’n 15-30% behaalt, terwijl ook nitrofiele grassen als Poa trivialis en Lolium perenne prominent aanwezig zijn.

• 6: Fritillario-Alopecuretum pratensis (Alopecurion pratensis): In twee hooilanden langs het Zwarte Water is B. racemosus aangetroffen in de Kievitsbloem-associatie. In de winter worden deze standplaatsen voor korte tijd overstroomd door de rivier. De bodem is kalkarmer en zuurder dan in de overige uiterwaarden. • 7: Sanguisorbo-Silaetum (Alopecurion pratensis): Dit soortenrijke, zeldzame vegetatietype is aangetroffen in de Hengstpolder langs de Nieuwe Merwede, en in de Gansooiense uiterwaard langs de Maas. Bromus racemosus behaalt hogere bedekkingen in drogere delen van deze percelen, die in de winter korter overstromen en in de zomer oppervlakkig uitdrogen. Hier is behalve subsp. racemosus ook subsp. commutatus aangetroffen (WEEDA 1991). • 8: Fragment van Sanguisorbo-Silaetum & Calthion palustris: Deze tamelijk soortenrijke, onbeschreven gemeenschap zonder kensoorten is aangetroffen in uiterwaarden langs de Lek en Nederrijn. De bodem is voedselrijker dan bij de vorige gemeenschap. De soortensamenstelling lijkt het meest op het Sanguisorbo-Silaetum vanwege het samen voorkomen van Alopecurus pratensis, Lathyrus pratensis, Symphytum officinale en Crepis biennis. Daarnaast vertoont de gemeenschap gelijkenis met het Calthion palustris, hetgeen wordt gekenmerkt door de aanwezigheid van Calliergonella cuspidata, Lychnis flos-cuculi and Caltha palustris (SCHAMINÉE et al. 2007). • 9: Een onbeschreven gemeenschap van Rhinanthus angustifolius en Lysimachia vulgaris binnen het Calthion palustris: Dit betreft een soortenrijke vegetatie in een kleiput in de Dertienmorgenwaard langs de Lek, waarin plaatselijk B. racemosus subsp. commutatus voorkomt; subsp. racemosus is niet aangetroffen. • 10: Angelico-Cirsietum oleracei (Calthion palustris): In een zeer soortenrijk hooiland op de flank van het beekdal van Het Merkske is deze associatie gevonden, met slechts plaatselijk enkele exemplaren van B. racemosus. De standplaats wordt gekenmerkt door kalkrijke kwel en een veenlaag van ca. 30 cm bovenop klei. • 11: Ranunculo-Senecionetum aquatici juncetosum articulati (Calthion palustris): Met name in polders in het rivierengebied wordt deze subassociatie aangetroffen. Het betreft vooral hooiweiden op klei- en klei-op-veengronden, die tot 25-50 jaar geleden landbouwgrond waren, maar minder zwaar bemest werden dan graslanden van gemeenschap 5. Sommige opnamen vertonen overeenkomsten met het Lolio-Cynosuretum loletosum uliginosi. Trosdravik behaalt hogere bedekkingen op de vaak hogere perceelsranden, en is zeldzaam tot afwezig in lage delen die ‘s winters lang onder water staan. Daarnaast zijn er in de LVD nog drie andere vegetatietypen aangetroffen, die tijdens het onderzoek niet werden waargenomen: • Bromus racemosus groeit soms in ontziltende graslanden met Alopecurus bulbosus, Juncus gerardii en enkele kweldersoorten. Deze 20 opnamen zou men kunnen toedelen aan het Trifolio frageri-Agrostietum stoloniferae, dat behoort tot het Lolio-Potentillion anserinae. Veel opnamen zijn omstreeks 1940 op kwelders langs de voormalige Zuiderzeekust gemaakt, en als Associatie van Alopecurus bulbosus en Bromus racemosus binnen het Armerion maritimae, onderscheiden door BOER (1955). Hij beschreef een successie vanuit het Armerieto-Festucetum litoralis via bovengenoemde associatie naar het LolioCynosuretum en/of Arrhenatheretum elatioris. Een opname uit 2010 uit het Merrevliet op Voorne is het enige recente voorbeeld. • 17 opnamen uit laagveengebieden vallen binnen de Parvocaricetea (voornamelijk Carici curtae - Agrostietum caricetosum diandrae) met soorten als Pedicularis palustris, Carex diandra, Lathyrus palustris, Eriophorum angustifolium, Hierochloe odorata en Carex elata. Dit indiceert zeer natte, voedselarme omstandigheden, en de bedekking van B. racemosus is dan ook laag. 68

â&#x20AC;˘ Ten slotte zijn er 51 opnamen uit de periode 1936-1978, die tot de subassociatie RanunculoSenecionetum caricetosum paniceae zijn te rekenen, met onder meer Galium uliginosum, Dactylorhiza majalis, Carex panicea, Comarum palustre, Jacobaea aquatica, Deschampsia cespitosa en Luzula multiflora. De Ellenberg-waarden wijzen op condities die wat natter, voedselarmer en zuurder zijn dan in de opnamen van type 11, maar droger en voedselrijker dan in het bovengenoemde type. De meeste opnamen zijn afkomstig uit Drenthe, o.a. uit het Drentsche Aagebied, waar de Trosdravik de laatste jaren niet meer is teruggevonden.

Conclusie Bromus racemosus kan beschouwd worden als een kensoort van de Molinio-Arrhenatheretea, met optimum in een rompgemeenschap op vrij vochtige en tamelijk voedselrijke grond. Daarnaast komt hij voor in beide associaties van het Alopecurion, in relatief droge vormen van Calthion-gemeenschappen, op tamelijk vochtige groeiplaatsen van het Arrhenatherion en in vrij vochtige hooiweiden van het Cynosurion. Als eenjarige soort zonder zaadbank is de Trosdravik kwetsbaar voor wijzigingen in het beheer. Succesvolle voortplanting is gebaat bij een constant beheer met maaien na 15 juni en nabeweiding. Daarnaast lijkt ofwel een relatief hoge, maar niet te hoge grondwaterstand in de winter, ofwel regelmatige overstroming met rivierwater nodig. Onder die voorwaarden kan B. racemosus goed concurreren met nitrofiele grassen, ook op voormalige landbouwgrond. Omdat de soort zo zeldzaam is geworden en zich nauwelijks over grote afstanden verspreidt, is herintroductie het overwegen waard. Aangezien subsp. commutatus slechts op drie locaties is aangetroffen kunnen geen harde uitspraken over de ecologische verschillen tussen de ondersoorten worden gedaan. Het lijkt er op dat subsp. commutatus gemiddeld op drogere plaatsen in hogere vegetaties groeit, al is er ook overlapping in de standplaatsen.

Erweiterte deutsche Zusammenfassung Habitatpräferenz und pflanzensoziologische Position von Bromus racemosus L. in den Niederlanden und umliegenden Ländern Einleitung – Bromus racemosus L. ist eine seltene Grasart der Feuchtwiesen. In den Niederlanden wird sie in zwei Unterarten unterteilt, subsp. racemosus und subsp. Commutatus, die in anderen Ländern manchmal auch als separate Arten angesehen werden. B. racemosus ist heimisch in weiten Teilen Europas. In vielen Ländern ist sie in der Roten Liste aufgenommen, weil sie aufgrund der Intensivierung der Landwirtschaft in den letzten Jahrzehnten stark abgenommen hat. Ihr winterannueller Lebenszyklus ist bemerkenswert für eine Art von Dauergrünland. Die Samen keimen direkt nach der Reifung, sobald sie feucht werden. Dies verhindert die Bildung einer Samenbank. Das Ziel dieser Nachforschung ist es das Wissen über die Habitatpräferenz von Bromus racemosus in den Niederlanden zu vergrößern. Der Einfluss von abiotischen Bedingungen und Grünlandbewirtschaftung auf ihre Abundanz und ihre syntaxonomische Position werden untersucht. Untersuchungsgebiet – Die Studie berücksichtigt 28 Naturschutzgebiete in den Niederlanden, wo Bromus racemosus vorkommt. Die meisten Fundorte befinden sich in den Auen und Poldern der niederländischen Flusslandschaft, ein kleiner Teil in den Poldern der Watteninsel Texel. Der Boden wird dominiert durch minerale Schichten die während des Holozäns durch Flüsse, Gezeiten-Flüsse oder das Meer abgelagert sind. Einige Standorte haben torfige Böden. Daten aus den umliegenden Ländern wurden analysiert und mit den Daten aus den Niederlanden verglichen. Ohne Ausnahme sind alle ausländische Standorte im Tiefland gelegen, in der Nähe eines Flusses oder Baches. Die Standorte befinden sich im Bremer Raum, vier Naturschutzgebieten in Belgien, vier in Nordfrankreich und zwei in Südengland. Material und Methoden – Es wurde die Braun-Blanquet-Methode angewendet. Vegetationsaufnahmen von 9 m2 wurden gemacht auf Standorten mit Bromus racemosus. Der Aufbau der Boden- und Humusprofile wurden mit einem Erdbohrer und einem Brotmesser sichtbar gemacht und weiter untersucht. Für jede Vegetationsaufnahme wurde für eine bodenchemische Analyse eine Bodenprobe aus zehn Teilproben bis 20 cm Tiefe entnommen. Für die Datenanalyse wurden Klassifikation in Juice (mit Twinspan), Ordination mit Canoco und univariate Statistiken angewendet. Vegetationsaufnahmen aus der niederländischen Vegetationsdatenbank und aus den umliegenden Ländern wurden mit Vegetationsaufnahmen aus eigener Feldarbeit verglichen. Die Hydrologie wurde beschrieben mit Hilfe hydromorpher Bodeneigenschaften und Online-Daten über Grundwasserschwankungen und Überschwemmungen. Ergebnisse – In Wiesen ohne Flussüberschwemmungen wurde Bromus racemosus gefunden auf Standorten mit einem mittleren höchsten Grundwasserspiegel höher als 40 cm unter der Bodenoberfläche. Die Art fehlt auf Standorten, die jeden Winter über einen längeren Zeitraum überschwemmt werden. Sie erreicht die höchste Deckung auf relativ trockenen Standorten. Auf Standorten, die mindestens einmal in zehn Jahren von Flusswasser überflutet werden,

wächst die Art auch auf trockeneren Böden mit einem mittleren höchsten Grundwasserspiegel weniger als 80 cm unter der Bodenoberfläche. Bromus racemosus erreicht die höchste Deckung auf relativ nährstoffreichen Standorten, aber kommt nicht in sehr nährstoffreichen Standorten vor. Sie wächst vor allem auf Böden mit Mull als Humusform, gekennzeichnet durch eine gute Mineralisierung. Bromus racemosus erreicht auf Wiesen mit Nachweide eine höhere Deckung als auf Wiesen die nur gemäht werden. Die meisten Wiesen mit B. racemosus werden ab der zweiten Junihälfte gemäht. Nur selten tritt die Art auf in Weiden, wobei ein niedriger Viehbesatz im Frühjahr Voraussetzung scheint zu sein. Die Klassifizierung ergab elf Gruppen von Vegetationsaufnahmen, die in der PCA sichtbar werden. Die erste Achse kann als Feuchtigkeitsgradient interpretiert werden, die zweite als pH-gradient und die dritte als Nährstoffgradient. Alle Vegetationsaufnahmen können innerhalb der Klasse Molinio-Arrhenatheretea eingeordnet werden. Bromus racemosus hat ihre Schwerpunkt in mäßig nährstoffreichen Basalgesellschaften, Alopecurion-Assoziationen, relativ trockenen Calthion- Gesellschaften, feuchten Arrhenatherion-Wiesen und feuchten, gemähten Cynosurion-Gesellschaften. Die meisten Vegetationsaufnahmen aus den umliegenden Ländern ähneln sich den niederländischen Vegetationsaufnahmen und werden den gleichen Verbänden zugeordnet. In den meisten Gesellschaften kann B. racemosus eine hohe Deckung erreichen, aber im Calthion erscheint die Deckung in allen untersuchten Gebieten relativ gering. Diskussion – Bromus racemosus hat eine ziemlich begrenzte hydrologische Amplitude. Sie wächst nur an Standorten mit feuchten Bedingungen im Winter oder kurzzeitiger Überflutung; Überschwemmung zum Beispiel schafft Lebensraum in der Vegetation. Wahrscheinlich profitiert B. racemosus von leichter Sommeraustrocknung in der Konkurrenz mit anderen Arten, da sie als Samen überleben kann bis feuchte Bedingungen auftreten. Die hydrologische Amplitude ist kleiner innerhalb als außerhalb der Deiche. Die Deckung innerhalb der Deiche ist aber durchschnittlich höher als außerhalb. Ein konstanter Wasserhaushalt ermöglicht der Art eine große Population mit hoher Deckung aufzubauen. Wenn die Hydrologie sich ändert, ist das Risiko des Aussterbens im flachen Polder höher als in Auen reich an Höhenunterschieden. Dort kann die Population pendeln. In den Auen besteht aber die Gefahr von Überschwemmung in der Blütezeit, wodurch die Saatproduktion verhindert wird. Das Fehlen von Bromus racemosus in sehr nährstoffreichen Wiesen kommt wahrscheinlich wegen der starken Konkurrenz in Zusammenhang mit der Tatsache, dass diese Wiesen zu früh gemäht werden. Wenn der Nährstoffreichtum von ehemaligen Wirtschaftsgrünland verringert wird durch mähen, kann B. racemosus die Wiesen innerhalb einiger Jahre wiederbesiedeln. Da die Samen sich nicht weit verbreiten, könnte in solchen Wiesen auch Wiedereinführung eine Option sein. Auf längere Sicht ist eine niedrige Dosis von Dünger alle paar Jahre wohl von Vorteil für B. racemosus, da ihre Deckung in nährstoffarmen Standorten niedriger ist. Das Fehlen einer Samenbank und eine geringe Verbreitungsfähigkeit gefährden die Art. Voraussetzung für den erhalt der Population ist erfolgreiche Samenreifung und Etablierung der Keimlinge in jedem Jahr. Eine Bewirtschaftung mit Mahd nach dem 15. Juni und Nachweide eignet sich am besten, da sie die Samenreifung ermöglicht und eine ausreichend offene Krautschicht erhält. Auch in der Literatur wird bestätigt, dass Bromus racemosus in Flusslandschaften sowohl in den feuchtesten Hochwasserwiesen als auch in den trockensten Standorten fehlt. Dazwischen erreicht sie ihren Schwerpunkt im Alopecurion pratensis, in feuchten Arrhenatherion elatioris-Gesellschaften und in Basalgesellschaften. Die britische Assoziation MG4 Alopecurus pratensis - Sanguisorba officinalis grassland ist vergleichbar mit dem

niederlĂ¤ndischen Alopecurion. In Deutschland gilt Bromus racemosus subsp. racemosus vor allem als Charakterart des Calthion palustris. In Frankreich und Wallonien werden vergleichbare Vegetationen oft eingeordnet im Verband Bromion racemosi. B. racemosus hat aber ihren Schwerpunkt in trockeneren Gesellschaften, wo sie eine hĂśhere Deckung erreicht; nur lokal kann sie als Charakterart des Calthion verwendet werden. Im GroĂ&#x;en und Ganzen sollte sie als Charakterart der Molinio-Arrhenatheretea betrachtet werden. Weiterhin kommt Bromus racemosus manchmal vor auf Standorten wo Entsalzung stattfindet.

References •

• •

•

• •

•

• • •

AGGENBACH, C.J.S., GRIJPSTRA, J., JALINK, M.H. & NOOREN, M.J. (2007): Indicatorsoorten voor verdroging, verzuring en eutrofiëring van plantengemeenschappen in uiterwaarden. Deel 10 uit de serie ‘Indicatorsoorten’ (Indicator species for desiccation, acidification and eutrophication of plant communities in floodplains. Part 10 of the series 'Indicator species') [in Dutch]. – Staatsbosbeheer, Driebergen: 194 pp. ALTERRA (2013): Bodemkundig Informatie Systeem (BIS) Nederland (Soil Information System (BIS) of the Netherlands) [in Dutch]. – URL: http://www.bodemdata.nl [accessed 2014-01-05]. ANTHEUNISSE, A.M., LOEB, R., LAMERS, L.P.M. & VERHOEVEN, J.T.A (2006): Regional differences in nutrient limitation in floodplains of selected European rivers: Implications for rehabilitation of characteristic floodplain vegetation. – River Research and Applications 22: 1039-1055. ARNOLDS, E., DIJK, E., DIJKSTRA, A., HEINEMEIJER, H., HOENTJEN, B., DE KONING, R., KOOIJ, A. & OOSTERVELD, P. (1999): Atlas van de Drentse flora (Atlas of the flora of Drenthe) [in Dutch]. – Stichting Werkgroep Florakartering Drenthe / Schuyt, Haarlem: 798 pp. BAKKER, J.P., DEKKER, M. & DE VRIES Y. (1980): The effect of different management practices on a grassland community and the resulting fate of seedlings. – Acta Botanica Neerlandica 29: 469-482. BAL, D., BEIJE, H.M., FELLINGER, M., HAVEMAN, R., VAN OPSTAL, A.J.F.M. & VAN ZADELHOFF, F.J. (2001): Handboek natuurdoeltypen. 2e geheel herziene editie (Handbook nature target types. 2nd completely revised edition) [in Dutch]. – EC-LNV, Wageningen: 832 pp. BARKMAN, J.J., DOING, H. & SEGAL, S. (1964): Kritische Bemerkungen und Vorschläge zur quantitativen Vegetationsanalyse. – Acta Botanica Neerlandica 13: 394-419. BUCKART, M., DIERSCHKE, H., HÖLZEL, N., NOWAK, B. & FARTMANN, T. (2004): Molinio-Arrhenatheretea (E1). Kulturgrasland und verwandte Vegetationstypen. Teil 2: Molinietalia. Synopsis der Pflanzengesellschaften Deutschlands 9. – Floristischsoziologische Arbeitsgemeinschaft, Göttingen: 103 pp. BFN (BUNDESAMT FÜR NATURSCHUTZ) (2013): Datenbank FlorKart aus deutschlandflora.de (NetPhyD). Verbreitung der Farn- und Blütenpflanzen in Deutschland. Aggregiert im Raster der Topographischen Karte 1:25000. Datenstand 2013. – URL: http://www.floraweb.de/webkarten/ [accessed 2014-01-05]. BOBBINK, R. & HETTELINGH, J.P. (Eds.) (2011): Review and revision of empirical critical loads and dose-response relationships. – Coordination Centre for Effects, Rijksinstituut voor Volksgezondheid en Milieu, Bilthoven: 243 pp. BOER, A.C. (1955): Plant succession on former tidal lands in the Northeastern Polder. – Acta Botanica Neerlandica 4: 161-166. BÖHLING, N., GRIESE, J., KLEINSTEUBER, A., LANGE, D., PHILIPPI, G., RÖSCH, M., ROSENBAUER, A., ROSENBAUER, S., SEYBOLD, S., VOGGESBERGER, M., WITSCHEL, M., WÖRZ, A., WOLFF, P. & WULF, F. (1998): Die Farn- und Blütenpflanzen BadenWürttembergs. Bd. 7: Spezieller Teil (Spermatophyta, Unterklassen Alismatidae, Liliidae Teil 1, Commelinidae Teil 1): Butomaceae bis Poaceae. – Ulmer, Stuttgart: 595 pp. 73

• • • • •

•

• • • • • •

• •

BOTTA-DUKÁT, Z., CHYTRÝ, M., HÁJKOVÁ, P. & HAVLOVÁ M. (2005): Vegetation of lowland wet meadows along a climatic continentality gradient in Central Europe. – Preslia, Praha 77: 89-111. BRAUN-BLANQUET, J. (1964): Pflanzensoziologie: grundzüge der vegetationskunde. Zweite, umgearbeitete und vermehrte Auflage. – Springer-Verlag, Wien: 865 pp. CHEFFINGS, C.M. & FARRELL, L. (Eds.) (2005): The Vascular Plant Red Data List for Great Britain. Species Status 7. – Joint Nature Conservation Committee, Peterborough: 116 pp. CHYTRÝ, M., TICHÝ, L., HOLT, J. & BOTTA-DUKÁT, Z. (2002): Determination of diagnostic species with statistical fidelity measures. – Journal of Vegetation Science 13: 79-90. DE FOUCAULT, B. (1980): Les prairies permanentes du bocage Virois (BasseNormandie, France). Typologie phytosociologique et essai de reconstitution des séries évolutives herbagères (Permanent grassland of Bocage Virois (Basse-Normandy, France). Phytosociological typology and attempt to reconstruct the evolutionary grassland series) [in French]. – Doc. Phyt. 5: 5-109. DE FOUCAULT, B. (1986): Quelques données phytosociologiques peu connues sur la végétation du Boulonnais et de la Côte d’Opale (Pas-de-Calais, France) (Some poorly known phytosociological data on the vegetation of Boulonnais and Côte d’Opale (Pasde-Calais, France)) [in French]. – Doc. Phyt. X(II): 93-116. DE FOUCAULT, B. (1988): Synsystematique des prairies mesophiles d’Europe (Ordre des Arrhenatheretalia elatioris) (Synsystematics of mesophilic meadows of Europe (Order Arrhenatheretalia elatioris)) [in French]. – Colloques phytosociologiques 16: 695-707. DENGLER, J., JANSEN, F., GLOCKLER, F., PEET, R.K., DE CACERES, M., CHYTRY, M., EWALD, J., OLDELAND, J., LOPEZ-GONZALEZ, G., FINCKH, M., MUCINA, L., RODWELL, J.S., SCHAMINEE, J.H.J. & SPENCER, N. (2011): The Global Index of Vegetation-Plot Databases (GIVD): a new resource for vegetation science. – Journal Of Vegetation Science 22: 582-597. DIEKMANN, M. (2003): Species indicator values as an important tool in applied plant ecology – a review. – Basic and Applied Ecology 4: 493-506. DROK, W.J.A. (1992): De zone met Grote vossestaart in het rivierengebied (The zone with Meadow foxtail in the river area) [in Dutch]. – Stratiotes 5: 15-21. ELLENBERG, H., WEBER, H.E., DÜLL, R., WIRTH, V. & WERNER, W. (2001): Zeigerwerte von Pflanzen in Mitteleuropa. Scr. Geobot. 18. – Goltze, Göttingen: 262 pp. EUROPEAN ENVIRONMENT AGENCY (2013): EUNIS biodiversity database. – URL: http://eunis.eea.europa.eu/index.jsp [accessed 2013-03-02]. FLORON (2014): Bromus racemosus L. – URL: http://www.verspreidingsatlas.nl/1610 [accessed 2014-07-10]. FRILEUX, P.N., DE FOUCAULT, B. & ROY, J. (1989): Etude de la végétation prairiale de la basse vallée de la Seine entre Rouen et l’estuaire (Seine-Maritime, France) (Study of grassland vegetation of the lower valley of the Seine between Rouen and the estuary (Seine-Maritime, France)) [in French]. – Colloques Phytosociologiques 16: 233-240. GALL, M. (1995): Untersuchungen zur Überstauungstoleranz ausgewählter Grünlandarten. – Diplomarbeit Universität Osnabrück, Osnabrück: 88 pp. GDN (TNO GEOLOGISCHE DIENST NEDERLAND) (2013): DINOloket. – URL: http://www.dinoloket.nl [accessed 2013-11-02].

• •

•

• •

• • • •

• •

GEHU, J.M. (1961): Les groupements végétaux du Bassin de la Sambre française (Plant communities of the French Bassin de la Sambre) [in French]. – Vegetatio 10: 69-372. GOWING, D.J.G., TALLOWIN, J.R.B., DISE, N.B., GOODYEAR, J., DODD, M.E. & LODGE, R.J. (2002): A review of the ecology, hydrology and nutrient dynamics of floodplain meadows in England. No. 446- English Nature Research Reports. – English Nature, Peterborough: 88 pp. GREVILLIOT, F. (1996): Les ecosystemes prairiaux de la plaine alluviale de la Meuse Lorraine: phytosociologie, dynamique et fonctionnement, en relation avec les gradients hydriques et les modifications des pratiques agricoles (The grassland ecosystems of the alluvial plain of the Meuse in Lorraine: phytosociology, dynamics and functioning, in relation to hydrological gradients and changes in agricultural practices) [in French]. – Université de Metz, Metz: 217 pp. GREVILLIOT, F., KREBS, L. & MULLER, S. (1998): Comparative importance and interference of hydrological conditions and soil nutrient gradients in floristic biodiversity in flood meadows. – Biodiversity and Conservation 7: 1495-1520. HAARTSEN, A.J. & HARTEN, J.D.H. (1999): Rivierkleilandschap (Riverclay landscape). – In: BARENDS, S. (Ed.): Het Nederlandse landschap – Een historisch-geografische benadering (The Dutch landscape – A historical-geographical approach) [in Dutch]: 114-131. Uitgeverij Matrijs, Utrecht. HELLSTRÖM, K., BULLOCK, J.M. & PYWELL, R.F. (2011): Testing the generality of hemiparasitic plant effects on mesotrophic grasslands: a multi-site experiment. – Basic and Applied Ecology 12: 235-243. HENNEKENS, S.M. & SCHAMINÉE, J.H.J. (2001): TURBOVEG, a comprehensive database management system for vegetation data. – Journal of Vegetation Science 12: 589-591. HET WATERSCHAPSHUIS (2013): Actueel Hoogtebestand Nederland (Actual height model of the Netherlands) [in Dutch]. – URL: http://www.ahn.nl [accessed 2013-1122]. HOUBA, V.J.G., VAN DER LEE, J.J. & VOVOZANSJY, I. (1995): Soil analysis procedures –Other procedures (soil and plant analysis, part 5B). Syllabus 1995, sixt edition. – Department of Soil Science and Plant Nutrition, Wageningen Agricultural University, Wageningen: 262 pp. HULTEN, E. & FRIES, M. (1986): Atlas of North European vascular plants – North of the tropic of cancer. – Koeltz Scientific Books, Koenigstein: 1172 pp. IBM CORP. (2013): IBM SPSS Statistics for Windows, Version 22.0. – Armonk, New York. JENSEN, K. (2004): Dormancy patterns, germination ecology and seed-bank types of twenty temperate fen grassland species. – Wetlands 24: 152-166. JONGMANS, A.G., VAN DEN BERG, M.W., SONNEVELD, M.P.W., PEEK, G.J.W.C., VAN DEN BERG VAN SAPAROEA, R.M. (2013): Landschappen van Nederland: geologie, bodem en landgebruik (Landscapes of the Netherlands: geology, soil and land use) [in Dutch]. – Wageningen Academic Publishers, Wageningen: 942 pp. JOYCE, C.B. & WADE, P.M. (1998): European wet grasslands: biodiversity, management and restoration. – John Wiley, Chichester: 340 pp. KALUSOVÁ, V., LE DUC, M.G., GILBERT, J., LAWSON, C.S., GOWING, D.J.G. & MARRS, R.H. (2009): Determining the important environmental variables controlling plant species community composition in mesotrophic grasslands in Great Britain. – Applied Vegetation Science 12: 459-471.

• •

•

• • • • • • • • • • •

KLIJN, F. & WITTE, J.P.M. (1999): Eco-hydrology: Groundwater flow and site factors in plant ecology. – Hydrogeology Journal 7: 65-77. KLOOSTERHUIS, J.L., STEUR, G.G.L. & HEIJINK, W. (1986): Bodemkaart van Nederland schaal 1 : 50.000 – toelichting bij het kaartblad Texel (Soil Map of Netherlands scale 1: 50,000 – notes on the map sheet of Texel) [in Dutch]. – Stichting voor Bodem Kartering, Wageningen: 133 pp. KNMI (KONINKLIJK NEDERLANDS METEOROLOGISCH INSTITUUT) (2011): Klimaatatlas – Langjarige gemiddelden 1981-2010 (Climate atlas – Long-term averages 19812010) [in Dutch]. – URL: http://www.klimaatatlas.nl/ [accessed 2014-04-12]. KOOMEN, A.J.M. & MAAS, G.J. (2004): Geomorfologische Kaart Nederland (GKN) – Achtergronddocument bij het landsdekkende digitale bestand (Geomorphological Map of the Netherlands (GKN) – Background document to the country-wide digital file) [in Dutch]. – Alterra, Wageningen: 38 pp. KRUIJNE, A.A., DE VRIES, D.M. & MOOI, H. (1967): Bijdrage tot de oecologie van de Nederlandse graslandplanten (A contribution to the ecology of the Dutch grassland plants) [in Dutch]. – Centrum voor Landbouwpublicaties en Landbouwdocumentatie, Wageningen: 65 pp. LNV (MINISTERIE VAN LANDBOUW, NATUUR EN VOEDSELKWALITEIT) (2004): Besluit Rode lijsten flora en fauna (Decree Red lists of flora and fauna) [in Dutch]. – Staatscourant 218: 21. LOEB, R., KUIJPERS, L., PETERS, C.J.H., LAMERS, L.P.M. & ROELOFS, J.G.M. (2009): Nutrient limitation along eutrophic rivers? Roles of N, P and K input in a species-rich floodplain hay meadow. – Applied Vegetation Science 12: 362-375. LOEB, R., VAN DAALEN, E., LAMERS, L.P.M. & ROELOFS, J.G.M. (2007): How soil characteristics and water quality influence the biogeochemical response to flooding in riverine wetlands. – Biogeochemistry 85: 289-302. LUDWIG, G., MAY, R. & OTTO, C. (2007): Verantwortlichkeit Deutschlands für die weltweite Erhaltung der Farn- und Blütenpflanzen; vorläufige Liste. – Bundesamt für Naturschutz (BfN), Bonn-Bad Godesberg: 32 pp. LUDWIG, G. & SCHNITTLER, M. (Eds.) (1996): Rote Liste gefährdeter Pflanzen Deutschlands. – Schriftenreihe für Vegetationskunde 28: 189-306. LUTZ, S. (1996): Soziologisches, ökologisches und populationsbiologisches Verhalten von Bromus racemosus L. im Bremer Raum – Diplomarbeit. – University of Bremen, Bremen: 83 pp. MCDONALD, A.W., BAKKER, J.P. & VEGELIN, K (1996): Seed bank classification and its importance for the restoration of species-rich flood-meadows. – Journal of Vegetation Science 7: 157-164. OBERDORFER, E. (1957): Süddeutsche Pflanzengesellschaften. Pflanzensoziologie Bd. 10. – Jena, Fischer: 564 pp. OBERDORFER, E. (1994): Pflanzensoziologische Exkursionsflora. – Ulmer, Stuttgart: 1050 pp. ODÉ, B., VAN DER MEIJDEN, R. & BAL, D. (2006): Toelichting op de Rode Lijst vaatplanten (Notes on the Red List vascular plants) [in Dutch]. – Ministerie van LNV, Directie Kennis, Ede: 78 pp. OFFH (OBSERVATOIRE DE LA FAUNE, DE LA FLORE ET DES HABITATS) (2013): La biodiversité en Wallonie; Espèces; Flore (Biodiversity in Wallonia; species; flora) [in French]. – URL: http://biodiversite.wallonie.be/fr/flore.html?IDC=805 [accessed 2013-12-12].

• • •

•

• • • • • • • • •

•

OOMES, M.J.M. (1992): Yield and species density of grasslands during restoration management. – Journal of Vegetation Science 3: 271-274. PASSARGE, H. (1964): Pflanzengesellschaften des nordostdeutschen Flachlandes 1. Pflanzensoziologie 13. – Fischer, Jena: 371 pp. PÄZOLT, J. & JANSEN, F. (2004): Molinio-Arrhenatheretea Tx. 1937 – Wirtschaftsgrünland. – In: BERG, C., DENGLER, J., ABDANK, A. & ISERMANN, M.: Die Pflanzengesellschaften Mecklenburg-Vorpommerns und ihre Gefährdung – Textband: 336-353. Weissdorn-Verlag, Jena. POSCHLOD, P., BAUMANN, B. & KARLIK, P. (2009): Origin and development of grasslands in Central Europe. – In: VEEN, P., JEFFERSON, R., DE SMIDT, J. & VAN DER STRAATEN, J.: Grasslands in Europe of high nature value: 15-25. KNNV Uitgeverij, Zeist. PRESTON, C.D., PEARMAN, D.A. & ARNOLD, H.R. (2002): New atlas of the British and Irish flora: an atlas of the vascular plants of Britain, Ireland, the Isle of Man and the Channel Islands. – Oxford University Press, Oxford: 910 pp. PROJECTTEAM WATWASWAAR.NL (2013): WatWasWaar (What Was Where) [in Dutch]. – URL: www.watwaswaar.nl [accessed 2013-11-12]. RICE, K.J. & DYER, A.R. (2001): Seed aging, delayed germination and reduced competitive ability in Bromus tectorum. – Plant Ecology 155: 237-243. RIJKSWATERSTAAT (2013): Live Waterbase. – URL: http://live.waterbase.nl/ [accessed 2013-12-12]. RODWELL, J.S. (Ed.) (1992): British Plant Communities Volume 3: Grasslands and montane communities. – Cambridge University Press, Cambridge: 540 pp. ROLEČEK, J., TICHÝ, L., ZELENÝ, D. & CHYTRÝ, M. (2009): Modified TWINSPAN classification in which the hierarchy respects cluster heterogeneity. – Journal of Vegetation Science 20: 596-602. ROSENTHAL, G. (2003): Selecting target species to evaluate the success of wet grassland restoration. – Agriculture, Ecosystems and Environment 98: 227-246. SCHAMINÉE, J.H.J., HENNEKENS, S.M. & OZINGA, W.A. (2007): Use of the ecological information system SynBioSys for the analysis of large datasets. – Journal of Vegetation Science 18: 463-470. SCHAMINÉE, J.H.J, JANSEN, A.J.M., AGGENBACH, C.J.S., HAVEMAN, R., SIERDSEMA, H., SMITS, N. & VAN ‘T VEER, R. (2001): Wegen naar natuurdoeltypen 2 : ontwikkelingsreeksen en hun indicatoren voor herstelbeheer en natuurontwikkeling (sporen B en C) (Pathways to nature target types 2: development series and their indicators for restoration management and nature development (tracks B and C)) [in Dutch]. – Expertisecentrum LNV, Wageningen: 364 pp. SCHAMINÉE, J.H.J., JANSSEN, J.A.M., HAVEMAN, R., HENNEKENS, S.M., HEUVELINK, G.B.M., HUISKES, H.P.J. & WEEDA, E.J. (2006): Schatten voor de natuur: achtergronden, inventaris en toepassingen van de Landelijke Vegetatie Databank (Treasures for nature: backgrounds, inventory and applications of the Dutch National Vegetation Database) [in Dutch]. – Alterra, Wageningen: 112 pp. SCHAMINÉE, J.H.J., STORTELDER, A.H.F. & WEEDA, E.J. (1996): De Vegetatie van Nederland, deel 3: Plantengemeenschappen van graslanden, zomen en droge heiden (The vegetation of the Netherlands, volume 3: Plant communities of grasslands, fringes and dry heaths) [in Dutch]. – Opulus Press, Uppsala, Leiden: 356 pp. SCHAMINÉE, J.H.J., STORTELDER, A.H.F., WEEDA, E.J., WESTHOFF, V. & HOMMEL, P.W.F.M. (1995): De vegetatie van Nederland, deel 1: Inleiding tot de plantensociologie – grondslagen, methoden, toepassingen (The vegetation of the

•

• • •

• • • • • • • • • • • •

Netherlands, volume 1: Introduction to phytosociology – principles, methods, applications) [in Dutch]. – Opulus, Uppsala, Leiden: 296 pp. SCHILT, A. & CORPORAAL, A. (2012): Vossenstaartgraslanden in Salland door de eeuwen heen (Foxtail Grasslands in Salland through the ages). – In: SCHAMINÉE, J.H.J. & JANSSEN, J.A.M.: Vegetatiekundige Monografieën 4: Geboeid door het verleden – Beschouwingen over historische ecologie (Phytosociological Monographs 4: Fascinated by the past – Reflections on historical ecology) [in Dutch]: 36-55. KNNV Uitgeverij, Zeist. SIEBEL, H., DURING, H., SOTIAUX, A. & LANDWEHR, J. (2006): Beknopte mosflora van Nederland en België (Concise moss flora of the Netherlands and Belgium) [in Dutch]. – KNNV Uitgeverij, Utrecht: 559 pp. SMITH, P.M. (1973): Observations on some critical Bromegrasses. – Watsonia 9: 319332. SOUGNEZ, N. & LIMBOURG, P. (1963): Les herbages de la Famenne et de la Fagne (Grasslands of the Famenne and Fagne) [in French]. – Centre de cartographie phytosociologique et centre de recherché ecologique et phytosociologique de Gembloux, communicaton 43: 359-412. SPALTON, L.M. (2002): An analysis of the characters of Bromus racemosus L., B. commutatus Schrad. and B. secalinus L. (Poaceae). – Watsonia 24: 193-202. SPARRIUS, L.B., ODÉ., B. & BERINGEN, R. (2013): Basisrapport voor de Rode Lijst vaatplanten 2012 (Basic report for the 2012 Red List vascular plants) [in Dutch]. – FLORON, Nijmegen: 88 pp. TER BRAAK, C.J.F (1988): Canoco – an extension of DECORANA to analyze speciesenvironment relationships. – Vegetatio 75: 159-160. TER BRAAK, C.J.F. & ŠMILAUER, P. (2012): Canoco reference manual and user’s guide: software for ordination (version 5.0). – Microcomputer Power, Ithaca: 496 pp. TICHÝ, L. (2002): JUICE, software for vegetation classification. – Journal of Vegetation Science 13: 451-453. TUTIN, T.G., HEYWOOD, V.H., BURGES, N.A., VALENTINE, D.H., WALTERS, S.M. & D.A. WEBB (Eds.) (2001): Flora Europaea (5 volumes). – Cambridge University Press, Cambridge: 2392 pp. USDA, ARS, NATIONAL GENETIC RESOURCES PROGRAM (2013): Germplasm Resources Information Network (GRIN). – URL: http://www.ars-grin.gov/cgibin/npgs/html/taxon.pl?7855 [accessed 2013-01-31]. VAN DELFT, B., DE WAAL, R., KEMMERS, R., MEKKINK, P. & SEVINK. J. (2006): Field guide humus forms: description and classification of humus forms for ecological applications. – Wageningen, Alterra: 92 pp. VAN DER MAAREL, E. (1979): Transformation of cover-abundance values in phytosociology and its effects on community similarity. – Vegetatio 39: 97-114. VAN DER MEIJDEN, R. (2005): Heukels’ Flora van Nederland, 23e druk (Heukels' Flora of the Netherlands, 23th edition) [in Dutch]. – Noordhoff Uitgevers bv, Groningen/Houten: 685 pp. VAN DUUREN, L., BAKKER, J.P. & FRESCO, L.F.M. (1981): From intensively agricultural practices to hay-making without fertilization: effects on moist grassland communities. – Vegetatio 46/47: 241-258. VAN ECK, W.H.J.M., LENSSEN, J.P.M., VAN DE STEEG, H.M., BLOM, C.W.P.M. & DE KROON, H. (2006): Seasonal dependent effects of flooding on plant species survival and zonation: a comparative study of 10 terrestrial grassland species. – Hydrobiologia 565: 59-69.

•

• • • •

• •

• • • • •

VAN GOETHEM, T.M.W.J. & VAN ROOIJEN, N.M. (2011): Anacamptis morio grasslands (Rhinantho-Orchietum morionis) in the Netherlands – its current ecological position, historical context and future prospects. – Radboud University, Nijmegen: 58 pp. VAN LANDUYT, W., VANHECKE, L. & HOSTE, I. (2006): Rode Lijst van de vaatplanten van Vlaanderen en het Brussels Hoofdstedelijk Gewest (Red List of vascular plants of Flanders and the Brussels Capital Region). – In: VAN LANDUYT, W. (Ed.): Atlas van de flora van Vlaanderen en het Brussels Gewest (Atlas of the flora of Flanders and the Brussels Capital Region) [in Dutch]: 70-80. INBO en Nationale Plantentuin van België, Brussel. VAN VLIET, A. (2014): Jaaroverzicht 2013 – De Natuurkalender; het fenologisch waarnemersnetwerk van Nederland (Annual overview 2013 – The Nature Calander; the phenological observers network of the Netherlands) [in Dutch]. – URL: http://www.natuurkalender.nl/jaaroverzichten/jaaroverzicht_2013.asp# [accessed 2014-04-04]. WAMELINK, W., GOEDHART, P.W., FRISSEL, J.Y., WEGMAN, R.M.A., SLIM, P.A. & VAN DOBBEN, H.F. (2007): Response curves for plant species and vegetation types. – Alterra, Wageningen: 20 pp. WEEDA, E.J. (1987): Nederlandse oecologische flora: wilde planten en hun relaties 2 (Dutch ecological flora: wild plants and their relationships 2) [in Dutch]. – KNNV Uitgeverij, Utrecht & IVN, Amsterdam: 304 pp. WEEDA, E.J. (1988): Nederlandse oecologische flora: wilde planten en hun relaties 3 (Dutch ecological flora: wild plants and their relationships 3) [in Dutch]. – KNNV Uitgeverij, Utrecht & IVN, Amsterdam: 302 pp. WEEDA, E.J. (1991): Het Sanguisorbo-Silaetum Klapp ex Hundt 1964 en verwante graslandvegetaties in het Middennederlandse rivierengebied (The SanguisorboSilaetum Klapp ex Hundt 1964 and related grassland vegetation types in the Middle Dutch river landscape) [in Dutch]. – Stratiotes 3: 3-32. WEEDA, E.J. (1994): Nederlandse Oecologische Flora: wilde planten en hun relaties 5 (Dutch ecological flora: wild plants and their relationships 5) [in Dutch]. – KNNV Uitgeverij & IVN, Amsterdam: 400 pp. WEEDA, E.J., SCHAMINÉE, J.H.J. & VAN DUUREN, L. (2002): Atlas van plantengemeenschappen in Nederlands, deel 2: graslanden, zomen en droge heiden (Atlas of plant communities in the Netherlands, volume 2: grasslands, fringes and dry heaths) [in Dutch]. – KNNV Uitgeverij, Utrecht: 223 pp. WESTHOFF, V. & VAN OOSTEN, M.F. (1991): De plantengroei van de Waddeneilanden (The vegetation of the Wadden islands) [in Dutch]. – KNNV Uitgeverij, Utrecht: 416 pp. WILLNER, W., SAUBERER, N., STAUDINGER, M., GRASS, V., KRAUS, R., MOSER, D., RÖTZER, H. & WRBKA T. (2013): Syntaxonomic revision of the Pannonian grasslands of Austria – Part II: Vienna Woods (Wienerwald). – Tuexenia 33: 421-458. WILSON, D. (1956): Cytogenetic studies in the genus Bromus – Ph.D. thesis. – University of Wales, Aberystwyth: 88 pp. ZELENÝ, D. & SCHAFFERS, A.P. (2012): Too good to be true: pitfalls of usingmean Ellenberg indicator values in vegetation analyses. – Journal of Vegetation Science 23: 419-431. ZWAENEPOEL, A. (2006): Bromus commutatus Schrad. Grote trosdravik & Bromus racemosus L. Trosdravik. – In: VAN LANDUYT, W., HOSTE, I., VANHECKE, L., VAN DEN BREMT, P., VERCRUYSSE, W. & DE BEER, D.: Atlas van de flora van Vlaanderen

en het Brussels Gewest (Atlas of the Flora of Flanders and the Brussels Capital Region) [in Dutch]: 194-199. INBO en Nationale Plantentuin van BelgiĂŤ, Brussel.

Appendices Appendix 1: Relevé table, header data and soil profile description This appendix is presented in a separate excel document.

Appendix 2: Study sites Table 7: The study sites. Relevé names start with two letters that abbreviate the location name, as shown below. Coordinates: Rijksdriehoekscoördinaten. Abbreviations for the nature organisations that manage the areas: sbb: Staatsbosbeheer; ul: Utrechts Landschap; nm: Natuurmonumenten; gl: Geldersch Landschap; zhl: Zuid-Hollands Landschap. Management: the numbers refer to the number of mowings per year, N means ‘no grazing’, G means ‘grazing’ (after mowing, if applicable). Number XYNatura of coordinate coordinate 2000- Mana- Manage- Dike Relevé names Location relevés (x 1000) (x 1000) area ger ment position AB Amerongse Bovenpolder 9 158.000 444.500 66 sbb G1 Outside AH Achterbergse Hooilanden 6 170.100 444.600 sbb G2 Inside BB Brakelse Benedenwaard 5 132.100 425.200 71 sbb G0 Outside BG Bolgerijen 6 136.200 441.200 ul G1/G2 Inside BR De Brommert 6 204.000 510.300 36 sbb N2 Outside CO Cortenoever 12 211.500 457.500 38 sbb G1 Outside DB De Bol 7 121.300 568.400 2 nm G1 Inside (Texel) DH Dijksmanshuizen 4 119.440 564.050 2 nm G1 Inside (Texel) DW Dertienmorgenwaard 3 130.400 442.900 sbb G1 Outside GB Genninger Buitenlanden 5 203.700 508.400 36 sbb N2 Outside GU Gansooiense uiterwaard 5 132.700 413.800 nm G1 Outside HP Hengstpolder 5 115.300 424.100 112 sbb G1 Outside Komgrondenreservaat KB Bommelerwaard 9 140.750 422.000 sbb N2 Inside Kleiput KD Dertienmorgenwaard 3 129.600 443.350 sbb N1 Outside LA Lingeuiterwaard Asperen 4 135.900 433.100 70 sbb N1 Outside LL Lekuiterwaard bij Lopik 1 123.701 441.576 sbb G1 Outside Het Merkske, MK Waterbeemd 2 116.260 381.695 sbb N1 Inside MW Middelwaard 2 132.970 444.320 ul G1 Outside OH Overheicop 4 135.350 438.400 zhl G1/G2 Inside PA Polder Achthoven 8 128.200 441.000 105 zhl G2 Inside RG De Regulieren 12 146.100 437.600 gl G2 Inside SB Sonsbrug 5 137.800 437.200 70 zhl G1 Inside SC Schaayk 6 136.200 434.700 70 zhl G1 Inside SW Scharperswijk 1 131.114 440.012 zhl G1 Inside WA De Waai 9 138.600 438.600 ul G1 Inside WB Waal en Burg 12 116.300 566.500 2 nm G1 Inside (Texel) WL Willige Langerakse Waard 9 120.150 439.200 sbb G1 Outside ZK Zijlkolk 2 204.950 507.690 36 sbb N2 Outside Total: 162

Fig. 15: The location of the sites visited for fieldwork. Abbreviations are explained in table 7. The map was made with Google earth. In fig. 16 the central river area is shown more precise.

Fig. 16: The location of the sites visited for fieldwork within the central river area. Abbreviations are explained in table 7. The map was made with Google earth. Fig. 15 is a map with all locations.

Appendix 3: Synoptic table Synoptic table with percentage frequency values based on the relevés from field work (as presented in the relevé table in appendix 1). All relevés are from the Netherlands and belong to the Molinio-Arrhenatheretea. The relevés were ordered by Twinspan, this classification was improved by hand, all two relevés of cluster 16 and one relevé of cluster 14 were removed from the table (see chapter 3.3 and 4.3.1). Species were assigned to clusters according to fidelity, measured by the phi coefficient (using presence/absence data, within the whole dataset, and the size of all groups was standardised to equal size). Only species with a phi coefficient value of 25.0 or higher for one community or a group of communities were assigned to clusters. Species that had a slightly higher fidelity to small clusters of less than five relevés than to a big cluster were assigned to the big cluster. Diagnostic species are arranged in an order that creates a diagonal structure (first criterion) and further they were ordered by strength of fidelity (second criterion). Companions were arranged by decreasing frequency. Species that are diagnostic for one community are in dark grey, species that are diagnostic for a group of communities are in light grey. Species occurring in less than three relevés are excluded. Layer: h: herb, m: moss. Communities: 1: Fragment of Rhinantho-Orchietum morionis 2: Rhinantho-Orchietum morionis 3: Lolio-Cynosuretum 4: Arrhenatheretum elatioris typicum 5: BC Bromus racemosus-Alopecurus pratensis-[Alopecurion pratensis] 6: Fritillario-Alopecuretum pratensis 7: Sanguisorbo-Silaetum 8: Fragment of Sanguisorbo-Silaetum & Calthion palustris 9: BC Rhinanthus angustifolius-Lysimachia vulgaris-[Calthion palustris] 10: Angelico-Cirsietum oleracei 11: Ranunculo-Senecionetum aquatici juncetosum articulati

Community Number of relevĂŠs

1 4

2 4

3 15

4 28

5 42

6 10

7 10

8 8

68 21

100

9 3

10 2

11 33

100

97 3

Bromus racemosus subsp. racemosus Bromus racemosus subsp. commutatus

h h

100

Hydrocotyle vulgaris Leontodon taraxacoides subsp. taraxacoides Juncus articulatus Juncus conglomeratus Rhytidiadelphus squarrosus

h h h h m

100 100 75 100 100

. 25 25 50 50

. . 7 20 13

. . . . 4

. . . 5 .

. . . . .

. . 20 . .

. . . . .

. . . . 33

. . . . .

. . 3 6 9

Danthonia decumbens Luzula campestris Carex ovalis Ophioglossum vulgatum Hypochoeris radicata Orchis morio Dactylorhiza majalis subsp. praetermissa Carex flacca

h h h h h h h h

50 50 25 25 25 . . .

50 100 50 75 100 100 75 50

7 20 . 7 27 . 7 .

. 7 . . . . . .

. . . . . . . .

. 10 . . 10 . . .

. . . . . . . .

. . . 33 . . . .

. . . . . . . .

. 3 . . . . . .

1-3

Odontites verna subsp. serotina Juncus gerardi Lotus tenuis Sagina procumbens Euphrasia stricta Cynosurus cristatus Agrostis capillaris Rhinanthus minor Carex nigra Triglochin maritima Ranunculus bulbosus Vulpia bromoides Ranunculus sardous Potentilla anserina Plantago major

h h h h h h h h h h h h h h h

75 50 50 50 50 100 75 75 25 . . . 25 . .

. . 25 25 25 75 100 100 50 25 25 50 . . .

47 13 7 7 27 93 67 80 20 7 20 47 33 27 27

. . . . . 11 11 43 . . . . . 7 4

. . . . . 31 14 . 2 . . . . 7 7

. . . . . . . . . . . . . . .

. . . . . . . 30 . . . . . . 10

. . . . . . . 13 . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . 100 . . . . . . .

. . . . . 79 18 . 3 . . . . 3 .

100

Community Number of relevĂŠs

1 4

2 4

3 15

4 28

5 42

6 10

7 10

8 8

9 3

10 2

11 33

3-5

Bromus hordeaceus Lolium perenne Cirsium arvense Hordeum secalinum Elymus repens

h h h h h

. 75 . . .

50 50 . . .

73 100 53 13 60

71 96 61 25 64

62 93 31 10 5

30 30 . . .

. 20 40 . 30

25 75 25 . 13

33 67 33 . 100

. . . . .

12 79 . . 6

Dactylis glomerata Achillea millefolium Trisetum flavescens Tragopogon pratensis subsp. pratensis Equisetum arvense Arrhenatherum elatius Anthriscus sylvestris Senecio jacobaea Peucedanum carvifolia Galium mollugo Heracleum sphondylium Potentilla reptans Crepis biennis Pimpinella major Convolvulus arvensis Eryngium campestre Carex spicata Medicago lupulina

h h h h h h h h h h h h h h h h h h

. . . . . . . . . . . . . . . . . .

. . . 7 . . 7 . . . . . . . . . . .

93 57 71 39 43 79 32 21 21 29 29 39 46 14 7 7 7 7

10 2 7 . 2 14 5 . 2 . . 5 17 . . . . .

10 . . . . 30 . . . . 20 . . . . . . .

. 10 20 . . . . . . 20 . 20 30 10 . . . .

13 . 13 . 13 . . . . . 13 38 63 . . . . .

100 . . . . 100 . . . . . 33 33 . . . . .

. . . . . 50 . . . . . . . . . . . .

9 . 18 . 3 3 . . . . . . 6 . . . . .

Fritillaria meleagris Poa palustris Leontodon autumnalis Ranunculus ficaria Eleocharis palustris Poa pratensis Senecio aquaticus Leontodon hispidus

h h h h h h h h

. . . . . 25 . .

. . 100 . . 25 . .

. . 67 . . 27 . .

. . 4 7 . 36 . .

. . 2 7 5 2 . .

80 40 100 80 50 70 20 20

. . 30 . 30 . . .

. . . 25 13 . . .

. . . . . . . .

. . . 100 . . . .

. . . . 6 3 . .

Community Number of relevĂŠs

1 4

2 4

3 15

4 28

5 42

6 10

7 10

8 8

9 3

10 2

11 33

6-7

Festuca pratensis Sanguisorba officinalis Stellaria palustris Centaurea jacea

h h h h

. . . .

7 . . .

61 . . 29

26 . . 5

100 50 40 50

100 70 60 80

25 . . .

100 . . 33

100 . . .

48 . 6 27

Symphytum officinale Achillea ptarmica Mentha aquatica Silaum silaus

h h h h

. . . .

14 . 7 .

7 . 2 .

30 . 20 .

80 60 60 10

25 . 13 .

100 33 . .

. . 50 .

6 . 9 .

6-10

Rumex crispus Thalictrum flavum Filipendula ulmaria Galium palustre Myosotis scorpioides Carex acutiformis Valeriana officinalis

h h h h h h h

25 . . 50 . . .

. . . . . . .

27 . . . . . .

21 . 4 . . . .

33 2 2 7 2 . 2

20 . 20 60 50 . 10

80 60 90 80 60 90 30

88 25 50 38 . . .

33 67 33 . . . 67

. . 100 100 100 100 100

21 6 18 27 3 3 3

6-11

Carex acuta Carex disticha Calliergonella cuspidata Equisetum palustre Lychnis flos-cuculi Lysimachia nummularia Cardamine pratensis Polygonum amphibium Lathyrus pratensis Phalaris arundinacea Agrostis canina Vicia cracca Caltha palustris Phragmites australis Carex riparia

h h m h h h h h h h h h h h h

. . 100 . 75 . 50 . . . 100 50 . 100 .

. . 50 . 75 . 75 . . . 25 50 . 75 .

. . 13 . 33 . 60 . . . . 53 . 27 .

4 18 7 14 4 29 68 21 46 4 . 39 . 7 .

. 5 17 43 5 2 83 52 21 48 2 12 . 5 .

80 90 80 50 70 70 100 70 60 50 70 30 30 40 .

40 100 100 100 80 90 100 100 100 40 . 80 . 60 20

100 75 88 88 63 25 100 63 75 38 . 63 38 88 13

33 100 100 100 67 67 100 67 100 67 . 100 . 100 .

100 100 100 100 100 50 100 . 100 . . 100 100 100 .

88 48 85 88 67 45 100 61 36 64 39 58 6 15 9

9 10

Community Number of relevĂŠs Lysimachia vulgaris Juncus acutiflorus Cirsium oleraceum Crepis paludosa Primula elatior Stellaria uliginosa Plagiomnium affine

Equisetum fluviatile Pedicularis palustris Lotus pedunculatus Cirsium palustre Glyceria fluitans Rhinanthus angustifolius Myosotis laxa subsp. caespitosa Ranunculus flammula Carex hirta Juncus effusus Prunella vulgaris Companions Poa trivialis Cerastium fontanum subsp. vulgare Rumex acetosa Ranunculus acris Taraxacum officinale s.l. Ranunculus repens Brachythecium rutabulum Trifolium pratense Alopecurus pratensis Trifolium repens Plantago lanceolata Holcus lanatus Agrostis stolonifera 11

1 4 .

2 4 .

3 15 .

4 28 .

5 42 .

6 10 .

7 10 .

8 8 .

9 3 100

10 2 .

11 33 3

h h h h h m

. . . . . .

10 . . . . .

. . . . . .

100 50 100 100 100 100

. . . . . .

h h h h h h h h h h h

. . . 100 . . . . . 100 .

. . . . . 75 . . . 25 25

. . 7 . . 73 . . . 27 7

4 . 14 . . 7 . . 18 . .

5 . 12 19 33 26 . 2 38 19 2

10 . . . 20 . 10 10 . . 30

. . 40 . . . . . . 10 30

. . . . 13 13 25 . . . .

. . 67 . . 100 . . 33 . 33

100 100 100 100 50 50 . . . . .

18 33 64 61 42 76 42 33 55 79 55

h h h h h h m h h h h h h

100 75 50 100 25 75 25 . . 100 50 100 75

75 75 100 75 75 . 75 100 . 50 100 100 .

100 87 87 80 60 40 60 73 . 87 93 100 73

96 89 79 93 89 68 86 100 71 79 96 39 79

100 93 76 74 95 98 83 69 95 76 50 81 81

70 80 100 100 80 80 90 80 100 80 90 70 70

100 90 100 90 80 100 30 60 100 100 90 30 90

100 100 88 75 100 100 75 100 88 75 88 13 88

67 100 100 100 100 . 100 67 100 . 67 100 100

100 100 50 100 50 100 . . 50 . 100 100 .

97 82 97 88 70 94 91 73 85 67 64 100 61

Community Number of relevĂŠs Anthoxanthum odoratum Trifolium dubium Bellis perennis Phleum pratense subsp. pratense Festuca rubra Glechoma hederacea Veronica arvensis Kindbergia praelonga Alopecurus geniculatus Glyceria maxima Geranium dissectum Allium vineale Lotus corniculatus Cerastium glomeratum Veronica serpyllifolia Poa annua Fraxinus excelsior Leucanthemum vulgare x Festulolium loliaceum Urtica dioica Lythrum salicaria Stellaria graminea Galium aparine Juncus bufonius Galium verum Crataegus monogyna Festuca arundinacea Rumex obtusifolius Myosotis discolor Stellaria media Veronica chamaedrys Cardamine hirsuta Epilobium tetragonum Oenanthe fistulosa Drepanocladus aduncus Dactylorhiza incarnata

h h h h h h h m h h h h h h h h h h h h h h h h h h h h h h h h h h m h

1 4 100 75 . . 100 . . 25 . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 4 100 50 25 . 75 . . . . . . . 50 . . . . . . . . . . . . . . . 25 . . . . . . .

3 15 93 80 53 20 40 7 . 7 40 . 7 . . . . 20 . . . . . 7 . 13 . . . . 7 7 . . . . . .

4 28 39 71 68 25 50 39 18 14 . . 25 21 25 . 4 . 4 7 14 7 . 11 4 . 11 11 7 . . . 7 . . . . .

5 42 50 64 48 24 . 26 40 29 33 17 10 2 . 10 12 10 7 2 . 10 2 . 2 . . . . 7 . 5 . 5 5 2 . .

6 10 100 20 80 70 50 40 . 10 20 20 . 30 . . 30 . . 20 20 . . . . . . 10 . . . . . . . 10 . .

7 10 50 20 20 40 50 30 . . . . . 10 20 . . . . 10 . . 10 . 10 20 10 . . . . . . . . . . .

8 8 13 75 75 25 . 38 25 25 . 25 . 13 . . . . . . . . 13 . . . . . 25 . . . . . . . 25 .

9 3 33 . . 100 67 . . 67 . . 67 67 . . . . . . . . . . 33 . . . . . . . . . . . . 33

10 2 100 50 . . 100 . 100 . 50 . . . . 50 . 50 50 . . . 50 . 50 . . . . . . . . . 50 . . .

11 33 100 58 27 33 12 27 15 15 9 15 3 . . 15 . . 9 6 . . 6 3 . . . . . 3 3 . 3 3 . 3 3 6

Appendix 4: Soil profile diagrams and photos • • • • • •

Header data: name relevé (code syntaxon) Humus type; Soil type The humus type is determined by Rein de Waal. The lay-out of the diagrams is taken from his work. The soil type is according to the soil map 1:50.000 (ALTERRA 2013), but not verified. Photos of the depicted soil profiles are presented below the diagrams. There are photos from all soil profiles, but only ¼ is presented here. The soil and humus terminology is in Dutch, since the Dutch terms often don’t have English equivalents. The names of humus profiles and layers follows the Field guide Humus Forms (VAN DELFT et al. 2006). A few names of humus types are new and will be published by Rein de Waal.

Appendix 5: Groundwater table graphs Groundwater graphs order per community. Only near a few relevés there was a groundwater measurement point. Source: www.dinoloket.nl (GDN 2013). x-axis: date y-axis: height of the groundwater table in cm above NAP (a vertical datum in use in large parts of Western Europe). Height of the relevés: www.ahn.nl (Het Waterschapshuis 2013).

Community 1 Dijkmanshuizen Identification: B09B0205 Coordinates: 119287, 564218 Near relevé DH3: 119.405, 564.072, -47 cm, community 1

Community 3 De Bol Identification: B09E0044 Coordinates: 121.480, 568.410 Near relevĂŠ DB1: 121.357, 568.508, -60 cm, community 3

De Bol Identification: B09E0043 Coordinates: 121485, 568190 Near relevĂŠ DB3: 121.287, 568.274, -67 cm, community 3

Waal en Burg Identification: B09B0218 Coordinates: 116.225, 566.800 Near relevĂŠ WBa, 116.233, 566.806, -44 cm, community 3

Community 4 Amerongse Bovenpolder Identification: B39B1382 Coordinates: 159.166, 444.378 Near relevĂŠ ABf: 159.169, 444.390, 636 cm, community 4

100

Amerongse Bovenpolder Identification: B39B1384 Coordinates: 158.527, 444.674 Near relevĂŠ AB1 158.600, 444.677, 530 cm, community 8 and ABc 158.573, 444.663, (530 cm), community 4

Community 5 Lingeuiterwaard Asperen Identification: B38H2325 Coordinates: 135.876, 432.980 Near relevĂŠ LA1: 135.802, 443.044, 60 cm, community 5

101

Community 11 Komgrondenreservaat Bommelerwaard Identification: B45A0420 Coordinates: 140.647, 422.247 Near relevĂŠ KBc: 140.668, 422.181, 125 cm, community 11

Komgrondenreservaat Bommelerwaard Identification: B45A0419 Coordinates: 140.429, 422.443 Near relevĂŠ KB7: 140.559, 422.432, 127 cm, community 11

102

De Regulieren Identification: B39C0292 Coordinates: 145570, 437485 Near relevĂŠ RG5, 145.618, 437.380, 129 cm, community 11

De Regulieren Identification: B39A0420 Coordinates: 146.125, 438.340 Near relevĂŠ RG7, 146.154, 438.316, 148 cm, community 11

103

De Regulieren Identification: B39A0421 Coordinates: 146.530, 437.930 Near relevĂŠ RG8, 146.578, 437.938, 131 cm, community 11

104

Appendix 6: Ordination results PCA with fieldwork data Summary Analysis ‘PCA final 1 Unconstrained-suppl-vars’ Method: PCA with supplementary variables Total variation is 34051.234, supplementary variables account for 61.4% (adjusted explained variation is 48.6%) Summary Table: Statistic Eigenvalues Explained variation (cumulative) Pseudo-canonical correlation (suppl.)

Axis 1 Axis 2 Axis 3 Axis 4 0.1236 0.0923 0.0773 0.0555 12.36 21.59 29.32 34.88 0.974

0.9585

0.9501

0.8856

* Correlation matrix * Resp Ax1 Resp Ax2 Resp Ax3 Resp Ax4 Expl Ax1 Expl Ax2 Expl Ax3 Expl Ax4 Cov_tot Cov_herb 1

1 0 0 0 0.974 -0.0127 -0.0305 -0.0025 -0.2315 -0.3854

1 0 0 -0.0125 0.9585 -0.0197 -0.048 0.1069 0.1344

1 0 -0.0298 -0.0196 0.9501 0.018 0.1126 -0.0461

1 -0.0023 -0.0444 0.0167 0.8856 0.0816 0.0684

Cov_moss Cov_lit Ht_herb pH_H2O pH_KCl OM K Ca Pw N_tot P_tot C:N Br_cov E_Light E_Moist E_pH E_N

0.3505 0.1068 0.1539 -0.272 -0.2776 0.2237 0.0958 -0.0159 -0.0903 0.1423 -0.0004 0.0825 -0.3114 0.0956 0.8547 -0.131 -0.4809

0.0074 -0.2072 0.2641 0.5834 0.5514 -0.1949 0.0864 0.4057 -0.0875 -0.2209 0.1168 -0.0367 -0.228 -0.1192 0.0065 0.7468 0.3797

0.0292 -0.0361 0.008 0.1877 0.1984 -0.4376 -0.572 -0.3349 -0.698 -0.5303 -0.7695 0.3875 -0.3262 0.5248 -0.3465 -0.2116 -0.5831

-0.3235 0.2131 -0.0542 0.0798 0.1226 -0.0019 -0.2824 -0.1641 -0.0915 -0.0712 -0.1841 0.2109 -0.0363 0.2676 0.1641 -0.0753 -0.0275

2 2 3 3 23 3 3 3 3 1 3 1 2 1

105

E_Mow W_Ca W_GLG W_GVG Richness Mowings

1 1 1 1 1

Mow_date Clay Silt Loam VII II IV VI I III Gwt Outside Texel Inside RM No_RM Twinspan Twinspan No_Graze Graze

3 3 3 3 1 1 1 1 1 1 1 2 2 2

-0.8518 0.8133 -0.8424 -0.8393 0.4363 0.2491

-0.0088 0.1502 -0.0158 0.2464 0.3794 -0.2111

-0.1926 -0.3876 0.3362 0.0365 0.2831 -0.1911

-0.0772 0.0105 -0.104 -0.1683 -0.3957 -0.0341

0.1899 -0.034 0.3994 -0.1429 0.2489 -0.2258 -0.401 -0.2386 0.2768 -0.1942 -0.3908 -0.1374 0.2756 -0.2108 -0.406 -0.176 -0.4808 0.4076 0.1286 -0.1131 0.5316 -0.3526 -0.1337 0.0285 -0.218 0.103 -0.0128 0.0668 -0.1283 0.1096 -0.0127 -0.032 0.2167 0.0857 0.0873 0.1053 -0.2041 0.0239 0.0481 0.0049 -0.6275 0.4542 0.1248 -0.1134 -0.1302 0.8284 -0.0192 -0.0503 -0.1198 -0.4662 0.6167 0.3488 0.2166 -0.4997 -0.4217 -0.1989 0.1609 0.012 0.0094 0.0411 -0.1609 -0.012 -0.0094 -0.0411 0.6513 0.0228 -0.3868 -0.3891 0.6769 -0.0278 -0.3736 -0.3959 2 0.1457 0.3245 0.0057 -0.0069 2 -0.1457 -0.3245 -0.0057 0.0069 Resp Ax1 Resp Ax2 Resp Ax3 Resp Ax4

RDA with fieldwork data: Analysis ‘Copy of RDA final 1 Interactive-forward-selection final’, step ‘Global permutation test’ Method: RDA Total variation is 34051.234, explanatory variables account for 38.4% (adjusted explained variation is 31.4%) Summary Table: Statistic Eigenvalues Explained variation (cumulative) Pseudo-canonical correlation Explained fitted variation (cumulative)

Axis 1 Axis 2 Axis 3 Axis 4 0.1117 0.0697 0.0554 0.0328 11.17 18.14 23.69 26.97 0.9573 0.9242 0.8237 0.8172 29.08

47.22

61.64

70.19

106

Permutation Test Results: On First Axis On All Axes

pseudo-F=17.7, P=0.002 pseudo-F=5.5, P=0.002

Analysis ‘Copy of RDA final 1 Interactive-forward-selection final’, step ‘Forward Selection’ Method: RDA Total variation is 34051.234, explanatory variables account for 37.7% (adjusted explained variation is 31.1%) Summary Table: Statistic Eigenvalues Explained variation (cumulative) Pseudo-canonical correlation Explained fitted variation (cumulative)

Axis 1 Axis 2 Axis 3 Axis 4 0.1117 0.0695 0.0553 0.0325 11.17 18.12 23.65 26.9 0.9571 0.9228 0.8226 0.8164 29.63

48.07

62.75

71.36

Table 8: The explanatory value of each included environmental value. Interactive-forwardselection with the Monte Carlo permutations test was used. Layer covers, soil variables and management variables were used as environmental variables. Further the Ellenberg values for moisture and nitrogen were included, since moisture and nitrogen are important factors that had not been measured (N-total was correlated to strong to OM%). Only variables with P adj. (false discovery rate) < 0.05 were included. Forward Selection Results: Explains Contribution pseudoName P P(adj) % % F Moisture av. Ellenberg value (based on ordinal 10.3 26.8 17.9 0.002 0.002 cover) Nitrogen av. Ellenberg value (based on ordinal 6.5 16.9 12.1 0.002 0.00213 cover) pH_H2O 4.5 11.7 8.8 0.002 0.00213 Cover moss layer (%) 2.4 6.2 4.8 0.002 0.00213 Av. height high herb layer (cm) 2.2 5.6 4.4 0.002 0.00246 P-total (mmol/kg) 1.9 5 4 0.002 0.002 Date of first mowing 1.5 3.8 3.1 0.002 0.002 Number of mowings 1.3 3.4 2.8 0.002 0.00213 Clay (%) 1.3 3.3 2.7 0.002 0.00213 Organic Matter (%) 1.2 3 2.5 0.002 0.00229 Pw (mg P/kg) 1.1 2.8 2.4 0.002 0.00246 Cover herb layer (%) 1 2.5 2.2 0.002 0.00229 K (mg/kg) 1 2.5 2.2 0.002 0.00229 Ca (mg/kg) 1 2.7 2.3 0.002 0.00213 Cover litter layer (%) 0.8 2 1.8 0.004 0.00427

107

* Correlation matrix * Cov_herb -0.3875 -0.1343 0.0214 -0.2238 Cov_moss 0.3313 0.0198 0.0734 0.4763 Cov_lit 0.1397 0.1228 -0.2086 -0.1891 Ht_herb 0.1179 -0.1608 0.196 0.2872 pH_H2O -0.3312 -0.2689 0.6143 -0.0928 OM 0.2703 -0.2095 -0.3945 0.0481 K 0.1089 -0.4776 -0.3134 0.3353 Ca -0.0395 -0.5362 0.1436 0.2427 Pw -0.0433 -0.4652 -0.5226 0.051 P_tot 0.0318 -0.6643 -0.3925 0.1436 E_Moist 0.8858 -0.2743 -0.05 -0.1009 E_N -0.4772 -0.7033 -0.1476 -0.0067 Mowings 0.2781 0.0028 -0.3286 0.0266 Mow_date 0.1538 0.3205 0.2598 0.2159 Clay 0.2835 -0.1466 -0.4272 0.2931 Resp Ax1 Resp Ax2 Resp Ax3 Resp Ax4

108

Fig. 17: Redundancy Analysis (RDA) of 158 relevĂŠs from fieldwork, that have been assigned to eleven communities with help of Twinspan, as shown in the relevĂŠ table (appendix 1) and fig. 10. The communities are depicted by symbols, as explained in the legend. Axis 1 and 2 are plotted, eigenvalues are 0.1117 and 0.0697 respectively. All used environmental variables are shown. Abbreviations: E: Ellenberg value, Mow_date: date on which the first mowing is allowed at the earliest. Mowings: number of mowings per year (0-2). Ht_herb: average height of the high herb layer.

Fig. 18: see fig. 17 for explanation. Here the first and the third axis are depicted, eigenvalues are 0.1117 and 0.0554 respectively.

109

DCA with data from fieldwork and surrounding countries: Analysis â&#x20AC;&#x2DC;Unconstrained-suppl-varsâ&#x20AC;&#x2122; Method: DCA with supplementary variables Total variation is 7.30027, supplementary variables account for 16.6% (adjusted explained variation is 12.6%) Summary Table: Statistic Eigenvalues Explained variation (cumulative) Gradient length Pseudo-canonical correlation (suppl.) Correlation matrix Richness 0.0225 Shannon 0.1546

Axis 1 Axis 2 Axis 3 Axis 4 0.3274 0.2209 0.1798 0.145 4.48 7.51 9.97 11.96 3.24 2.72 2.54 2.34 0.9371 0.8477 0.6999

0.4094 0.4406

0.3024 0.2579

-0.0738 0.1473

Evenness

0.2037

0.3099

0.1022

0.2767

Continen Light

0.3858 0.2153

0.3122 0.2057

0.435 -0.0001

0.1084 0.0784

Moisture Nitrogen Ph Salt

-0.8817 0.1693 0.3982 0.0297

-0.1898 -0.7685 -0.0502 0.0715

0.1292 -0.2446 0.312 -0.4075

0.0787 0.2078 0.2679 0.1039

Temperat

-0.0216

-0.0353

0.2439

0.1097

Br_cover

0.558

0.2041 -0.026 -0.3245 -0.1279 Resp Resp Resp Resp Ax1 Ax2 Ax3 Ax4

110

Appendix 7: Vegetation photos This appendix contains several photos from all 11 communities, as described in chapter 4.3.1. Photos: Max Simmelink

111

1&2: Fragment of Rhinantho-Orchietum morionis & Rhinantho-Orchietum morionis (Calthion palustris)

Below and above: DB4 (De Bol), community 2

112

Below and above: WB 6 (Waal en Burg, Westerkolk), community 2

113

Below and above: DH1 (Dijkmanshuizen), community 1

114

3: Lolio-Cynosuretum (Cynosurion cristati)

Below and above: WB 7 (Waal en Burg, Westerkolk), community 3

115

Below and above: DB1 (De Bol), community 3

116

Below and above: WB8 (Waal en Burg), community 3

117

4: Arrhenatheretum elatioris typicum (Arrhenatherion elatioris)

Below and above: ABb (Amerongse Bovenpolder), community 4

118

Below and above: CO1 (Cortenoever), community 4

119

Below and above: BB1 (Brakelse Benedenwaard), community 4, subsp. commutatus, grazed

120

Below and above: KB1 (Komgrondenreservaat Bommelerwaard, donk het Delwijse Loo), community 4, subsp. commutatus

121

Below and above: GB3 (Genninger Buitenlanden), community 4

122

5 BC Bromus racemosus-Alopecurus pratensis-[Molinio-Arrhenatheretea]

Below and above: AH2 (Achterbergse Hooilanden), community 5

123

Below and above: DW3 (Dertienmorgenwaard), community 5

124

Below and above: KB3 (Komgrondenreservaat Bommelerwaard), community 5

125

Below and above: PA1 (Polder Achthoven), community 5

126

Below and above: PA7 (Polder Achthoven), community 5, grazed

127

6: Fritillario-Alopecuretum pratensis (Alopecurion pratensis)

Below and above: GB4 (Genninger Buitenlanden), community 6

128

Below and above: BR5 (De Brommert), community 6

129

Below and above: BR2 (De Brommert), community 6

130

7: Sanguisorbo-Silaetum (Alopecurion pratensis)

Below and above: GU5 (Gansooiense Uiterwaard), community 7

131

Below and above: HP4 (Hengstpolder), community 7

132

Below and above: HP2 (Hengstpolder), community 7

133

8: Fragment of Sanguisorbo-Silaetum & Calthion palustris

Below and above: AB1 (Amerongse Bovenpolder), community 8

134

Below and above: WL5 (Willige Langerakse Waard), community 8

135

9: BC Rhinanthus angustifolius-Lysimachia vulgaris-[Calthion palustris]

Below and above: KD2 (Kleiput Dertienmorgenwaard), community 9, subsp. commutatus

136

Below and above: KD1 (Kleiput Dertienmorgenwaard), community 9, subsp. commutatus

137

10: Angelico-Cirsietum oleracei (Calthion palustris)

Below and above: MK1 (Het Merkske, Kromme Hoek, Waterbeemd), community 10

138

Below and above: MK2 (Het Merkske, Kromme Hoek, Waterbeemd), community 10

139

11: Ranunculo-Senecionetum aquatici juncetosum articulati (Calthion palustris)

Below and above: LA4 (Lingeuiterwaard Asperen), community 11

140

Below and above: RG5 (De Regulieren), community 11

141

Below and above: RG8 (De Regulieren), community 11

142

Below and above: SB4 (Sonsbrug), community 11

143

Below and above: SC6 (Schaayk), community 11

144

145

146