Analysis of soil seed bank patterns in an oxbow system of a disconnected floodplain

Ecological Engineering 100 (2017) 46–55

Contents lists available at ScienceDirect

Ecological Engineering journal homepage: www.elsevier.com/locate/ecoleng

Analysis of soil seed bank patterns in an oxbow system of a disconnected floodplain André Schwab ∗ , Kathrin Kiehl Vegetation Ecology & Botany, University of Applied Sciences Osnabrück, Oldenburger Landstr. 24, 49090 Osnabrück, Germany

a r t i c l e

i n f o

Article history: Received 18 September 2015 Received in revised form 16 November 2016 Accepted 20 November 2016 Keywords: Soil seed bank Floodplain Fluctuating water Dynamic Ecological restoration Riparian habitats

a b s t r a c t Soil seed banks have a high potential for vegetation re-establishment in restoration projects. We studied the soil seed bank in an oxbow system of a disconnected floodplain of the Danube River in Southern Germany. The aim of the study was to analyze if floodplain target species were still present in the seed bank after more than 150 years of embankment and disconnection from fluvial dynamics. In this context we investigated seed density, seed bank species richness and species composition in four broad habitat types with and without water-level fluctuations during the time of embankment (permanent water, fluctuating water, reed bed, hardwood floodplain forest). In addition, the similarity between seed bank and above-ground vegetation in these habitat types was studied in order to predict the success of future restoration measures. In total, 124 vascular plant species were determined in the seed bank samples. More than 50 % (66 species) were target species typical for floodplain habitats and 26 of these target species were lost or very rare in the above-ground vegetation. The four habitat types differed significantly in mean seed density and mean species richness. Mean species richness and the number of target species in the seed bank as well as the mean seed density were greatest in the habitats with fluctuating water level whereas mean seed density was much lower in the parts with more or less stable conditions like permanently standing water and hardwood floodplain forest. Sørensen similarity between seed bank and above-ground vegetation was very low in habitats with more or less stable water levels and desirable floodplain target species were very rare or completely absent. Our results indicate that the soil seed bank can be an important seed reservoir for the ecological restoration of floodplain plant communities especially for habitats with unstable environmental conditions during the period of disconnection. Restoration of water level dynamics is important to maintain the seed bank of populations of floodplain target species. © 2016 Elsevier B.V. All rights reserved.

1. Introduction The ecological restoration of floodplains is a challenge on almost every river affected by humans in the last centuries (Jensen et al., 2006; Woolsey et al., 2007; Feld et al., 2011). Restoration targets aim to regain a functional floodplain as well as to preserve and enhance floodplain biodiversity (Mant et al., 2012; Doi et al., 2013; Mitsch and Gosselink, 2015). In the European Union, the Water Framework Directive and the Habitats Directive make river and floodplain restoration obligatory to reach a predefined “favorable ecological status” (European Union, 2000, 2004).

∗ Corresponding author. E-mail addresses: [email protected] (A. Schwab), [email protected] (K. Kiehl). http://dx.doi.org/10.1016/j.ecoleng.2016.11.068 0925-8574/© 2016 Elsevier B.V. All rights reserved.

Due to small scale relief changes and their influence on hydrological regime natural floodplains of large rivers are characterized by different habitat types, which are important for many specialized plant and animal species (Ward et al., 1999, Mitsch and Gosselink, 2015). Plant communities in these habitats vary according to the gradient of natural river dynamics caused by fluctuating water levels and mechanical disturbance (Tockner et al., 1999; Ward et al., 2002). Today, the connection between rivers and floodplains is lacking in many anthropogenically transformed rivers and floodplain communities are degraded with strong negative effects on biodiversity (Gumiero et al., 2013). To stop this development fluvial dynamics can be restored by re-connecting rivers with their floodplains (Ward et al., 1999; Amoros, 2001; Bunn and Arthington, 2002; Nilsson and Svedmark, 2002). In such restoration projects, however, it is often not clear if restoration success depends only on the applied restoration measures or also on the initial conditions

A. Schwab, K. Kiehl / Ecological Engineering 100 (2017) 46–55

concerning abiotic factors, pre-restoration plant communities and soil seed banks. Soil seed banks are repositories from past and present vegetation and have great potential in restoration projects to enable the re-establishment of plant communities (e.g. Grime, 1979; Bakker et al., 1996; Baskin and Baskin, 1998; Liu et al., 2009; Lu et al., 2010). While Goodson et al. (2001) pointed out the general lack of knowledge concerning the role of soil seed banks in riparian systems, some recent studies have analyzed the seed banks of riparian habitats in relation to the burial depth (O’Donnell et al., 2014), under the influence of non-native tree species (Skowronek et al., 2014) and according to flow regulation (Greet et al., 2013). In France, Abernethy and Willby (1999) investigated the composition and density of propagule banks in aquatic habitats which changed along a gradient of disturbance. These authors found that species richness and seed density of the seed bank decreased with decreasing disturbance and that seed bank similarity to the above-ground vegetation is higher in regularly disturbed than in stable habitats. Little is known, however, if and how long floodplain species, which are adapted to regular disturbance, can survive in the soil seed bank of disconnected floodplains without regular flooding. To our knowledge no study exists, which compares soil seed banks under fluctuating and stable conditions for aquatic as well as for terrestrial and semi-terrestrial habitats in a disconnected floodplain. The river Danube is one of the largest rivers in Europe, with only a few natural or near-natural floodplain areas. In Germany, a large scale restoration project planned to re-connect the old oxbows of a Danube floodplain forest with the river and to enhance fluvial dynamics more than 150 years after embankment (Stammel et al., 2012). Before restoration, this area included habitat types with rather stable conditions, such as terrestrial hardwood floodplain forest or old deep backwaters with permanently standing water but also habitat types with fluctuating environmental conditions such as temporary backwaters with fluctuating water levels or reed beds (Margraf, 2004; Lang et al., 2011). The aim of our study was to investigate the soil seed bank of these major habitat types to answer the following questions:

47

1. Are target species typical for floodplains still present in the seed bank more than 150 years after embankment and disconnection from fluvial dynamics? 2. Are there differences in seed bank species richness and species composition between habitat types differing in initial conditions like water regime (area, depth, duration and timing) and due to influence of the construction works? 3. How similar is the species composition of the soil seed bank and the above-ground vegetation in different habitat types of a disconnected floodplain? 2. Materials and methods 2.1. Study area The study area, a floodplain forest of 1200 ha is located on the Danube River between Neuburg (48◦ 43 49.0” N, 11◦ 11 19.6” E) and Ingolstadt (48◦ 45 59.5” N, 11◦ 25 32.7“E), Southern Germany (Fig. 1) at elevations between 368 m and 378 m a.s.l. Average annual rainfall is 700–750 mm and mean annual temperature 8.8 ◦ C (1990–2012, Deutscher Wetterdienst, Karlshuld). Important changes of this part of the river Danube started around 1820 with river straightening, embankment and incising of the river bed. Around 1970 two hydropower dams in Bergheim and Ingolstadt were constructed (Stammel et al., 2012). Due to the altered connection between river and floodplain, which resulted in altered hydrology, large areas of the floodplain with former channels and oxbows degenerated to terrestrial forest (Margraf, 2004; Lang et al., 2011). Only a few backwaters with more or less stable water level by 1–2.5 m depth remained. Areas with frequently fluctuating water levels were rare in the floodplain and water level changes were mostly due to ground- or rainwater influence. Recent larger floods of the Danube River (HQ 10), which covered a bigger part of the study area, only occurred in 1999 and 2005. Restoration planning started in the early 1990s aiming to restore hydrological dynamics within the floodplain (Stammel et al., 2012). We focus our study area on these parts, which are affected by the two main measures (Fig. 1): A new floodplain water course with dynamic

Fig. 1. Study area and restoration measures; sector: division of the study area depending on the water regime before start of the restoration (Table 1, Lang et al., 2013).

48

A. Schwab, K. Kiehl / Ecological Engineering 100 (2017) 46–55

Table 1 Number of study sites in the five sectors and environmental parameters defining the pre-restoration state of the sectors in the study area (Fig. 1).Hydrological information from Schlegel GmbH (2000); for more details see Stammel et al. (2012) and Lang et al. (2013). Sector

State of water body (before restoration) Ground water level (before restoration) Impact of construction work Number of Sites

Total

2

3

4

5

6

no water

temporary

temporary lotic; lentic

temporary lentic

lentic (big backwater)

low

medium

medium

medium

high

high

medium

none

none

none

13

15

11

12

14

water amount of 0.5–5 m3 /s and ecological flooding events (up to 25 m3 /s) should bring back the missing fluctuating water levels in the future (Stammel et al., 2012). For the scientific monitoring the project area was divided in sectors (Fig. 1, Lang et al., 2013) in order to take into account the water regime before start of the restoration (Table 1; hydrological information from Schlegel GmbH (2000); for details see Stammel et al., 2012). After construction works for the new water course (08/2009–03/2010) with top soil removal, but before the reconnection of the oxbows with the river (06/2010), analyses of the above-ground vegetation and the soil seed bank were carried out in the sectors 2–6 of the floodplain along 16 transects, which had been installed perpendicular to the new water course to cover different habitat types in the floodplain zonation (generally three per sector, four in sector 2). Each transect extended over the floodplain parts that would be directly affected by the new water course and the ecological flooding as predicted by hydrological models (transect length: 20–109 m, Lang et al., 2013).

2.2. Field methods For soil seed bank analysis the 16 transects of the vegetation monitoring (see 2.1) were stratified by habitat type based on elevation and water level measurements (Table 1, Schlegel GmbH (2000)). Four habitat types were distinguished, i.e. PW: ‘permanent water’; FW: “fluctuating water”; RB: “reed bed ‘/including also sedge bed & tall herbs; HW: “hardwood floodplain forest’. Due to the long disconnection of the study area from the Danube River softwood floodplain forests (SW), which are rare and an endangered habitat type in Europe, were completely missing. Therefore this habitat type could not be included in the sampling design, but in the data analysis plant species typical for softwood-forest communities were analyzed separately (see 2.4). In total we selected 65 sites (n = 16 transects *∼4 habitat types) for seed bank sampling in sectors 2–6 of the study area (Table 1). At each of the 65 sites, four replicate soil samples were taken randomly in winter 2009/2010, after the construction measures with top soil removal and prior to flooding of the new water course. As pilot surveys in summer 2009 had revealed that most seeds occurred in the upper 15–20 cm of the soil (Table A1 Appendix) we decided to take samples from 0 to 17 cm. According to the method of Ter Heerdt et al. (1996) each sample consisted of 1 l soil taken as a mixed sample from three cores (5 cm diameter) on 1 m2 plots close to the transects. In total 260 samples (780 cores) were taken. The above-ground vegetation had been analyzed in August and September 2009 prior to the seed bank sampling. Vegetation relevés were carried out along the 16 transects on contiguous 1 m2 plots from one edge to the other across the river bed (Lang et al., 2013). For each 1 m2 plot percentage cover of all vascular plant species was recorded using the decimal scale of Londo (1976). Plant nomenclature is based on Wisskirchen and Haeupler (1998). As it was not possible to differentiate between vegetative Viola reichen-

65

bachiana and V. riviniana plants, these two species were recorded together as well as Lamium galeobdolon and L. montanum. 2.3. Germination in the greenhouse In a pilot survey in summer 2009 we compared the seed concentration method of Ter Heerdt et al. (1996) with the common seedling-emergence method of Thompson and Grime (1979) and found no differences in seed numbers between the two methods (Table A1, Appendix). Therefore we used the less time-consuming seed washing and seed concentration method of Ter Heerdt et al. (1996) in our study. Soil seed bank samples were concentrated under flowing water with two sieves of 4 mm and 200 ␮m in the laboratory. All samples were stratified for six weeks at 0 ◦ C and spread out with maximum 0.5 cm thickness afterwards onto a sterile culture substrate in a 15 cm × 20 cm plant tray. The substrate was a 2.5 cm layer of peat (TKS 1, Floragard) and 0.4 cm washed quartz sand as separating layer. The trays were brought to the greenhouse to germinate, for the first six weeks with at least 12 h light and temperature between 25 ◦ C in daytime and 15 ◦ C at night. The trays were watered regularly from above to keep mostly waterlogged conditions (saturated but not inundated, see Boedeltje et al., 2002; Gurnell et al., 2007; Kenow and Lyon, 2009). To prevent contamination by external seeds they were covered with gauze and trays without seed bank samples were put randomly in the rows to detect any seed drift. Seedlings were determined (in the beginning every second day later once a week) and removed as soon as possible. Seedlings, which could not be identified initially, were cultivated in the greenhouse until determination was possible. After the germination period had finished the upper layer of the substrate with the seed bank samples was broken up with a fork, turned upside down and emerging seedlings were detected again. The detection of seedlings stopped, when no new germination occurred (maximally after seven months). 2.4. Data analysis Before further data analysis mean seed density (number of seeds) and species richness (number of species) of the four replicate samples per site and the frequency of the seed bank species were calculated. For better comparison with literature data seed density was converted to the number per 1 m2 by multiplying the area of the corer. For each of the 65 sites the species of the above-ground vegetation were listed with relative frequency per homogeneous habitat type (Fig. 2). To analyze the influence of the construction works, the samples of sector 2 were analyzed separately for 2a) with and 2b) without topsoil removal (Fig. 4). We defined those plant species as target species, which are typical for plant communities of floodplain habitats according to Ellenberg et al., 1992 (Table 2). The nomenclature of the plant communities follows Oberdorfer (2001). For the interpretation of plant species composition we calculated qualitative mean Ellenberg indicator values (Diekmann, 2003)

A. Schwab, K. Kiehl / Ecological Engineering 100 (2017) 46–55

49

frequency of species. The proportion of explained variance was calculated according to McCune and Mefford (2011) with an after-the fact evaluation using the relative Euclidean distance. Rare Species present in only one or two plots were excluded from the analysis (McCune and Mefford, 2011). 3. Results 3.1. Seed density and species richness of the seed bank

Fig. 2. Example for the distribution of the four habitat types along a transect (PW: “permanent water”; FW: “fluctuating water”; RB: “reed bed ‘/including also sedge bed & tall herbs; HW: “hardwood floodplain forest’).

Table 2 Allocation of target species of different phytosociological units according to Oberdorfer (2001) to the four habitat types ‘permanent water’ (PW), “fluctuating water” (FW), “reed beds and tall herb stands” (RB), and “hardwood floodplain forest” (HW). * The fifth habitat type softwood floodplain forest (SW) was not present in the study area before restoration, but softwood species are also target species and were found occasionally in our study. Abbreviation

Habitat type

Phytosociological units as basis for the definition of target species

PW

Permanent Water

• Lemnetea minoris Tx. 55 • Potamogetonetea pectinati Klika 41 ap. Nov. et Klika 41

FW

Fluctuating Water

• Isoëto-Nanojuncetea Br.-Bl. et Tx. 43 • Bidentetea tripartitae Tx., Lohm. et Prsg in Tx. 50 • Agrostietea stoloniferae Oberd. et Müll. ex Görs 68 • Scheuchzerio-Caricetea fuscae (Nordh. 36) Tx. 37

RB

Reed Beds & tall herb stands

• Phragmiti-Magnocaricetea Klika in Klika et Novak 41 • Convolvuletalia sepium Tx. 50 • Molinietalia caeruleae W. Koch 26

HW

HardWood floodplain forest

• Alno-Ulmion minoris Br.-Bl. et Tx. 43

(SW)*

(SoftWood floodplain forest)

• Salicetea purpureae Moor 58

for light (L) and moisture (F) for the above-ground vegetation and the seed bank of each of the 65 sites. Sørensen similarity index between seed bank and above-ground vegetation was first calculated per site using presence – absence data and then we calculated means per habitat type or sector (Sørensen, 1948). As the prerequisites for parametric statistic were not met we used non-parametric Kruskal-Wallis-ANOVA followed by nonparametric multiple comparison of means with STATISTICA 7.1 (StatSoft, Inc., 2005, www.statsoft.com) for the statistical analysis of differences of species richness and seed density between sectors (Fig. 1 and Table 1) and habitat types (Fig. 2). We analyzed the species composition of the seed bank and the above-ground vegetation by calculating a Detrended Correspondence Analysis (Ordination), with square-root transformed relative

In total, 12 606 seedlings of 124 different vascular plant species were determined in the seed bank samples of the 65 sites. The highest number of seedlings was found in a sample from a periodically flooded oxbow with 1 112 individuals per 1 l soil which is equivalent to almost 190 000 germinable seeds as calculated per 1 m2 . The average number of seedlings over all samples was 47 (±6.3) individuals per l soil (∼8000 seeds per 1 m2 ). The four habitat types differed significantly in mean seed density (Kruskal-Wallis H: 18.602, P < 0.001, df 3, n = 65) and mean species richness (Kruskal-Wallis H: 9.029, P = 0.029, df 3, n = 65) of the seed bank. Mean seed density was greatest in the habitats with fluctuating water level (∼27 500/m2 , Fig. 3A) whereas it was much lower in the parts with more or less stable conditions like permanently standing water (∼1900/m2 ; P = 0.02) and hardwood floodplain forest (∼3800/m2 ; P = 0.001). Mean species richness was also significantly higher in habitats with fluctuating water with 22 species (Fig. 3B) than in habitats permanently covered by water with eight species (P = 0.04), but differences to hardwood floodplain forest habitats with 14 species were not significant. Total species richness per floodplain habitat type increased with decreasing water influence (Table 3). Only 21 species were found in the seed bank of permanent water habitats compared to a total species number of 108 species in the hardwood floodplain forest, whereas reed beds showed intermediate values (Table 3). Both, mean seed density (Kruskal-Wallis H: 32.1525, P < 0.001, df 5, n = 45) and species richness (H: 33.387, P < 0.001, df 5, n = 45) differed not only between habitat types but also between floodplain sectors (for definitions of sectors see Table 1). In sector 2, where the ground water level before restoration was very low and only hardwood floodplain forest habitats were present, mean species richness of the seed bank (5 species per site) was significantly lowest as well as total species richness (35 species, Table 4). The seed bank of the other four sectors was very similar concerning total species richness (71–74 species) and showed fewer differences in mean species richness, which was highest in sector 5 with 23 species (Table 4). As parts of the hardwood floodplain forest dominated sector 2 had been affected by topsoil removal due to construction works all plots with hardwood floodplain forest habitats of the whole study area were analyzed separately (Fig. 4). There were no significant differences, however, in seed density or species richness between the samples of sector 2a which had been influenced by the construction work and the samples of sector 2b (without topsoil removal). Seed density and mean species richness was significantly higher in hardwood floodplain forest habitats of sectors 3–5, where all four habitat types occur adjacently. 3.2. Above-ground vegetation Mean species richness of the above-ground vegetation also showed clear differences (Kruskal-Wallis H: 32.962, P < 0.001, df 3, n = 65) between the different habitat types. Mean species richness was significantly higher in hardwood floodplain forest habitats than in the other habitat types (RB: P = 0.006; PW: P = 0.001; FW: P < 0.001), which were more affected by water (Fig. 5). Total species

50

A. Schwab, K. Kiehl / Ecological Engineering 100 (2017) 46–55

Fig. 3. Mean seed density (A) and mean species richness (B) of the seed bank for four habitat types. Boxplots indicate: 䊐 mean, standard error (box) and standard deviation (whiskers). Equal letters indicate no significant differences (P > 0.05) between habitat types (non-parametric multiple comparison according to Siegel and Castellan, 1988).

Fig. 4. Mean seed density (A) and mean species richness (B) of the seed bank shown only for the hardwood floodplain forest habitats of the five sectors (2a = topsoil removal due to construction works, 2b = no top soil removal). Boxplots indicate: 䊐 mean, standard error (box) and standard deviation (whiskers). Similar letters indicate no significant differences (P > 0.05) between sectors (multiple comparison according to Siegel and Castellan, 1988).

Table 3 Total species richness of the seed bank, the above-ground vegetation and both together within the four habitat types (PW = “permanent water”, FW = “fluctuating water”, RB = “reed bed & tall herb stands”, HF = “hardwood floodplain forest”). Data also include the mean number of species (±1SE) only represented in the seed bank or vegetation respectively and the mean Sørensen similarity index per sites. Species richness

Total

Habitat type PW

FW

RB

HF

124 53 139 68 192 71

21 17 15 11 32 4

60 44 31 15 75 16

83 49 74 40 123 34

108 56 114 62 170 52

Mean species richness Vegetation Seed bank

19.0 ± 1.1 15.6 ± 1.1

5.5 ± 2.1 7.8 ± 2.8

8.4 ± 1.2 22.0 ± 2.2

15.3 ± 2.0 17.8 ± 2.0

24.0 ± 1.0 14.3 ± 1.4

Similarity between vegetation and seed bank Mean Sørensen index (%)

20.8 ± 1.7

7.1 ± 4.2

20.7 ± 3.9

33.8 ± 2.5

16.9 ± 1.9

Total species richness Seed bank (all species) Species only found in the seed bank Above-ground vegetation (all species) Species only found in above-ground vegetation Total (either in vegetation or seed bank or both) Species present in vegetation and seed bank

richness was also highest in hardwood floodplain forest habitats (114 species) and lowest in permanent water habitats (15 species) (Table 3). Although total species richness of the above-ground vegetation was lowest in sector 2 with only 54 species, the highest mean species richness (24) was found here (Table 4).

3.3. Sørensen similarity between above-ground vegetation and seed bank A comparison of the total species richness between seed bank and above-ground vegetation showed that from a total of 192 vascular plant species 68 species were only found in the above-ground

A. Schwab, K. Kiehl / Ecological Engineering 100 (2017) 46–55

51

Table 4 Total and mean species richness of the above-ground vegetation, the seed bank and both together within the five sectors. Data also include the number of species only represented in the seed bank or vegetation respectively and the mean Sørensen similarity index per sites. (SE) = Standard error. Species richness

Total

Sector 2

3

4

5

6

192 71 139 68 124 53

75 14 54 40 35 21

118 30 75 45 73 43

117 34 80 46 71 37

106 31 66 35 71 40

121 27 74 47 74 47

Mean species richness Vegetation Seed bank

19.0 ± 1.1 15.6 ± 1.1

24.2 ± 1.4 4.9 ± 1.2

17.6 ± 2.7 16.6 ± 1.7

20.9 ± 3.0 19.6 ± 1.4

17.7 ± 1.9 22.7 ± 1.3

15.4 ± 2.6 15.1 ± 2.3

Similarity between vegetation and seed bank Mean Sørensen index (%)

20.8 ± 1.7

5.1 ± 1.8

25.7 ± 2.9

25.7 ± 3.4

27.8 ± 3.3

20.4 ± 3.4

Total species richness Total (either in vegetation or seed bank or both) Species present in vegetation and seed bank Vegetation (all species) Species only in vegetation Seed bank (all species) Species only in seed bank

Fig. 5. Mean species richness per 1 m2 of the above-ground vegetation for the four habitat types. Boxplots indicate: 䊐 mean, standard error (box) and standard deviation (whiskers). Different letters indicate significant differences (P < 0.05, Siegel and Castellan, 1988).

vegetation, 53 species only in the seed bank and 71 species both in the seed bank and in the vegetation (Tables 3 and 4). This resulted in a total Sørensen similarity index of 54% between the above-ground vegetation and the soil seed bank for the whole study area. The similarity between soil seed bank and above-ground vegetation per site differed clearly between the four habitat types (Table 3) und between the sectors (Table 4). The Sørensen similarity index was higher for reed bed & tall herb habitats (34%) and for fluctuating water habitats (21%) than for hardwood floodplain forest habitats (17%). Hardly any similarity between seed bank and aboveground vegetation was found in the permanent water habitats (7%) or in sector 2 (5%), where only hardwood floodplain forest habitats existed. 3.4. Species composition A Detrended Correspondence Analysis (DCA) indicated that the species composition of vegetation and seed bank differed clearly (Fig. 6). Differences between habitat types were more pronounced in the above-ground vegetation than in the seed bank. The first two axes of the DCA explained more than 50% of the variance in the data set (Axis 1 27.0% and Axis 2 25.8%). The number of target species typical for hardwood floodplain forests was significantly negatively correlated with the first axis (r = −0.84, P < 0.001) whereas the number of target species for permanent water habitats correlated with axis 2 (r = 0.62, P < 0.001) (Table A2, Appendix). These two species

groups dominated in the above-ground vegetation, whereas the number of species typical for fluctuating water habitats was higher in the seed bank (Table 5). Species of reed bed and tall herb stands were distributed more or less similar in the vegetation and their seed bank and showed no correlation with the two axes of the DCA (Table A2, Appendix). The mean Ellenberg indicator value for light showed a strong correlation (r = 0.90, P < 0.001) with the first axis, whereas the mean moisture value correlated with both axes (Axis 1: r = 0.59, axis 2: r = 0.56, P < 0.001) (Table A2, Appendix; Fig. 6B). Urtica dioica was the most frequent species of the whole study (overall 50.6% frequency; Table A3, Appendix). It was found in 136 seed bank samples with 3 397 individuals in total. The second frequent-species Lycopus europaeus (35.1%) was detected with only 335 individuals but in 83 seed bank samples, whereas Veronica catenata (33.7%) could be determined with 3 347 individuals in 78 seed bank samples. In contrast, Deschampsia cespitosa as third individual-rich species was detected with 425 individuals in 67 samples with a total frequency of 16.1%. Only four target species of permanent-water plant communities were found in the seed bank (Table 5). Callitriche palustris agg. was rather common (seed bank and vegetation) with a maximum frequency in the seed bank of fluctuating water habitats of 35.7% (Table A3 Appendix). Ranunculus trichophyllus and Sparganium emersum were the two target species of the permanent water habitat in the seed bank that could not be determined in the above-ground vegetation. Seventeen species of the fluctuating water habitats were found in the seed bank but not in the above-ground vegetation. The frequency of Juncus articulatus and Ranunculus sceleratus, for example, was >50% in the seed bank of these habitats. Most floodplain target species detected in both seed bank and above-ground vegetation were reed bed species and tall herbs. Phragmites australis showed the highest frequency in the vegetation of reed bed habitats (88%) and Lycopus europaeus was most frequent in the seed bank of fluctuating water habitats (75%) (Table A3, Appendix). The lack of softwood floodplain forest habitats in the study area was mirrored in the occurrence of only one target species, Salix alba, which was absent in the seed bank and found in the above-ground vegetation with a minor frequency (0.25%). Although 22 target species of hardwood floodplain forest could be determined in the vegetation only eight species were found in the seed bank, e.g. Rumex sanguineus (21.4%), Stachys sylvatica (17.1%) and Carduus personata (11.2%). Annual species as Epilobium tetragonum, Chenopodium polyspermum and Poa annua were found almost exclusively in the seed bank (Table A3, Appendix). Eight species of the Red List of Bavaria could be found only in the seed bank (Table 5). In total we found 26 target species of natural undegraded floodplains in the seed bank, which were lost or very rare in the above-ground vegetation. Most of them

52

A. Schwab, K. Kiehl / Ecological Engineering 100 (2017) 46–55

Fig. 6. (A): Results of a DCA (Detrended Correspondence Analysis) calculated from species relative frequencies in the above-ground vegetation and seed bank. Plots were grouped according to habitat types. The gradient length of axis 1 was: 4.5. (B): Vectors indicate correlations between site scores and the number of target species (see Table 2) of hardwood floodplain forest (HF), permanent water (PW), fluctuating water (FW), and for the mean Ellenberg indicator values for F: moisture and L: light.

were species typical for fluctuating water habitats (17) or reed bed stands (7); almost no species was typical for habitats with stable water levels. In sector 2 only four of the target species were found (Table 5, Table A3, Appendix). In contrast, only few non-native species were found with low frequencies (Table 5).

4. Discussion 4.1. Persistence of target species after more than 150 years of embankment Although the Danube floodplain in our study area was embanked and disconnected from fluvial dynamics for more than 150 years (Margraf, 2004; Stammel et al., 2012) target species of natural undegraded floodplains were found in the above-ground vegetation and in the seed bank. In addition we determined seed bank species, which had not been detected in the above-ground vegetation. Most of these species are typical for plant communities of regularly disturbed habitats or habitats with clear difference in water regime like lake shore or river edges (Bidentetea, Isoeto-

Nanojuncetea; Oberdorfer, 2001; Ellenberg and Leuschner, 2010). Several of them (e.g. Ranunculus sceleratus, Juncus bulbosus) are known to have a long-term persistent seed bank (Thompson et al., 1997). The other species with transient or short-term persistent seed bank types (e.g. Veronica anagallis-aquatica, V. catenata or Cyperus fuscus), however, depend on regular seed production within the embanked floodplain or on seed dispersal from external sites (Bakker et al., 1996; Thompson et al., 1997). As seed dispersal by water from external sites was highly limited in our disconnected floodplain without regular floods the survival and seed production of the remnant populations was only possible in suitable habitats within the floodplain.

4.2. Effects of habitat conditions on soil seed bank The results of our seed bank analyses indicated clear differences between the four habitat types and the five sectors of the study area. The embankment, more than 150 years ago, reduced the mechanical disturbance by flooding in all habitat types to a minimum of few extraordinarily high floods (HQ 10) which occurred e.g. 1999

Table 5 Total species richness and number of species in different species groups found in above-ground vegetation only, seed bank only and in both sources. Red-list species and non-native species are according to Scheuerer and Ahlmer (2003).

Species richness Target species

Red List species of Bavaria Non-native species

Permanent water species Fluctuating water species Reed bed & tall herb species Softwood forest species Hardwood forest species

Vegetation

Both

Seed bank

Total

68 7 1 11 1 14 15 3

71 2 10 20 0 8 13 2

53 2 17 7 0 0 8 4

192 11 28 38 1 22 37 9

A. Schwab, K. Kiehl / Ecological Engineering 100 (2017) 46–55

and 2005 (Stammel et al., 2012). In addition, the building of two hydropower stations at the beginning of the 1970s totally prevented ground water level fluctuations in large parts of the study area (Stammel et al., 2012; Lang et al., 2013). Before restoration, habitats of plant communities adapted to fluctuating water levels occurred only in depressions with rainfall-induced ground- and surface water dynamics (Margraf, 2004). In our study, habitats with changing water levels from complete inundation to periodically occurring drought periods showed significantly higher seed density and species richness of the seed bank compared to habitats with more stable conditions like permanently flooded oxbows or dry hardwood floodplain forests (Fig. 3). This may imply that habitats with regularly fluctuating water levels seem to be sufficient for reproduction and longevity of several floodplain typical plant communities (Baldwin et al., 2001; Ficken and Menges, 2013), but these habitats were restricted only to small areas in our disconnected floodplain, which is not sufficient for a fully functional floodplain system. Although our results are from a disconnected floodplain, they are similar to findings of Abernethy and Willby (1999), who studied seed density and species richness in aquatic floodplains, which were still connected to or directly affected by running water. Other studies indicate that the importance of the soil seed bank for vegetation dynamics under fluctuating habitat conditions is typical not only for aquatic (e.g. Ye et al., 2013; O’Donnell et al., 2014) but also for terrestrial systems (Middleton, 2003; Dölle and Schmidt, 2009; Wang et al., 2013) as frequent disturbance promotes species with high seed production (Grime, 1974, 1979). This is also confirming to the intermediate disturbance hypothesis of Connel (1978). Although we found low seed densities and mean species richness of the seed bank in disconnected hardwood floodplain forest habitats with long-term stable environmental conditions, total species richness of the seed bank and the above-ground vegetation of this habitat type was high, which is according to results of Baldwin et al. (2001). This seems to be in contrast to the hypothesis that under stable conditions seed bank diversity decreases (Abernethy and Willby, 1999). In our study, a comparison of the five sectors showed, however, that effects of neighboring habitats on seed bank species composition have to be considered. In sectors 3–5 seed density and mean species richness were higher than in Sector 2 even in hardwood floodplain forest habitats, because hardwood floodplain forest occurred together with habitats of fluctuating water levels due to small-scale relief changes. Effects of neighboring habitats on seed density and species composition have also been found in other studies, which stated that in habitat mosaics − which are typical for floodplains − the seed bank is often influenced by seed dispersal (Bakker et al., 1996; Lanta and Leps, 2009; O’Donnell et al., 2014). In our study, ordination results indicated more differences in species composition in the above-ground vegetation than in the seed bank and there was a strong overlap in the seed bank species composition of the different habitats. Bossuyt and Honnay (2008) also state that ‘soil seed bank communities are [generally] dominated by a low number of species, and that the dominating genera are rather similar across different communities’. Frequent species in our seed bank analysis as Urtica dioica, Veronica catenata, Lycopus europaeus or Deschampsia cespitosa are known to produce a large number of seeds, which spread in the neighboring habitats (Krämer and Fartmann, 2007; Williams et al., 2008; Gurnell et al., 2008; Greet et al., 2013). In contrast, typical forest species or aquatic macrophytes reproduce mainly vegetatively or with low seed numbers (Barrat-Segretain, 1996; Combroux and Bornette, 2004; Rodrigo et al., 2013). In addition, most of these species build up only a transient seed bank (Skowronek et al., 2014; Greet et al., 2013). This explains the fact that no hardwood floodplain forest

53

species and only two species typical for permanent water were found only in the seed bank. Similarity between seed bank and above-ground vegetation indicated by Sørensen index showed clear differences between habitats with stable conditions and habitats with fluctuating water levels. An explanation for this finding is the fact that species from the seed bank were not able to germinate under stable conditions without disturbance, because they need bare ground and light for their germination (Nishihiro et al., 2004; Duncan et al., 2009; Rivera et al., 2012). This is also indicated by the high correlation of the Ellenberg L-value with the first ordination axis separating the seed bank samples from the hardwood floodplain forest vegetation samples.

4.3. Potential significance of the seed bank for the restoration of floodplain vegetation Previous studies have found that seed banks can provide a high potential for the restoration of dynamic riparian vegetation communities (Combroux and Bornette, 2004; Boudell and Stromberg, 2008; Greet et al., 2013). In our study this became particularly clear, when target species in a stricter sense like species of the Red list of Bavaria were considered. Eight endangered species were found only in the seed bank and most of them are target species of fluctuating water habitats like Cyperus fuscus, Ranunculus sceleratus and Rumex maritimus. These species are typical for frequently disturbed floodplain habitats and have adapted their reproduction with a great number of seeds, fast reproduction and a persistent seed bank (Thompson et al., 1997; Baskin and Baskinm, 1998). This means that it should be possible to enhance plant communities with these species in the above-ground vegetation by restoring dynamic processes with flooding-induced erosion and sedimentation processes (Stammel et al., 2012; Lang et al., 2013). In contrast, the lack of germinable seeds of target species in communities typical for habitats with stable conditions like hardwood floodplain forests or permanently flooded oxbows, indicates that the seed bank has only limited potential for the restoration of these habitats (see also Skowronek et al., 2014). The generally low number and frequency of non-native species in the seed bank and in the vegetation can be explained by the long history of more than 150 years of disconnection of the floodplain from the Danube River (Margraf, 2004), which resulted in a limited dispersal of invasive non-native species. In the future, increasing disturbance due to restoration measures might favor, however, not only the re-colonization of target but also the spreading of invasive non-native species, which often use streams as dispersal corridors (Pyˇsek and Prach, 1993; Hood and Naiman, 2000).

5. Conclusions In our study seed density, species richness and species composition of the soil seed bank in a disconnected floodplain and oxbow system could be clearly related to pre-restoration dynamics of water levels. As habitats with fluctuating (ground-) water levels were still present in our disconnected floodplain future restoration success may be enhanced by the contribution of their soil seed bank to the re-establishment of rare target plant species and communities after restoring fluvial dynamics. More natural dynamics including erosion and sedimentation may activate the seed bank of remnant populations of floodplain target species. Species of softwood floodplain forests will also be favored by increased fluvial dynamics but due to their transient seed bank they depend on seed dispersal from small remnant populations in the above-ground vegetation.

54

A. Schwab, K. Kiehl / Ecological Engineering 100 (2017) 46–55

For habitats without fluctuating water levels and therefore relatively stable environmental conditions before restoration like hardwood floodplain forests and permanent waters the relation between seed bank species composition and future vegetation development is largely unpredictable due to the lack of habitat specific species with persistent seed bank and a possible influence of neighboring habitats. Acknowledgments This study was a part of the research project (MONDAU) which was supported by the German Federal Agency for Nature Conservation (BfN) with funding from the German Federal Ministry for the Environment, Nature Conservation and Nuclear Safety (BMU). We are grateful to Jörg Ewald of the University of Applied Sciences Weihenstephan-Triesdorf for providing greenhouse facilities. Bärbel Stammel and the anonymous reviewers provided helpful comments on earlier versions of the manuscript. Thanks to all colleagues from the Aueninstitut Neuburg, family and friends for supporting the works in the field, laboratory and greenhouse. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ecoleng.2016.11. 068. References Abernethy, V.J., Willby, N.J., 1999. Changes along a disturbance gradient in the density and composition of propagule banks in floodplain aquatic habitats. Plant Ecol. 140, 177–190. Amoros, C., 2001. The concept of habitat diversity between and within ecosystems applied to river side-arm. Restor. Environ. Manage. 28, 805–817. Bakker, J.P., Poschlod, P., Strykstra, R.J., Bekker, R.M., Thompson, K., 1996. Seed banks and seed dispersal: important topics in restoration ecology. Acta Botanica Neerlandica 45, 461–490. Baldwin, A.H., Egnotovich, M.S., Clarke, E., 2001. Hydrologic change and vegetation of tidal freshwater marshes: field, greenhouse, and seed-bank experiments. Wetlands 21 (4), 519–531. Barrat-Segretain, M.H., 1996. Strategies of reproduction, dispersion, and competition in river plants: a review. Plant Ecol. 123, 13–37. Baskin, C.C., Baskin, J.M., 1998. Seeds: Ecology, Biogeography, and Evolution of Dormancy and Germination. Academic Press/School of Biological Sciences. Lexington, Kentucky, US. Boedeltje, G., Ter Heerdt, G.N.J., Bakker, J.P., 2002. Applying the seedling-emergence method under waterlogged conditions to detect the seed bank of aquatic plants in submerged sediments. Aquat. Bot. 72, 121–128. Bossuyt, B., Honnay, O., 2008. Can the seed bank be used for ecological restoration? An overview of seed bank characteristics in European communities. J. Veg. Sci. 19, 875–884. Boudell, J.A., Stromberg, J.C., 2008. Propagule banks: potential contribution to restoration of an impounded and dewatered riparian ecosystem. Wetlands 28, 656–665. Bunn, S.E., Arthington, A.H., 2002. Basic principles and ecological consequences of altered flow regimes for aquatic biodiversity. Environ. Manage. 30, 492–507. Combroux, I.C.S., Bornette, G., 2004. Propagule banks and regenerative strategies of aquatic plants. J. Veg. Sci. 15, 13–20. Connel, J.H., 1978. Diversity in tropical rain forests and coral reefs. Science 199, 1302–1310. Dölle, M., Schmidt, W., 2009. The relationship between soil seed bank, above-ground vegetation and disturbance intensity on old-field successional permanent plots. Appl. Veg. Sci. 12, 415–428. Diekmann, M., 2003. Species indicator values as an important tool in applied plant ecology − a review. Basic Appl. Ecol. 4, 493–506. Doi, H., Katano, I., Negishi, J.N., Sanada, S., Kayaba, Y., 2013. Effects of biodiversity, habitat structure, and water quality on recreational use of rivers. Ecosphere 4 (article 102). Duncan, R.P., Diez, J.M., Sullivan, J.J., Wangen, S., Miller, A.L., 2009. Safe sites, seed supply, and the recruitment function in plant populations. Ecology 90, 2129–2138. Ellenberg, H., Leuschner, C., 2010. Vegetation Mitteleuropas mit den Alpen. Ulmer. Ellenberg, H., Weber, H.E., Düll, R., Wirth, V., Werner, W., Paulissen, D., 1992. Zeigerwerte von Pflanzen in Mitteleuropa. Scripta Geobotanica 18, 2 (Auflage). European Union, 2000. Directive 2000/60/EC of the European Parliament and of the Council of 23 October 2000 establishing a framework for communities in the field of water policy. Off. J. Eur. Commun. L327 (1) (22.12.2000).

European Union, (2004), Council Directive 92/43/EEC of 21 May 1992 on the conservation of natural habitats and of wild fauna and flora. Office for Official Publications of the European Communities, CONSLEG: 1992L0043–01/05/2004. Feld, C.K., Birk, S., Bradley, D.C., Hering, D., Kail, J., Marzin, A., Melcher, A., Nemitz, D., Pedersen, M.L., Pletterbauer, F., Pont, D., Verdonschot, P.F.M., Friberg, N., 2011. From natural to degraded rivers and back again: a test of restoration ecology theory and practice. Adv. Ecol. Res. 44, 119–209. Ficken, C.D., Menges, E., 2013. Seasonal wetlands on the Lake Wales Ridge, Florida: does a relict seed bank persist despite long term disturbance? Wetl. Ecol. Manage. 21, 373–385. Goodson, J.M., Gurnell, A.M., Angold, P.G., Morrissey, I.P., 2001. Riparian seed banks: structure: process and implications for riparian management. Prog. Phys. Geogr. 25, 301–325. Greet, J., Cousens, R.D., Webb, J.A., 2013. Flow regulation is associated with riverine soil seed bank composition within an agricultural landscape: potential implications for restoration. J. Veg. Sci. 24, 157–167. Grime, J.P., 1974. Vegetation classification by reference to strategies. Nature 250, S26–S31. Grime, J.P., 1979. Plant Strategies and Vegetation Processes. Wiley, Chichester, pp. 222S. Gumiero, B., Mant, J., Hein, T., Elso, J., Boz, B., 2013. Linking the restoration of rivers and riparian zones/wetlands in Europe: sharing knowledge through case studies. Ecol. Eng. 56, 36–50. Gurnell, A., Goodson, J., Thompson, K., Mountford, O., Clifford, N.J., 2007. Three seedling emergence methods in soil seed bank studies: implications for interpretation of propagule deposition in riparian zones. Seed Sci. Res. 17, 183–199. Gurnell, A., Thompson, K., Goodson, J., Moggridge, H., 2008. Propagule deposition along river margins: linking hydrology and ecology. J. Ecol. 96, 553–565. Hood, W.G., Naiman, R.J., 2000. Vulnerability of riparian zones to invasion by exotic vascular plants. Plant Ecol. 148, 105–114. Jensen, K., Trepel, M., Merritt, D., Rosenthal, G., 2006. Restoration ecology of river valleys. Basic Appl. Ecol. 7, 383–387. Kenow, K.P., Lyon, J.E., 2009. Composition of the seed bank in drawdown areas of navigation pool 8 of the upper Mississippi river. River Res. Appl. 25, 194–207. Krämer, D., Fartmann, T., 2007. Syntaxonomy and ecology of riverine pioneer vegetation of the Chenopodion glauci (class: bidentetea tripartitae) in the Lower Oder valley. Tuexenia 27, 195–219. Lang, P., Frey, M., Ewald, J., 2011. Waldgesellschaften und Standortabhängigkeit der Vegetation vor Beginn der Redynamisierung der Donauauen zwischen Neuburg und Ingolstadt. Tuexenia 31, 39–57. Lang, P., Schwab, A., Stammel, B., Ewald, J., Kiehl, K., 2013. Long-term vegetation monitoring for different habitats in floodplains. Sci. Ann. Danube Delta Inst. 19, 39–48. Lanta, V., Leps, J., 2009. How does surrounding vegetation affect the course of succession: a five-year container experiment. J. Veg. Sci. 20, 686–694. Liu, W.Z., Zhang, Q.F., Liu, G.H., 2009. Seed banks of a river–reservoir wetland system and their implications for vegetation development. Aquat. Bot. 90, 7–12. Londo, G., 1976. The decimal scale for releves of permanent quadrats. Vegetatio 33, 61–64. Lu, Z.J., Li, L.F., Jiang, M.X., Huang, H.D., Bao, D.C., 2010. Can the soil seed bank contribute to revegetation of the drawdown zone in the Three Gorges Reservoir Region? Plant Ecol. 209, 153–165. Margraf, C., 2004. Die Vegetationsentwicklung der Donauaue zwischen Ingolstadt und Neuburg. Hoppea. Denkschriften der Regensburgischen Botanischen Gesellschaft 65, 295–704. Mant, J., Gill, A.B., Janes, M., Di, H., 2012. Restoration of rivers and floodplains. In: van Andel, J., Aronson, J. (Eds.), Restoration ecology - the new frontier. Wiley Blackwell, Oxford. McCune, B., Mefford, M.J., 2011. PC-ORD. Multivariate Analysis of Ecological Data. Version 6.13. MjM Software, Gleneden Beach, Oregon, U.S.A. Middleton, B.A., 2003. Soil seed banks and the potential restoration of forested wetlands after farming. J. Appl. Ecol. 40, 1025–1034. Mitsch, W.J., Gosselink, J.G., 2015. Wetlands, 5th ed. Wiley, New York. Nilsson, C., Svedmark, M., 2002. Basic principles and ecological consequences of changing water regimes: riparian plant communities. Environ. Manage. 30, 468–480. Nishihiro, J., Araki, S., Fujiwara, N., Washitani, I., 2004. Germination characteristics of lakeshore plants under an artificially stabilized water regime. Aquat. Bot. 79, 333–343. O’Donnell, J., Fryirs, K., Leishman, M.R., 2014. Digging deep for diversity: riparian seed bank abundance and species richness in relation to burial depth. Freshw. Biol. 59, 100–113. Oberdorfer, E., 2001. Pflanzensoziologische Exkursionsflora für Deutschland und angrenzende Gebiete, 8th ed. Ulmer Stuttgart, Germany. Pyˇsek, P., Prach, P., 1993. Plant invasions and the role of riparian habitats: a comparison of four species alien to central Europe. J. Biogeogr. 20, 413–420. Rivera, D., Jauregui, B.M., Peco, B., 2012. The fate of herbaceous seeds during topsoil stockpiling: restoration potential of seed banks. Ecol. Eng. 44, 94–101. Rodrigo, M.A., Rojo, C., Alonso-Guillén, J.L., Vera, P., 2013. Restoration of two small Mediterranean lagoons: the dynamics of submerged macrophytes and factors that affect the success of revegetation. Ecol. Eng. 54, 1–15. Sørensen, T., 1948. A method of establishing groups of equal amplitude in plant sociology based on similarity of species content, and its application to analysis

A. Schwab, K. Kiehl / Ecological Engineering 100 (2017) 46–55 of the vegetation on Danish commons. Kongelige Danske Videnskabernes Selskabs Biologiske Skrifter 5, 1–34. Scheuerer, M., Ahlmer, W., 2003. Rote Liste gefährdeter Gefäßpflanzen Bayerns mit regionalisierter Florenliste. In: Schriftenreihe des Bayerischen Landesamtes für Umweltschutz (Augsburg, Germany) 165, 1–372. Schlegel GmbH, 2000. Renaturierung der Donauauen. Auswertungder Grundwassermessdaten. Berichte und Karten. München (unpublished). Siegel, S., Castellan, N.J., 1988. Nonparametric Statistics for the Behavioral Sciences, 2nd ed. McGraw-Hill, New York. Skowronek, S., Terwei, A., Zerbe, S., Molder, I., Annighofer, P., Kawaletz, H., Ammer, C., Heilmeier, H., 2014. Regeneration potential of floodplain forests under the influence of nonnative tree species: soil seed bank analysis in Northern Italy. Restor. Ecol. 22, 22–30. Stammel, B., Cyffka, B., Geist, J., Müller, M., Pander, J., Blasch, G., Fischer, P., Gruppe, A., Haas, F., Kilg, M., Lang, P., Schopf, R., Schwab, A., Utschick, H., Weissbrod, M., 2012. Floodplain restoration on the Upper Danube (Germany) by reestablishing back water and sediment dynamics: a scientific monitoring as part of the implementation. River Syst.: Integr. Landsc. Catchment Perspect. Ecol. Manage. 20, 55–70. StatSoft, Inc., 2005. STATISTICA for windows [software-system for data-analysis] Version 7.1. www.statsoft.com. Ter Heerdt, G.N.J., Verweij, G.L., Bekker, R.M., 1996. An improved method for seed bank analysis: seedling emergence after removing the soil by sieving. Funct. Ecol. 10, 245–248. Thompson, K., Grime, J.P., 1979. Seasonal variation in the seed banks of herbaceous species in 10 contrasting habitats. J. Ecol. 67, 893–921.

55

Thompson, K., Bakker, J., Bekker, R., 1997. The Soil Seed Banks of North West Europe: Methodology, Density and Longevity. Cambridge University Press, Cambridge, UK. Tockner, K., Schiemer, F., Baumgartner, C., Kum, G., Weigand, E., Zweimuller, I., Ward, J.V., 1999. The Danube restoration project: species diversity patterns across connectivity gradients in the floodplain system. Regul. Rivers: Res. Manage. 15, 245–258. Wang, N., Jiao, J.Y., Du, H.D., Wang, D.L., Jia, Y.F., Chen, Y., 2013. The role of local species pool, soil seed bank and seedling pool in natural vegetation restoration on abandoned slope land. Ecol. Eng. 52, 28–36. Ward, J.V., Tockner, K., Schiemer, F., 1999. Biodiversity of floodplain river ecosystems: ecotones and connectivity. Regul. Rivers: Res. Manage. 15, 125–139. Ward, J.V., Tockner, K., Arscott, D.B., Claret, C., 2002. Riverine landscape diversity. Freshw. Biol. 47, 517–539. Williams, L., Reich, P., Capon, S.J., Raulings, E., 2008. Soil seed banks of degraded riparian zones in southeastern Australia and their potential contribution to the restoration of understory vegetation. River Res. Appl. 24, 1002–1017. Wisskirchen, R., Haeupler, H., 1998. Standardliste der Farn- und Blütenpflanzen Deutschlands. Verlag Eugen Ulmer, Stuttgart, Germany. Woolsey, S., Capelli, F., Gonser, T., Hoehn, E., Hostmann, M., Junker, B., Paetzold, A., Roulier, C., Schweizer, S., Tiegs, S.D., Tockner, K., Weber, C., Peter, A., 2007. A strategy to assess river restoration success. Freshw. Biol. 52, 752–769. Ye, C., Zhang, K., Deng, Q., Zhang, Q., 2013. Plant communities in relation to ‘flooding and soil characteristics in the water level fluctuation zone of the Three Gorges Reservoir, China. Environ. Sci. Pollut. Res. 20, 1794–1802.