Ecological assessment of different riverbank revitalisation measures to restore riparian vegetation in a highly modified river

Abstract

Since anthropogenic activities have become concentrated along rivers, river regulations have strong-ly reduced the lateral connectivity by separating rivers from their floodplains. Consequently riparian habitat heterogeneity and the related species diversity are degrading, especially in highly modified prealpine rivers. Riverbank revitalisation measures aim at mitigating this degradation, but although river restoration projects have become widespread, little knowledge exists about their specific outcome, as standardised monitoring programs are missing. The aim of this study is to systematically compare vegetation change in response to three contrasting measures of bank diversification, i.e. embankment removal, sand input or gravel addition. Moreover, the influence of these measures on adjacent vegeta-tion is studied. Conclusions were drawn on the basis of three common goals of restoration projects: (i) improvement of vegetation structure, (ii) increase of species diversity, and (iii) characteristic species composition. The field work was done along River Inn northeast of Munich. Vegetation structure, species identity and cover as well as selected habitat variables were recorded in a stratified randomised sampling design; variation between measures was analysed using uni- and multivariate statistics. We detected great differences in the effect of the three measures two years after implementation. Embank-ment removal initiated highly dynamic habitats where plant establishment was difficult. The input of sand led to a rather homogenous species composition, at least partly because the habitats were produc-tive and therefore most likely will develop to tall reed stands or riparian forests. After gravel addition the restored sites remained relatively open, while riparian pioneer species could colonise. Vegetation structure and composition of adjacent reed stands were positively affected. The results indicate how restoration outcomes can vary depending on the specific measures chosen. This confirms the need for careful consideration of the pursued goals and site-specific conditions prior to implementation as well as long-term monitoring after implementation.

Full text

Tuexenia 35: xxx–xxx. Göttingen 2015.
doi: 10.14471/2015.35.005, available online at www.tuexenia.de
Ecological assessment of different riverbank revitalisation
measures to restore riparian vegetation in a highly
modified river
Ökologische Bewertung unterschiedlicher Uferstrukturierungsmaßnahmen
zur Renaturierung der Ufervegetation eines anthropogen geprägten Flusses
Katharina Strobl*, Anna-Lena Wurfer & Johannes Kollmann
Chair of Restoration Ecology, Technische Universität München, Emil-Ramann-Straße 6,
85354 Freising, Germany, katharina.strobl@tum.de; xxxxx@xxxx.de; jkollmann@wzw.tum.de
*Corresponding author
Abstract
Since anthropogenic activities have become concentrated along rivers, river regulations have strong-
ly reduced the lateral connectivity by separating rivers from their floodplains. Consequently riparian
habitat heterogeneity and the related species diversity are degrading, especially in highly modified
prealpine rivers. Riverbank revitalisation measures aim at mitigating this degradation, but although
river restoration projects have become widespread, little knowledge exists about their specific outcome,
as standardised monitoring programs are missing. The aim of this study is to systematically compare
vegetation change in response to three contrasting measures of bank diversification, i.e. embankment
removal, sand input or gravel addition. Moreover, the influence of these measures on adjacent vegeta-
tion is studied. Conclusions were drawn on the basis of three common goals of restoration projects: (i)
improvement of vegetation structure, (ii) increase of species diversity, and (iii) characteristic species
composition. The field work was done along River Inn northeast of Munich. Vegetation structure,
species identity and cover as well as selected habitat variables were recorded in a stratified randomised
sampling design; variation between measures was analysed using uni- and multivariate statistics. We
detected great differences in the effect of the three measures two years after implementation. Embank-
ment removal initiated highly dynamic habitats where plant establishment was difficult. The input of
sand led to a rather homogenous species composition, at least partly because the habitats were produc-
tive and therefore most likely will develop to tall reed stands or riparian forests. After gravel addition
the restored sites remained relatively open, while riparian pioneer species could colonise. Vegetation
structure and composition of adjacent reed stands were positively affected. The results indicate how
restoration outcomes can vary depending on the specific measures chosen. This confirms the need for
careful consideration of the pursued goals and site-specific conditions prior to implementation as well
as long-term monitoring after implementation.
Keywords: embankment removal, floodplain, gravel addition, monitoring, restoration, sand input
Erweiterte deutsche Zusammenfassung am Ende des Artikels
Manuscript received 26 November 2014, accepted 03 March 2015
Co-ordinating Editor: Norbert Hölzel
1

1. Introduction
River ecosystems are among the biologically most diverse ecosystems in the world
(WARD et al. 1999), and they provide important ecosystem services like flood mitigation,
nutrient exchange, drinking water supply and habitat provisioning (SCHOLZ et al. 2012). At
the same time river biodiversity and the associated ecosystem processes are negatively af-
fected (WARD et al. 1999), since anthropogenic activities have become concentrated in ripar-
ian zones (HABERSACK & PIÉGAY 2008). River regulations alter natural longitudinal and
lateral dynamics, which sustain high riverine habitat heterogeneity and species diversity
(WARD et al. 1999). Yet, as demonstrated by the Flood Pulse Concept (JUNK et al. 1989),
strong hydrological and ecological interconnections of rivers and their floodplains are the
basis for the functioning of riverine ecosystems.
Riparian habitats and communities are particularly affected by these changes, as rivers
and their floodplains have become separated by embankments and straightening (STAMMEL
et al. 2012). The riparian species richness generally decreases as habitat diversity gets lost
(WARD et al. 1999). Specialists like riparian pioneers are outcompeted by woody species
(KARRENBERG et al. 2002) or herbaceous perennials (ELLENBERG & LEUSCHNER 2010),
when connectivity and disturbance by floods are missing (TOCKNER et al. 1999). This often
results in a ‘single-channel’ morphology and straight bank lines with a dominance of monot-
onous tall reed stands dominated by herbaceous perennials, while there is a loss of diverse
open communities dominated by annual pioneers, which are the most threatened species in
floodplains (VON HEßBERG 2003).
In order to counteract these degradations, river restoration projects have increased expo-
nentially in the past years, including a large number of actions on riverbanks (BERNHARDT et
al. 2005). Nevertheless, although there is a large diversity of restoration measures, little is
known about their impact on flora, fauna and the surrounding landscape (BERNHARDT et al.
2005). This is mainly due to a lack of monitoring programs, which meet statistical and scien-
tific demands (CHAPMAN & UNDERWOOD 2000, BERNHARDT et al. 2005). Moreover, the
studies which fit scientific requirements focus mostly on the evaluation of the effectiveness
of single types of restoration measures in comparison to control sites (MÜLLER et al. 2014),
like removing bank fixation (JÄHNIG et al. 2009, JANUSCHKE et al. 2011), erosion control by
gravel addition (BARITEAU et al. 2013), or creation of new floodplain channels (SCHAICH et
al. 2010, STAMMEL et al. 2012). The publications, which compare the effects of different
types of restoration, focus mostly on aquatic organisms like fish and macroinvertebrates
(MÜLLER et al. 2014), or specific hydraulic aspects like the effect of embankment techniques
on biodiversity (CAVAILLÉ et al. 2013). In contrast, comparative studies take rarely into
account riparian vegetation, although it is a suitable indicator for restoration success (JÄHNIG
et al. 2009, JANUSCHKE et al. 2011). Plant species respond relatively fast to riverbank resto-
ration, due to effective dispersal or persistence in the soil seed bank (JÄHNIG et al. 2009), and
they are the first to indicate changes in abiotic conditions (ELLENBERG et al. 1992). As our
investigations as well as many monitoring programs are carried out only a few years after
restoration, riparian vegetation is a suitable indicator showing early developments.
Besides systematic comparative studies of contrasting types of restoration measures, it is
important to evaluate whether there are positive effects on adjacent vegetation and habitat
types. Even if some studies exist concerning the influence of land use on riparian forests
(cf. FERREIRA et al. 2005, FERNANDES et al. 2011), none of them deals with the specific
effects of restoration measures. However, such knowledge is of great practical relevance,
2

since it answers the question of possible positive effects of small-scale restoration measures
on larger areas. This is particularly important in highly modified rivers where space for
restoration measures is limited (DE NOOIJ et al. 2006).
The analysis of the restored communities often focuses on the presence of target species,
rare species or habitat specialists, and the absence of alien plants (HÖLZEL & OTTE 2001,
JÄHNIG et al. 2009, CAVAILLÉ et al. 2013, JANUSCHKE 2014). However, there is a problem
with defining target species in riparian vegetation, and a potential conflict between total
species diversity and presence of rare species, especially in ruderal sites of restored rivers.
Here we focus on the prealpine River Inn as suitable study system to better understand
the effects of contrasting measures that enhance natural dynamics and structural diversity of
riparian vegetation. The removal of embankment as well as the addition of gravel or sand
aim to (i) improve structural diversity, (ii) increase species diversity, and (iii) create a char-
acteristic species composition. The aim of our study is to compare these different revitalisa-
tion treatments (removal of embankment, sand input or gravel addition) by analysing vegeta-
tion change as indicator of restoration success. In addition, the influence of revitalisation
measures on adjacent vegetation types is investigated. We ask the following specific ques-
tions: 1) Is the removal of the embankment or sediment input better suited for increasing the
riparian diversity? 2) Does vegetation diversity (structure, small-scale variability, species
richness) in prealpine rivers respond differently when gravel or sand is added? 3) Can an
impact of the revitalisation measures be observed in adjacent vegetation?
2. Material and methods
2.1 Study area and revitalisation measures
The study area is located in the upper valley of the Bavarian Inn, ca. 60 km east from Munich, be-
tween the barrages Feldkirchen in the south (RW 4511890, HW 5313609, 440 m a. s. l.) and Gars in the
north (RW 4523404, HW 5335074, 412 m). The River Inn originates at about 2,500 m altitude in Swit-
zerland and flows into the Danube in Passau (300 m). Two thirds of its total length of 500 km is located
in the Alps resulting in a flow regime that is highly influenced by mountains.
In the study area, River Inn crosses the prealpine Inn-Chiemsee hilly region (Inn-Chiemsee-
Hügelland, MEYNEN et al. 1962), the end moraine and the Lower Inn Valley (Isar-Inn-Schotterplatten¸
MEYNEN et al. 1962). The studied river section is situated in the braided reach. It has a slope of
0.5-1.0‰ and its floodplain can be attributed to the group of (pre)alpine floodplains (gefällereiche
Flussaue der Alpen / Voralpen; KOENZEN et al. 2005). The natural substrate type of the river and its
floodplain is gravel (MEYNEN et al. 1962, KOENZEN 2005). According to ECKELMANN et al. (1996) this
grain fraction is round shaped and >2 mm. The mean discharge (1965–2012) is 221 m3 s-1 in winter and
490 m3 s-1 in summer in Wasserburg, which represents the centre of the investigated river section (LFU
2015). As many other European rivers the Inn is nowadays highly modified. A total of 24 barrages were
constructed; four of them located in the study area (Feldkirchen, Wasserburg, Teufelsbruck, Gars).
The barrages and other intensive hydro-engineering activities led to a loss of the natural river dy-
namics. The straightening caused a lowering of the riverbed and of the groundwater level (CONRAD-
BRAUNER 1994). Due to the construction of hydropower plants, the river carries almost no gravel
anymore, but mostly silt, sand and organic material (<2 mm; CONRAD-BRAUNER 1994). The mean
daily transport of suspended solid materials is 4.5 kg s-1. During flooding these fine sediments are
deposited along the riverbank, resulting in a strong reduction of floodplain dynamics, little turnover of
sand or gravel bars (CONRAD-BRAUNER 1994), and a disconnection between river and floodplain
(HABERSACK & PIÉGAY 2008).
3

In order to counteract these degradation processes, the revitalisation along River Inn aimed at im-
proving the ecological quality of the river and its floodplains as well as at increasing the natural dynam-
ics. Therefore, several restoration and revitalisation measures were carried out in 15 sites within a ca.
35 km long section of the river in winter 2012/13. For the purpose of this study only the riverbank
diversification measures were considered and solely their impact on riparian vegetation was studied.
They aimed at creating more diverse riparian habitats by adding or removing sediment. All sediments
originated from the study area to avoid introduction of non-native species and genotypes. In particular,
three measures can be distinguished: (i) embankment removal, (ii) sand input, or (iii) gravel addition.
i. Embankment removal consisted in the removal of constructed reinforcements or opening of natu-
rally fixed embankments by using excavators. The resulting open sites should facilitate subsequent
erosion. The substrate type favoured by this measure was mostly sand.
ii. Sand input was done in front of the bank lines. The sand originated from the digging of oxbow
lakes in the direct surroundings. It was deposited in form of piles of variable height (1.0–2.5 m) and
became eroded by river dynamics. The gravel cover in these sites was 0–10%.
iii. Gravel addition was also done in front of the straight or opened bank; the material was available
from local deposit. In most cases, gravel was introduced in form of flat groynes. The slope of these
sites is half as steep as the one of the two other measures. Gravel cover ranged from 70–100% in
these sites.
Three representative study sites were chosen for each of the three measure types. If more than three
sites were available, the choice was determined by accessibility. For the study of the adjacent vegeta-
tion types five locations were chosen, and reference sites were selected as a control for each of these
five sites. The references represent the regulated pre-restoration state and are located upstream from
each restored section. Thus, they are similar in terms of habitat conditions and exposed to similar dis-
turbance regimes, albeit without being affected by the restoration measure.
2.2 Study design and vegetation sampling
For the evaluation of floristic patterns, the survey was conducted on two spatial scales. On a small-
er scale, a map of riparian habitat types was obtained by studying aerial photographs of summer 2012.
As most of these pictures were taken prior to restoration, information was completed by more recent
pictures from GoogleEarth (GOOGLE INC. 2012) and verified in the field. Thus, homogenous structures
were delimited and assigned to three different categories. A predefined standard was established for the
mapping in order to prevent interpretation errors and to enable the reproducibility of the mapping:
− Reed: herbaceous vegetation, which can be distinguished from shrubs by a different shade of green
and missing shadows;
− Bare soil with pioneer vegetation: sparsely vegetated or vegetation free areas with clear characteris-
tics of exposed sediments;
− We further distinguished between sand and gravel dominated bare soil habitats on aerial photo-
graphs if possible and verified it in the field.
On a larger scale, vegetation and abiotic variables were studied in the field using a stratified-
randomised sampling design. The different strata corresponded to the habitat types delineated on the
aerial photographs and described above. In each stratum five units were randomly selected at each
location. The randomisation was carried out in two steps for each stratum. Firstly the placement on the
longitudinal axis was randomly selected. Secondly the plot was again placed randomly on a transverse
line through the first selected point. In narrow vegetation stripes, the plots were located in the middle of
these strata, in order to avoid edge effects. The plot size was adapted to each habitat type in order to
meet the requirements of the minimum area (MÜLLER-DOMBOIS & ELLENBERG 1974). In reed zones
the plots measured 1.4 x 1.4 m, in pioneer vegetation 1.0 x 1.0 m.
We chose three locations for each measure type (embankment removal, addition of gravel, intro-
duction of sand) and five for reed and control plots, thus investigating 15 plots per measure type (3 x 5
plots) and 25 plots per treatment (5 x 5 plots). Thus, the study design resulted in a total of 95 sampling
plots.
4

The fieldwork was carried out between mid-June and early August 2014, when riparian vegetation
was most developed. The order of the study sites was random in order to compensate for phenological
differences between early and late sampling. Vegetation was recorded according to the BRAUN-
BLANQUET (1964) method, which was slightly modified to fit our purposes. The vegetation cover was
sampled instead of the abundance and number of individuals. The cover of each species, the proportion
of vegetation, moss, litter and bare ground were estimated, and the mean vegetation height was meas-
ured. Nomenclature of plant species followed BUTTLER & HAND (2008).
In addition, soil pH, distance to the riverbank and the slope of the riverbank were recorded. The pH
value of the soil was measured using pH-indicator strips following the protocol of ECKELMANN et al.
(1996). The slope of the riverbank was studied with a level instrument (Theis Tecomat 5/8’’); because
of the limited visibility, it could not be studied in reed zones. An orthogonal transect from the water-
front towards the plot until the forest or reed edge was established for each plot.
2.3 Conservation status and habitat specialism
Plants were identified as target species for pioneer vegetation or reed zones according to the classi-
fication of ELLENBERG et al. (1992). They were described as target species for pioneer vegetation on
bare soil habitats when attributed to one of the following phytosociological classes: Isoëto-Littorelletea
Br.-Bl. Et Vlieger in Vlieger 1937, Scheuchzerio-Caricetea nigrae Tüxen 1937, Isoëto-Nanojuncetea
Br.-Bl. et Tx. ex Westhoff et al. 1946, Bidentetea tripartitae Tx. et al. ex von Rochow 1951 or Agrost-
ietea stoloniferae Oberdorfer et al. 1967 (= Polygono-Potentilletalia anserinae Tx. 1947).Species were
defined as target species for reed-zones when they were characteristic for the phytosociological class
Phragmito-Magnocaricetea Klika in Klika et Novák 1941, the orders Calystegietalia sepium Tx. ex
Moor 1958 or Molinietalia cearuleae W. Koch 1926 or the alliance Filipendulion ulmariae Segal ex
Lohmeyer in Oberd. et al. 1967. Species were called threatened when they were assigned to the catego-
ries 1, 2, 3 or V (V = likely to become endangered in the near future) in the Red List of Germany
(LUDWIG & SCHNITTLER 1996) or Bavaria (LFU 2003). Alien species were determined according to
the black list of invasive species (NEHRING et al. 2013), and the list of neophytes from the BFN (2014).
The definition of moisture indicating species was based on Ellenberg indicator values (ELLENBERG
et al. 1992). We calculated the median of the moisture value over all species to compare the average
moisture value of different vegetation units. In order to investigate whether the strategies of the riparian
vegetation differed between treatments, life-forms according to the classification of Raunkiaer
(MÜLLER-DOMBOIS & ELLENBERG 1974) were analysed. The amount of therophytes indicates the
disturbance regime, whereas the abundance of phanerophytes shows the regeneration of riparian forest.
2.4 Data analyses
Alpha-diversity was analysed directly via the number of species per plot as well as by the determi-
nation of the evenness index, in order to get information about the relative abundance of species. It was
calculated using the following equation:
=


where Hs is the alpha-diversity according to SHANNON (1948) and S is the total number of species.
Non-metric multidimensional scaling (NMDS) based on Bray-Curtis dissimilarities was computed
to visualise variation within and among types of measure. It aimed to investigate whether species com-
munities differ depending on the revitalisation measures. The appropriate number of dimensions was
determined by evaluating a scree plot choosing the number of axes, beyond which stress values do not
reduce considerably anymore (MCCUNE & GRACE 2002). Abiotic variables were subsequently overlaid
in the NMDS graph as explicating variables. Species used for emphasising results and statements could
clearly be assigned to one measure type on the ordination graphs. In order to test whether there are
significant differences in species composition PERMANOVA (permutational multivariate analysis of
variance; ANDERSON (2001) based on Bray-Curtis dissimilarities was applied.
5

Univariate statistics were investigated to test significant differences in the distribution of the
measures and to support the results of the multivariate analyses. The normality of the data and the
homogeneity of variances were tested using the Shapiro-Wilk test and the Levene test. In case of nor-
mality and homogeneity of variances a t-test was computed for pairwise comparisons and one-way
ANOVA for more than two variables (a posteriori Tukey-HSD). Yet the Mann-Whitney U test for
pairwise comparisons or the Kruskal-Wallis test (a posteriori Tamhane-T2) for multiple comparisons
was chosen if data were not normally distributed. Non-parametric Spearman-rank correlations (rho)
were calculated to test relations between different variables, since normality and homogeneity of vari-
ances could not be detected for all data.
The univariate statistics were carried out using the IBM SPSS Statistics 22 (IBM® 2013) software.
For multivariate analysis the statistical software program PCOrd (MCCUNE & MEFFORD 2011) was
used.
3. Results
3.1 Structural diversity
A total of 45 plots were analysed to compare the three types of riverbank revitalisation
measures: (i) embankment removal, (ii) sand input, and (iii) gravel addition. Vegetation
structure differed significantly between the three types of measure. The removal of the em-
bankment resulted in a significantly higher cover of bare soil (p < 0.001) in comparison to
sand input, whereas gravel addition showed intermediate values and no significant differ-
ences to the other types of measure. Vegetation cover and height were significantly lower on
plots with gravel than on those with sand (Fig. 1; p < 0.01, p < 0.001) and intermediate on
embankment removal. These results were also supported by the negative correlation of grav-
el cover with vegetation height (rho = -0.68, p < 0.01) and vegetation cover (rho = -0.64,
p < 0.01), as well as the positive correlation with the cover of bare soil (rho = 0.74,
p < 0.01).
3.2 Species diversity
In total 117 plant species were recorded. The median of the species number per plot after
embankment removal was 9 compared to 13 on sand and 10 on gravel. The species number
was significantly lower, when the embankment was removed than when sandy sediment was
added (Fig. 2; p < 0.05), even if the calculation of the evenness showed no significant differ-
ences between the three types of measure (p > 0.05). Both diversity indicators showed a
higher variance of values for the sites restored by embankment removal, while the other two
measures were more homogenous.
3.3 Species composition
The NMDS ordination showed clear differences in the species composition of the three
types of measure (Fig. 3) and PERMANOVA revealed that these differences are significant
(F = 3.80, p < 0.01). The NMDS ordination showed a clear distinction between sand and
gravel addition plots, where the species composition was mostly homogenous. However, the
plots recorded after removal of the embankment largely overlapped with the other two stud-
ied riverbank revitalisation measures. Their species composition was more variable.
6

Fig. 1. Comparison of vegetation cover (left) and vegetation height (right) for the three different types
of measure: embankment removal (black), sand input (dark grey) and gravel addition (light grey).
Gravel cover ranged 70–100% on sites with gravel addition, while it was between 0–10% on sites with
sand input and between 0–5% on sites with embankment removal. Different letters indicate significant
differences (p < 0.05; median; box, 25–75%; whiskers, min–max).
Abb. 1. Vergleich der Vegetationsdeckung (links) und der Vegetationshöhe (rechts) der drei Maßnah-
mentypen (vlnr): Uferanbruch (schwarz), Sandzugabe (dunkelgrau) und Kieszugabe (hellgrau). Die
Kiesdeckung betrug 70–100 % auf Flächen mit Kieszugabe, 0–10 % auf Flächen mit Sandzugabe und
0-5 % nach Uferanbruch. Unterschiedliche Buchstaben verweisen auf signifikante Unterschiede
(p < 0,05; Median; Box: 25–75 %; Whisker: Min–Max).
Fig. 2. Comparison of species number (left) and evenness (right) for the three different types of meas-
ure: embankment removal (black), sand input (dark grey) and gravel addition (light grey). Different
letters indicate significant differences (p < 0.05).
Abb. 2. Vergleich der Artenzahl (links) und des Evenness Index (rechts) der Maßnahmentypen Uferan-
bruch (schwarz), Sandzugabe (dunkelgrau) und Kieszugabe (hellgrau). Unterschiedliche Buchstaben
verweisen auf signifikante Unterschiede (p < 0,05).
7

Fig. 3. NMDS ordination based on Bray-Curtis dissimilarity for the comparison of species composition
of 15 plots sampled on three sites with embankment removal (grey crosses), 15 plots sampled on three
sites with sand input (grey circles) and 15 plots sampled on three sites with gravel addition (black
quadrats). It is based on the cover of 117 plant species and is visualised as a joint plot with environmen-
tal gradients represented as black vector lines (cutoff r² = 0.15; stress based on two dimensions: 0.18;
Monte-Carlo: p < 0.001). The vector lines of ‘vegetation height’ and ‘median moisture value’ as well as
‘distance to riverbank’ and ‘cover alien species’ overlap.
Abb. 3. NMDS Ordination basierend auf Bray-Curtis Unähnlichkeiten. Es wird die Artenzusammen-
setzung von je 15 Plots und je drei Standorten nach Uferanbruch (dunkelgraue Kreuze) nach
Sandzugabe (schwarze Kreise) und nach Kieszugabe (hellgraue Quadrate) verglichen. Die Ordination
basiert auf der Deckung von 117 Pflanzenarten und ist als ‚Joint Plot’ dargestellt, zusammen mit
Umweltgradienten, die hier als schwarze Vektorlinien angezeigt werden (cutoff r² = 0,15). Stresswert
basierend auf zwei Dimensionen: 0,18; Monte-Carlo: p < 0,001. Die Vektorlinien ‚Vegetationshöhe‘
und ‚Median Feuchtewert‘ sowie ‚Abstand zum Ufer‘ und ‚Deckung Neophyten‘ überschneiden sich.
Besides the variable heterogeneity, the three treatments could be distinguished by differ-
ent plant species. There was a significantly higher cover of therophyte species like Ara-
bidopsis arenosa, Arenaria serpyllifolia and Herniaria glabra on sites, where gravel was
added, compared to sites, where sand was added (p < 0.05). In addition, typical riparian plant
species like Clematis vitalba, Humulus lupulus and Solanum dulcamara were recorded on
sites with gravel addition (Fig. 4). Univariate statistics showed no significant difference in
the amount of pioneer target species (p > 0.05). On the NMDS graph (Fig. 4) all species
classified as target species for pioneer habitats were found in the centre of the graph on plots
with gravel and sand input. Sites after restoration by sand input were characterised by the
presence of pioneer species like Alopecurus aequalis, Isolepis setacea and Juncus inflexus
and those after gravel addition by species like Carex hirta, Juncus articulatus and Rumex
obtusifolius.
8

Fig. 4. NMDS ordination based on Bray-Curtis dissimilarity for the comparison of species composition
of 15 plots sampled on three sites with embankment removal (grey crosses), 15 plots sampled on three
sites with sand input (grey circles) and 15 plots sampled on three sites with gravel addition (black
quadrats). It is based on the cover of 117 plant species (stress based on two dimensions: 0.18; Monte-
Carlo: p < 0.001).
Abb. 4. NMDS Ordination basierend auf Bray-Curtis Unähnlichkeiten. Es wird die Artenzusammenset-
zung von je 15 Plots und je drei Standorten nach Uferanbruch (dunkelgraue Kreuze) nach Sandzugabe
(schwarze Kreise) und nach Kieszugabe (hellgraue Quadrate) verglichen. Die Ordination basiert auf der
Deckung von 117 Pflanzenarten. Stresswert basierend auf zwei Dimensionen: 0,18; Monte-Carlo:
p < 0,001.
Abbreviations: Acerpseu: Acer pseudoplatanus; Achimill: Achillea millefollium agg.; Alnuglut: Alnus glutinosa; Alnuinca: Alnus incana;
Alopaequ: Alopecurus aequalis; Arabaren: Arabidopsis arenosa; Arenserp: Arenaria serpyllifolia; Barbstri: Barbarea stricta; Beruerec: Berula
erecta; Betupend: Betula pendula; Bromerec: Bromus erectus; Calaepi g: Calamagrostis epigejos; Calapseu: Calamagrostis pseudophragmites;
Cardimpa: Cardamine impatiens; Cardpers: Carduus personata; Careelat: Carex el ata; Careflac: Carex flacca; Carehirt: Carex hirta; Carenigr:
Carex nigra; Careotru: Carex otrubae; Carepseu: Carex pseudocyperus; Centeryt: Centaurium erythraea; Chaeminu: Chaenorhinum minus;
Cirsarve: Cirsium arvense; Cirsoler: Cirsium oleraceum; Cirspalu: Cirsium palustre; Clemvita: Clematis vitalba; Conycana: Conyza canadensis;
Dactglom: Dactylis glomerata; Desccesp: Deschampsia cespitosa agg.; Echicrus: Echinochloa crus-galli; Eleopalu: Eleochari s palustris; Epilhirs:
Epilobium hirsutum; Epilparv: Epilobium parviflorum; Epiltetr: Epilobium tetragonum; Eq uiarve: Equisetum arvense; Equihyem: Equisetum
hyemale; Equipalu: Equisetum palustre; Erigacri: Erigeron acris; Erigannu: Erigeron annuus; Festrubr: Festuca rubra agg.; Filiulma: Filipendula
ulmaria; Fragvesc: Fragaria vesca; Galianis: Galium anisophyllon; Galimoll: Galium mollugo agg.; Gerarobe: Geranium robertianum; Hedeheli:
Hedera helix; Hernglab: Herniaria glabra; Hierpilo: Hieracium pilosella; Hierpilo: Hieracium piloselloides; Hippvulg: Hippuris vulgaris;
Humulupu: Humulus lupulus; Hypemacu: Hypericum maculatum; Hypeperf: Hypericum perforatum; Impaglan: Impatiens glandulifera; Isolseta:
Isolepis setacea; Juncarti: Juncus articulatus; Juncbulb: Juncus bulbosus; Junceffu: Juncus effusus; Juncinfl: Juncus inflexus; Junctenu: Jun cus
tenuis; Lycoeuro: Lycopus europaeus; Lythsali: Lythrum salicaria; Medilupu: Medicago lupulina; Melioffi: Melilotus officinalis; Mentaqua:
Mentha aquatica; Myosaqua: Myosoton aquaticum; Phalarun: Phalaris arundinacea; Phraaust: Phragmites australis; Planmajo: Plantago major
subsp. major; Poaannu: Poa annua; Poapalu: Poa palustris; Poaprat: Poa pratensis; Poatriv: Poa trivialis s. l.; Popunigr: Populus nigra; Poteanse:
Potentilla anserina; Poteerec: Potentilla erecta; Poteneum: Potentilla neumanniana; Prunvulg: Prunella vulgaris; Ranuacri: Ranunculus acris;
Ranurepe: Ranunculus repens; Reselute: Reseda lutea; Rhinglac: Rhinanthus glacialis; Roripalu: Rorippa palustris; Rubucaes: Rubus caesius;
Rumeobtu: Rumex obtusifolius; Saginodo: Sagina nodosal; Sagiproc: Sagina procumbens; Salialba: Salix alba; Saliauri: Salix aurita; Salicapr:
Salix caprea; Salicine: Salix cinerea s. l.; Salimyrs: Salix myrsinifolia; Salipurp: Salix purpurea; Salirube: Salix rubens; Salivimi: Salix viminalis;
Scronodo: Scrophularia nodosa; Scroumbr: Scrophularia umbrosa; Scutgale: Scutellaria galericulata; Seneaqua: Senecio aquaticus s. str.;
Silepalu: Silene palustris/ latifolia; Soladulc: Solanum dulcamara; Soncaspe: Sonchus asper; Sympoffi: Symphytum officinale s. str.; Taraoffi:
Taraxacum officinale agg.; Trifprat: Trifolium pratense; Tripmari: Tripleurospermum maritimum agg.; Tussfarf: Tussilago farfara; Typhlati:
Typha latifolia; Urtidioi: Urtica dioica s. l.; Verbdens: Verbascum densiflorum; Verbnigr: Verbascum nigrum; Verbthap: Verbascum thapsus;
Veroarve: Veronica arvensis; Verobecc: Veronica beccabunga; Verocate: Veronica catenata; Vicicrac: Vicia cracca s. str.
9

Plots with sand addition showed a significant higher median of moisture value than plots
with gravel input, revealing more moisture indicating species (p < 0.001). This result is
supported by the clear gradient in direction of ‘sand plots’ in the ordination graph (Fig. 3).
Furthermore, the amount of target species for reed stands was significantly higher on these
sites (p < 0.001) in comparison to the other measures. The NMDS ordination (Fig. 4)
showed that those are species like Carex elata, Lythrum salicaria and Phragmites australis.
None of the studied species groups could be exclusively attributed to the heterogeneous plots
after embankment removal. Thus, they were characterized by a significantly lower amount
of phanerophytes (p < 0.05).
There were no significant differences in univariate statistics in regard to the amount of
alien species (p > 0.05). Yet, the NMDS (Fig. 3) showed a distinct gradient of the cover of
alien species in the direction of two ‘gravel plots’ outlying the cluster. These two plots had a
comparatively higher cover (10 and 15%) of Impatiens glandulifera in proportion to their
main vascular plant cover (10 and 20% total cover).
3.4 Effect of restored riverbanks on adjacent reed stands
In order to detect the effects of structural bank diversification measures, 25 plots close to
restored riverbanks were compared to 25 plots in monotonous reed stands of regulated areas.
Although no significant differences in structural parameters could be found, there were
significant differences in species diversity. The species number per plot in restored areas had
a median of seven, which is significantly higher than in non-restored areas (four spp.;
p < 0.001). The evenness also revealed a significant higher diversity in restored areas
(p < 0.001).
In addition to differences in the species diversity, species composition varied significant-
ly between impact and control sites (PERMANOVA: F = 5.98, p < 0.001). The restored
areas were characterised by a significantly higher number of endangered species (p < 0.01)
and target species for reed stands (p < 0.01) like Carex elata, Equisetum palustre, Poa palus-
tris and Valeriana officinalis (Fig. 5). Although alien species (Conyza canadensis, Erigeron
annuus, Impatiens glandulifera) were found on both restored and unrestored sites, the cover
of alien plants was significantly higher on the control site (p < 0.05).
4. Discussion
4.1 Structural diversity
Our results showed that the removal of embankments and the addition of gravel are suit-
able measures to create open, sparsely vegetated areas. Our data revealed a lower vegetation
cover and height on these sites compared to those with addition of sand. It can be supposed
that these bare soil zones cannot be sustained on sites with sand input in absence of regular
natural disturbance or management. A high cover of reed species on these sites indicates fast
succession. This can have two explanations: Firstly, the sand used for creating the artificial
bank originated from the floodplain, so it can be suggested that it is nutrient-rich (FRIESE et
al. 2000) inducing high productivity. Secondly, the sand contains seeds that get activated
during disturbance (cf. HÖLZEL & OTTE 2001); these processes lead to dense and high vege-
tation cover, which is not suitable for riparian pioneer species.
10

Fig. 5. NMDS ordination based on Bray-Curtis dissimilarity for the comparison of species composition
of 25 plots sampled on five sites close to restored sites (black diamonds) and 25 plots sampled on five
non-restored control sites (grey triangles). It is based on the cover of 55 plant species and is visualised
as a joint plot with environmental gradients represented as red vector lines (cutoff r² = 0.15; stress
based on three dimensions: 0.10; Monte-Carlo: p < 0.05).
Abb. 5. NMDS Ordination basierend auf Bray-Curtis Unnähnlichkeiten. Es wird die Artenzusammen-
setzung von 25 Plots von fünf Standorten im direkten Umgriff der Renaturierungsmaßnahmen
(schwarze Rauten) verglichen mit 25 Kontrollflächen von fünf Standorten nicht-renaturierter
Flussabschnitte (graue Dreiecke). Die Ordination basiert auf der Deckung von 55 Pflanzenarten und
wird als ‚Joint Plot’ dargestellt, zusammen mit Umweltgradienten, die hier als rote Vektorlinien
angezeigt werden (cutoff r² = 0,15; Stresswert basierend auf drei Dimensionen: 0,10; Monte-Carlo:
p < 0,05).
11

Two years after restoration, the highly productive and dominant species can only hardly
settle under the circumstances created by gravel addition or embankment removal. The re-
moval of embankment structures leads to the development of instable banks with a dynamic
bare soil surface. It can be assumed that in general plant establishment is difficult due to
these sediment dynamics (WARD et al. 1999). This is shown by the significantly lower
amount of phanerophytes on these sites, for example. Though, even if in terms of habitat
creation these sites provide open spaces, it appears to be less suited for the development of
riparian vegetation.
Gravel introduction represents an intermediate situation with moderate species estab-
lishment leading to environmental conditions suitable for early-successional species, i.e.
nutrient-poor conditions, occasional inundation and frequent drought (VON HEßBERG 2003,
ELLENBERG & LEUSCHNER 2010). Thus, it increases the structural diversification in preal-
pine rivers by the establishment of another more open habitat type. Since gravel bars have
largely disappeared in the past decades, structural restoration measures should focus on their
reestablishment.
While considerable achievements could be obtained directly on measure sites for the
open structure, none of the studied measures revealed significant structural changes in the
adjacent reed zones. Nonetheless, this was not expected, because reed zones are naturally
characterised by a high and homogenous vegetation cover and height (ELLENBERG &
LEUSCHNER 2010). This implicates, that structural parameters are not a suitable indicator for
the monitoring of restoration impacts in these areas.
4.2 Species diversity
The results of our study indicated that sand and gravel input are more suitable measures
for an increase of species diversity than embankment removal. These results are following
the conclusions of JÄHNIG et al. (2009), who found out that gravel bars and loamy habitats
greatly increased species richness. Plots with embankment removal differ greatly in terms of
species diversity compared to the two other treatments. The reason for this might be the
difficult abiotic conditions only few species can deal with.
Concerning the regional diversity, ROSENZWEIG (1995) remarked a positive correlation
between the biotic diversity and habitat heterogeneity. As our results revealed very different
species compositions depending on the measures, we suggest that each measure leads to
another habitat type. In this case, implementing different revitalisation measures can increase
β-diversity. These findings follow the statements of WARD et al. (1999), who noticed that
regional diversity not only depends on the number of species per habitat, but also on the
number of habitats and the turnover between habitats. Compared to gravel input and em-
bankment removal the species composition of sand input is rather homogenous. Therefore, if
the creation of different habitats is defined as a goal and only one type of measure is imple-
mented, the input of sand cannot be recommended. To sum up, sand and gravel input are
most suitable for the enhancement of local species diversity, whereas the embankment re-
moval is more likely to increase the habitat diversity due to more heterogeneous species
compositions on these sites.
Furthermore, the results of this study showed that the implementation of structural diver-
sification measures enhances the species diversity of direct measure sites, but also the one of
the adjacent reed zones. Like JANUSCHKE et al. (2011) we recorded an increased α-diversity
in reed zones close to restoration sites in comparison to control sites. This could indicate
positive restoration effects by an easier colonisation due to a shorter distance to a diverse
12

species pool in restored areas (KAREIVA 1990). Nonetheless, this assumption cannot be ex-
clusively confirmed. Sites for revitalisation measures may have been chosen according to an
existing good conservation potential, whereas the control sites are situated upstream to the
revitalisation measures outside those better-preserved areas. Furthermore, the sites could
have been affected by disturbances during the implementation works and therefore support a
higher diversity. Intermediate disturbance events lead to a temporarily increase in species
diversity, because they create open spaces where other species can colonise, corresponding
to the well-known Intermediate Disturbance Hypothesis (CONNELL 1978).
4.3 Species composition
In our study, species composition greatly differs between all treatments as a result of a
successful creation of different habitat conditions. Moreover, our results showed that differ-
ent types of measures promote different species. The introduction of gravel leads to the
development of riparian pioneer species (therophytes and pioneer target species) shortly after
restoration. These results support the findings of JÄHNIG et al. (2009), who also described
that short-living taxa on gravel bars benefit from floodplain restoration. In floodplains,
where natural sediment deposition patterns are missing, gravel addition is an efficient means
to create open spaces that are suitable for many riparian species (RICHARDSON et al. 2007)
Nonetheless, the vegetation composition after gravel input can vary considerably. In our
results two plots differed greatly from the others due to a particularly high cover of Impati-
ens glandulifera. This may be problematic as invasive species can affect native species
communities if they reach high abundances (RICHARDSON et al. 2007), though HEJDA
& PYŠEK (2006) argued that I. glandulifera does not considerably affect native species; it
merely influences the proportional cover of other dominant species like Urtica dioica. Ac-
cording to this assumption, the occurrence of I. glandulifera could be neglected and thus
rather be seen as one component of a partially new system (RICHARDSON et al. 2007).
Compared to the gravel introduction and embankment removal the input of sand was
more homogenous. Over all studied plots, the species composition was quite similar with
few outliers. In comparison with embankment removal, the level of disturbance was lower
due to more gentle slopes on these plots. Moisture indicating plants as well as pioneer spe-
cies cover was increased on these plots. HÖLZEL & OTTE (2001) suggested that riparian
species increase after seed bank mobilisation due to sediment relocation. Nonetheless, as we
also recorded a high number of phanerophytes and reed species, the succession of these sites
cannot be predicted two years after restoration. In the absence of sediment relocating floods,
some of these sites will develop into riparian forest or into reed stands. The fine sediment
with a high capacity of water retention provides good conditions for germination of riparian
tree species (KARRENBERG et al. 2002), at least when the abundance of competitive herbs is
low. Otherwise, the sites are more likely to support reed stands in the near future, especially
when competitive reed species are found in the direct surroundings. The measure of em-
bankment removal leads to a highly variable species composition. It increases dynamics,
which causes a high variability of habitats and species. Nonetheless, our results also indicate
that this measure is not suitable, if quick establishment of riparian pioneer species or the
development of riparian forest is aimed.
The species composition in the surroundings of restored river sections is positively af-
fected by revitalisation measures. Our results show a higher conservation value in these
areas, which is indicated by a higher amount of endangered species and target species for
reed stands on these plots. This observation can be due to a different species pool in the seed
13

bank or the immediate surroundings of restored and degraded river sections (HÖLZEL
& OTTE (2001). It could also be due to the reflection of methodological problems in the
sampling design. Indeed, plots may differ in other factors than restoration impact when using
a time for space substitution design. Indeed, HEJDA & PYŠEK (2006) discussed similar pat-
terns with the same kind of spatial design.
Finally, one must note, that our results represent only are a ‘snap shot’ of the restoration
state two years after measure implementation. After some years, the conservation value of
sites in the surrounding of restored river sections may diminish again. Sites after embank-
ment removal might flatten, facilitating species establishment on less dynamic habitats.
Succession towards dominant reed stands, indicated by the presence of Phalaris arundina-
cea and Phragmites australis, may also be observed on the now more open sites after gravel
input and embankment removal, as natural disturbances are missing in highly modified
rivers like the Inn. Though, fluvial dynamics are crucial for the maintenance of different
successional stages enhancing the overall species diversity (WARD et al. 1999). Therefore,
JANUSCHKE (2014) argues that a monitoring should not only evaluate, whether the desired
habitat types and species are initially present, but also verify, if they are maintained after a
longer time period. If then undesired vegetation change is observed, adaptive restoration
measures or management have to be considered.
4.4 Perspectives and implications for management
The three studied types of revitalisation measures for riverbank diversification greatly
differ in their development since implementation. For future projects, this means that differ-
ent measures should be accomplished according to the predefined goals. Gravel introduction
leads to open habitat structures where riparian pioneer species can establish. Due to high
nutrient levels and suitable moisture conditions the input of sand causes a comparatively fast
succession into either desirable riparian forest or monotonous reed stands. The removal of
embankment leads to the creation of highly dynamic habitats. This may be difficult for
plants, but can benefit reptiles, ground beetles or kingfishers.
In our study we showed that rare species, pioneer target species and other desirable ripar-
ian plants increase with the creation of open soil areas, which nowadays are also the most
threatened habitats in floodplains (VON HEßBERG 2003). Therefore, management efforts need
to guarantee the conservation of these early successional sites, as there is a lack of dynamics
in highly modified rivers. Our results show that succession is particularly fast on fine sedi-
ments, whereas restored gravel bars or reopened embankment are slower in succession and
can therefore be considered more sustainable. Possible management strategies to counteract
succession are soil-disturbance and biomass reduction, initiated for instance by grazing
(SCHAICH et al. 2010). As outcomes vary considerably depending on the type of measure,
our results emphasise how important careful planning, evaluation of site-specific conditions
and post-implementation monitoring and management can be.
Erweiterte deutsche Zusammenfassung
Einleitung – Fließgewässer und ihre Auen gehören zu den artenreichsten Ökosystemen der Welt
(WARD et al. 1999). Mit der Intensivierung der anthropogenen Nutzung in Flussauen wurden Flüsse
zunehmend verbaut und ihre Gewässerstruktur sowie die natürliche Fließgewässerdynamik beeinträch-
tigt (HABERSACK & PIÉGAY 2008). Dabei sind das Auftreten regelmäßiger Störungen und eine intakte
laterale Konnektivität notwendig für den Artenreichtum der Aue (WARD et al. 1999). Besonders in den
14

stark verbauten Voralpenflüssen wird die Revitalisierung im Sinne einer Strukturverbesserung zur
Dynamisierung der Auen zu einer dringenden Notwendigkeit. Eine Studie von BERNHARDT et al.
(2005) belegt eine starke Zunahme von Renaturierungsprojekten an Flüssen. Standardisierte, langjähri-
ge Monitoringprogramme sind dabei bisher selten (CHAPMAN & UNDERWOOD 2000), aber unbedingt
nötig, um den Erfolg der Maßnahmen bewerten und zukünftige Projekte zielgerichteter durchführen zu
können. Ziel dieser Arbeit ist ein systematischer Vergleich unterschiedlicher Uferstrukturierungsmaß-
nahmen und die Untersuchung des Einflusses dieser Maßnahmen auf die Vegetation. Die abschließende
Bewertung erfolgte anhand von drei in Renaturierungsprojekten häufig verfolgten Zielen: (1) Verbesse-
rung der Vegetationsstruktur, (2) Erhöhung der Artenvielfalt und (3) charakteristische Artenzusammen-
setzung.
Material und Methoden – Das Untersuchungsgebiet liegt an einem stark anthropogen überprägten
Fluss des bayerischen Voralpenlandes, dem Inn südlich und nördlich von Wasserburg. Dort wurden mit
dem Ziel der Dynamisierung des Flusses die Ufer umstrukturiert. Die durchgeführten Maßnahmen
lassen sich in drei Gruppen einteilen: Uferanbruch (Entsteinung, Aufbrechen), Kies- und Sandzugabe.
In einem stratifiziert-randomisierten Aufnahmedesign wurden gezielt die Vegetationsstruktur, die
Deckung der Arten und ausgewählte abiotische Größen (pH-Wert, Uferentfernung, Uferneigung) erho-
ben. Als Straten dienten die vorgefundenen Strukturtypen ‚Röhricht‘ und ‚Pionierfluren‘. Die Daten-
auswertung erfolgte anhand uni- und multivariater Analysen.
Ergebnisse – Alle drei Maßnahmentypen unterschieden sich bezüglich der Vegetationsstruktur. Die
Rohbodendeckung war maximal nach Revitalisierung durch Uferanbruch. Die Vegetationshöhe und -
deckung am Flussufer waren am höchsten nach Zugabe von Sand. Die Probeflächen mit Uferanbruch
wiesen eine starke Streuung der Artenzahl auf, und auch die Diversität der angrenzenden Röhrichte war
erhöht. Alle drei Maßnahmen unterscheiden sich deutlich in ihrer Artenzusammensetzung, wobei erneut
die Maßnahme des Uferanbruchs eine sehr heterogene Entwicklung bewirkte. An Renaturierungsflä-
chen angrenzende Röhrichte zeigten eine veränderte Artenzusammensetzung im Vergleich mit Kon-
trollflächen.
Diskussion – Zwei Jahre nach der Durchführung führten die drei untersuchten Maßnahmen zu gro-
ßen Unterschieden in der Vegetationsentwicklung. Durch die Entfernung der Uferbefestigung entstehen
dynamische Habitate, auf denen sich nur wenige Pflanzen etablieren. Wie auch in der Studie von JÄH-
NIG et al. (2009) führt die Kieszugabe zu einer vergleichsweise offenen Vegetationsstruktur, in der sich
auentypische Pionierarten ansiedeln. Auf Sandflächen stellt sich eine homogene Artenzusammenset-
zung ein; diese lässt sich zumindest teilweise durch die hohe Produktivität der Standorte erklären, die
sich vermutlich entweder in Richtung Röhricht (FRIESE et al. 2000) oder Auwald (KARRENBERG et al.
2002) entwickeln werden, jedenfalls wenn Störungsereignisse ausbleiben. Die beobachteten Entwick-
lungen zeigen die kurzfristige Reaktion der Auepflanzen auf die Uferstrukturierungsmaßnahmen. Für
eine abschließende Beurteilung der Wirksamkeit der Maßnahmen ist jedoch ein längerfristiges Monito-
ring nötig (JANUSCHKE 2014). Wenn die natürliche Dynamik des Flusses nicht ausreicht, müsste über
geeignete Managementstrategien zur Offenhaltung von Pionierstandorten nachgedacht werden.
Perspektiven für die Renaturierung – Die Ergebnisse zeigen, wie unterschiedlich die Wirkungen
einzelner Renaturierungsmaßnahmen sein können. Dies unterstreicht, wie wichtig es ist, vor den Maß-
nahmen abzuwägen, welche Renaturierungsziele unter den gegebenen lokalen Bedingungen verfolgt
werden sollen. Nach Maßnahmendurchführung ist ein Langzeitmonitoring wünschenswert.
15

Acknowledgements
We are grateful to VERBUND Innkraftwerke GmbH and in particular Georg Loy for financial support,
insights into important data, access to the study sites and local support. We thank the landscape plan-
ning consultancies Dr. H. M. Schober (Gesellschaft für Landschaftsarchitektur mbH) and Aquasoli
Ingenieurbüro for providing information about the study site, and Dr. Melanie Müller (TU München,
Chair of Aquatic Systems Biology) for the cooperation, concerning the selection of study sites.
Supplements - im Text fehlt der Verweis???
Additional supporting information may be found in the online version of this article.
Zusätzliche unterstützende Information ist in der Online-Version dieses Artikels zu finden.
Supplement E1. List of species and frequency of occurrence on all surveyed plots.
Anhang E1. Artenliste und Frequenz des Vorkommens auf allen aufgenommenen Flächen.
References
ANDERSON, M.J. (2001): A new method for non-parametric multivariate analysis of variance. – Austr.
Ecol. 26: 32–46.
BARITEAU, L., BOUCHARD, D., GAHNON, G., LEVASSEUR, M., LAPOINTE, S. & BÉRUBÉ, M. (2013): A
riverbank erosion control method with environmental value. – Ecol. Eng. 58: 384–392.
BERNHARDT, E.S., PALMER, M.A., ALLAN, J.D., ALEXANDER, G., BARNAS, K., BROOKS, S., CARR, J.,
CLAYTON, S., DAHM, C., FOLLSTAD-SHAH, J., GALAT, D., GLOSS, S., GOODWIN, P., HART, D.,
HASSETT, B., JENKINSON, R., KATZ, S., KONDOLF, G.M., LAKE, P.S., LAVE, R., MEYER, J.L.,
O’DONNELL, T.K., PAGANO, L., POWELL, B. & SUDDUTH, E.B. (2005): Synthesizing U.S. river res-
toration efforts. – Science 308 (5722): 636–637.
BFN = BUNDESAMT FÜR NATURSCHUTZ (2014): FloraWeb. Neophyten und Kulturpflanzen. – URL:
http://www.floraweb.de/pflanzenarten/neophyten.html [accessed 2014–09–19].
BRAUN-BLANQUET, J. (1964): Pflanzensoziologie, Grundzüge der Vegetationskunde. 3. Aufl. – Sprin-
ger, Wien: 365 pp.
BUTTLER, K.P. & HAND, R. (2008): Liste der Gefäßpflanzen Deutschlands. – Kochia Beih. 2: 1–107.
CAVAILLÉ, P., DOMMANGET, F., DAUMERGUE, N., LOUCOUGARAY, G., SPIEGELBERGER, T., TABAC-
CHI, E. & EVETTE, A. (2013): Biodiversity assessment following a naturality gradient of riverbank
protection structures in French prealps rivers. – Ecol. Eng. 53: 23–30.
CHAPMAN, M.G. & UNDERWOOD, A.J. (2000): The need for a practical scientific protocol to measure
successful restoration. – Wetlands 19: 28–49.
CONNELL, J.H. (1978): Diversity in tropical rainforests and coral reefs. – Sci. Total Environ.
199: 1302–1310.
CONRAD-BRAUNER, M. (1994): Naturnahe Vegetation im Naturschutzgebiet "Unterer Inn" und seiner
Umgebung. Eine vegetationskundlich-ökologische Studie zu den Folgen des Staustufenbaus. – Bay-
erische Akademie für Naturschutz und Landschaftspflege: 175 pp.
DE NOOIJ, R.J.W., VERBERK, W.C.E.P., LENDERS, H.J.R., LEUVEN, R.S.E.W. & NIENHUIS, P.H.
(2006): The importance of hydrodynamics for protected and endangered biodiversity of lowland riv-
ers. – Hydrobiologia 187: 153–162.
ECKELMANN, W., SPONAGEL, H., GROTTENTHALER, W., HARTMANN, K.-J., HARTWICH, R., JANETZKO,
P., JOISTEN, H., KÜHN, D., SABEL, K.-J. & TRAIDL, R. (1996): Bodenkundliche Kartieranleitung. Mit
103 Tabellen. 5. Aufl. – Schweizerbart, Stuttgart: 438 pp.
ELLENBERG, H. & LEUSCHNER, C. (2010): Vegetation Mitteleuropas mit den Alpen in ökologischer,
dynamischer und historischer Sicht. 6. Aufl. – Ulmer, Stuttgart: 1333 pp.
ELLENBERG, H., WEBER, H., DÜLL, R., WIRTH, V., WERNER, W. & PAULIßEN, D. (1992): Zeigerwerte
von Pflanzen in Mitteleuropa. – Scr. Geobot. 18: 1–248.
16

FERNANDES, M.R., AGUIAR, F.C. & FERREIRA, M.T. (2011): Assessing riparian vegetation structure
and the influence of land use using landscape metrics and geostatistical tools. – Landsc. Urban Plan.
99: 166–177.
FERREIRA, M.T., AGUIAR, F.C. & NOGUEIRA, C. (2005): Changes in riparian woods over space and
time: Influence of environment and land use. – Forest Ecol. Manag. 212: 145–159.
FRIESE, K., WITTER, B., MIEHLICH, G. & RODE, M. (Ed.) (2000): Stoffhaushalt von Auenökosystemen.
Böden und Hydrologie, Schadstoffe, Bewertungen. – Springer, Berlin: 434 pp.
GOOGLE INC. (2012): Google Earth. Version 6.2.2.6613.
HABERSACK, H. & PIÉGAY, H. (2008): 27 - River restoration in the Alps and their surroundings: past
experience and future challenges. – In: HABERSACK, H., PIÉGAY, H. & RINALDI, M. (Ed.): Gravel
bed rivers VI. From process understanding to river restoration. Developments in earth surface pro-
cesses: 703–735. Elsevier Science & Technology, Amsterdam.
HEJDA, M. & PYŠEK, P. (2006): What is the impact of Impatiens glandulifera on species diversity of
invaded riparian vegetation? – Biol. Conserv. 132: 143–152.
HÖLZEL, N. & OTTE, A. (2001): The impact of flooding regime on the soil seed bank of flood-
meadows. – J. Veg. Sci. 12: 209–2018.
IBM® (2013): IBM® SPSS® Statistics. Version 22.0.0.0. –.
JÄHNIG, S.C., BRUNZEL, S., GACEK, S., LORENZ, A.W. & HERING, D. (2009): Effects of re-braiding
measures on hydromorphology, floodplain vegetation, ground beetles and benthic invertebrates in
mountain rivers. – J. Appl. Ecol. 46: 406–416.
JANUSCHKE, K. (2014): River restorations. Morphological effects on colonization and succession of
aquatic and riparian organism groups. – PhD. Thesis, Universität Duisburg-Essen: 113 pp.
JANUSCHKE, K., BRUNZEL, S., HAASE, P. & HERING, D. (2011): Effects of river restorations on riparian
mesohabitats, floodplain vegetation and carabid beetles. – Biodiv. Conserv. 20: 3147–3164.
JUNK, W.J., BAYLEY, P.B. & SPARKS, R.E. (1989): The Flood Pulse Concept in river-floodplain sys-
tems. – In: DODGE, D.P. (Ed.): Proceedings of the international Large River Symposium
(LARS): 110–127. Canadian Special Publication of Fisheries and Aquatic Sciences 106.
KAREIVA, P. (1990): Population dynamics in spatially complex environments: theory and data. – Philos.
Trans. Biol. Sci. 330: 175–190.
KARRENBERG, S., EDWARDS, P.J. & KOLLMANN, J. (2002): The life history of Salicaceae living in the
active zone of floodplains. – Freshw. Biol. 47: 733–748.
KOENZEN, U. (2005): Fluss- und Stromauen in Deutschland: Typologie und Leitbilder. Ergebnisse des
F+E-Vorhabens "Typologie und Leitbildentwicklung für Flussauen in der Bundesrepublik Deutsch-
land" des Bundesamtes für Naturschutz: FKZ: 803 82 100. – Bundesamt für Naturschutz, Bonn:
327 pp.
LFU = Landesamt für Umweltschutz (2003): Rote Liste gefährdeter Gefäßpflanzen Bayerns mit regio-
nalisierter Florenliste. – Schriftenr. Bayer. Landesamt Umweltschutz 165: 1–374.
LFU = BAYERISCHES LANDESAMT FÜR UMWELT (2015): Hauptwerte Pegel Wasserburg am
Inn. – URL: http://www.gkd.bayern.de/fluesse/abfluss/stationen/statistik/index.php?thema=gkd&
rubrik=fluesse&produkt=abfluss&gknr=6&msnr=18003004 [accessed 2015–02–03].
LUDWIG, G. & SCHNITTLER, M. (1996): Rote Liste der Pflanzen Deutschlands. 224 pp.
MCCUNE, B. & GRACE, J.B. (2002): Analysis of ecological communities. – MjM Software, Gleneden
Beach, Oregon, U.S.A.
MCCUNE, B. & MEFFORD, M.J. (2011): PC-ORD. Multivariate analysis of ecological data. Version 6.0.
– MjM Software, Gleneden Beach, Oregon, U.S.A.
MEYNEN, E., SCHMITHÜSEN, J., GELLERT, J., NEEF, E., MÜLLER-MINY, H. & SCHULTZE, J.H. (Eds.)
(1962): Naturräumliche Gliederung Deutschlands. Band I. – Selbstverlag, Bad Godesberg: xx pp.
MÜLLER, M., PANDER, J. & GEIST, J. (2014): The ecological value of stream restoration measures: An
evaluation on ecosystem and target species scales. – Ecol. Eng. 62: 129–139.
MÜLLER-DOMBOIS, D. & ELLENBERG, H. (1974): Aims and methods of vegetation ecology. – John
Wiley & Sons, New York: 547 pp.
NEHRING, S., KOWARIK, I., RATITSCH, W.E.F. & ESSL, F. (Eds.) (2013): Naturschutzfachliche Invasivi-
tätsbewertung für die in Deutschland wild lebende gebietsfremde Gefäßpflanzen. – BfN Skripten
352: 1–202.
17

RICHARDSON, D.M., HOLMES, P.M., ESLER, K.J., GALATOWITSCH, S.M., STROMBERG, J.C., KIRKMAN,
S.P., PYŠEK, P. & HOBBS, R.J. (2007): Riparian vegetation: degradation, alien plant invasions, and
restoration prospects. – Divers. Distrib. 13: 126–139.
ROSENZWEIG, M.L. (1995): Species diversity in space and time. – Cambridge University Press, Cam-
bridge: 460 pp.
SCHAICH, H., RUDNER, M. & KONOLD, W. (2010): Short-term impact of river restoration and grazing
on floodplain vegetation in Luxembourg. – Agric. Ecosyst. Environ. 139: 142–149.
SCHOLZ, M., MEHL, D., SCHULZ-ZUNKEL, C., KASPERIDUS, H.D., BORN, W. & HENLE, K. (2012):
Ökosystemfunktionen von Flussauen. Analyse und Bewertung von Hochwasserretention, Nährstoff-
rückhalt, Kohlenstoffvorrat, Treibhausgasemissionen und Habitat-funktion. – Naturschutz Biol.
Vielfalt 124: 1–258.
SHANNON, C.E. (1948): A mathematical theory of communication. – Bell Syst. Tech. J. 27: 379–423,
623–656.
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., UTSCHIK, H. & WEIßBROCK, M.
(2012): Floodplain restoration on the Upper Danube (Germany) by re-establishing water and sedi-
ment dynamics: a scientific monitoring as part of the implementation. – River Syst. 20: 55–70.
TOCKNER, K., SCHIEMER, F., BAUMGARTNER, C., KUM, G., WEIGLAND, E., ZWEIMÜLLER, I. &
WARD, J. (1999): The Danube restoration project: Species diversity patterns River across connec-
tivity gradients in the floodplain system. – Regul. Rivers: Res. Manag. 15: 245–258.
VON HEßBERG, A. (2003): Landschafts- und Vegetationsdynamik entlang renaturierter Flussabschnitte
von Obermain und Rodach. – PhD. Thesis, Universität Bayreuth: 252 pp.
WARD, J.V., TOCKNER, K. & SCHIEMER, F. (1999): Biodiversity of floodplain river ecosystems: eco-
tones and connectivity. – Regul. Rivers: Res. Manag. 15: 125–139.
18