Ecological effects of rehabilitation measures at the Austrian Danube: A meta-analysis of fish assemblages

Abstract

Large rivers are worldwide under severe pressure and there is a lack of information on large river restoration. The present paper represents a meta-analysis of available data on river rehabilitation projects performed at the Austrian Danube River consisting of six rehabilitation projects addressing 19 sites. The overall goal was to analyse the response of fish assemblages to different rehabilitation types based on (1) morphological type (“Instream Habitat Enhancement”, “backwater Enhancement”, “extended Enhancement”), (2) length of rehabilitation measure (3) time after construction (4) applied monitoring design. Biological metrics evaluated included number of fish species and relative density, habitat guilds and Leitbild species. In total, number of species increases by 55% comparing rehabilitated with unrestored sites. The number of species of all habitat guilds is higher after rehabilitation. The proportion of rheophilic species increased and the community evolved toward a more type-specific community, according to the Leitbild. Significant differences between measure types were not detected. The rehabilitation success depends mainly on its spatial extent. Highest positive response of number of rheophilic species is achieved by a length >3.9 km. The results show that habitat rehabilitation of large rivers is effective if the spatial extent of the measure is in accordance with river size.

Full text

1 23
Hydrobiologia
The International Journal of Aquatic
Sciences
ISSN 0018-8158
Volume 729
Number 1
Hydrobiologia (2014) 729:49-60
DOI 10.1007/s10750-013-1511-z
Ecological effects of rehabilitation measures
at the Austrian Danube: a meta-analysis of
fish assemblages
Stefan Schmutz, Helga Kremser,
Andreas Melcher, Mathias Jungwirth,
Susanne Muhar, Herwig Waidbacher &
Gerald Zauner

1 23
Your article is protected by copyright and all
rights are held exclusively by Springer Science
+Business Media Dordrecht. This e-offprint
is for personal use only and shall not be self-
archived in electronic repositories. If you wish
to self-archive your article, please use the
accepted manuscript version for posting on
your own website. You may further deposit
the accepted manuscript version in any
repository, provided it is only made publicly
available 12 months after official publication
or later and provided acknowledgement is
given to the original source of publication
and a link is inserted to the published article
on Springer's website. The link must be
accompanied by the following text: "The final
publication is available at link.springer.com”.

WORLD’S LARGE RIVERS CONFERENCE
Ecological effects of rehabilitation measures at the Austrian
Danube: a meta-analysis of fish assemblages
Stefan Schmutz •Helga Kremser •Andreas Melcher •
Mathias Jungwirth •Susanne Muhar •Herwig Waidbacher •
Gerald Zauner
Received: 25 July 2012 / Accepted: 4 April 2013 / Published online: 14 April 2013
ÓSpringer Science+Business Media Dordrecht 2013
Abstract Large rivers are worldwide under severe
pressure and there is a lack of information on large
river restoration. The present paper represents a meta-
analysis of available data on river rehabilitation
projects performed at the Austrian Danube River
consisting of six rehabilitation projects addressing
19 sites. The overall goal was to analyse the response
of fish assemblages to different rehabilitation
types based on (1) morphological type (‘‘Instream
Habitat Enhancement’’, ‘‘backwater Enhancement’’,
‘‘extended Enhancement’’), (2) length of rehabilitation
measure (3) time after construction (4) applied mon-
itoring design. Biological metrics evaluated included
number of fish species and relative density, habitat
guilds and Leitbild species. In total, number of species
increases by 55% comparing rehabilitated with unre-
stored sites. The number of species of all habitat guilds
is higher after rehabilitation. The proportion of
rheophilic species increased and the community
evolved toward a more type-specific community,
according to the Leitbild. Significant differences
between measure types were not detected. The reha-
bilitation success depends mainly on its spatial extent.
Highest positive response of number of rheophilic
species is achieved by a length [3.9 km. The results
show that habitat rehabilitation of large rivers is
effective if the spatial extent of the measure is in
accordance with river size.
Keywords Rehabilitation Habitat Large rivers 
Fish Monitoring WFD Austria Danube
Introduction
Restoration of anthropogenic modified and heavily
impaired habitats has become a central effort in river
ecology particularly for large floodplain rivers. An
important role plays the restoration of large floodplain
rivers (Schiemer, 1995; Stanford et al., 1996; Tockner
& Schiemer, 1997; Schiemer et al., 1999; Tockner
et al., 1999). Although, large rivers have been
impacted more than any other aquatic ecosystem only
a few rehabilitation projects have been completed
Electronic supplementary material The online version of
this article (doi:10.1007/s10750-013-1511-z) contains
supplementary material, which is available to authorized users.
Guest editors: H. Habersack, S. Muhar & H. Waidbacher /
Impact of human activities on biodiversity of large rivers
S. Schmutz (&)H. Kremser A. Melcher 
M. Jungwirth S. Muhar H. Waidbacher
Institute of Hydrobiology and Aquatic Ecosystem
Management, Department of Water, Atmosphere and
Environment, University of Natural Resources and Life
Sciences, Max-Emanuel-Straße 17, 1180 Vienna, Austria
e-mail: stefan.schmutz@boku.ac.at
G. Zauner
ezb – eberstaller zauner bu
¨ros, Technische Bu
¨ros fu
¨r
Angewandte Gewa
¨ssero
¨kologie, Fischereiwirtschaft,
Kulturtechnik und Wasserwirtschaft, Marktstrasse 53,
4090 Engelhartszell, Austria
123
Hydrobiologia (2014) 729:49–60
DOI 10.1007/s10750-013-1511-z
Author's personal copy

worldwide and even fewer have been evaluated. Due
to the large size and complexity of river/floodplain
systems rehabilitation is expensive and the effects are
difficult to predict.
Ecological integrity of floodplain rivers is mainly
defined by the extent of the intersection between
geomorphology and hydrology, resulting in a reach-
specific level of habitat diversity, and habitat charac-
teristics (Junk et al., 1989; Bayley, 1995; Ward &
Stanford, 1995; Ward, 1998; Schiemer et al., 2000).
Fish are good indicators of ecological integrity in
floodplain rivers since the various guilds integrate a
wide range of riverine conditions from bed sediment
structure for egg development to longitudinal connec-
tivity for spawning migrations (Copp, 1989; Persat
et al., 1995; Schiemer et al., 1991,2000). Studies of
endangered fish species show that both, species pre-
ferring flowing water (rheophilic species) and stagnant
water (limnophilic), are at risk (Lelek, 1987; Schiemer
& Spindler, 1989; Schiemer & Waidbacher, 1992;
Guti, 1995; Wolter et al., 1999; IUCN, 2000; Buijse
et al., 2002).
The reference conditions, so-called Leitbild, are
well-defined for the Austrian Danube River (ADR),
and reveal many deficiencies of the current conditions
compared to those prior to channelization (Hohensin-
ner et al., 2004,2011) when the system was ana-
branched and consisted of more than 90% lotic water
bodies (Fig. 1). Over the long-term, erosion, and
aggradation presumably remained in a dynamic equi-
librium. River channelization primarily led to a
stabilization of the former morphodynamic processes
(Hohensinner et al., 2008). About 80% of the ADR has
been dammed in the second half of the twenties
century. This resulted in an alteration of the active
channel conditions, interruption of the longitudinal
and lateral continuity, and decoupling and loss of
floodplain habitats. Fish populations declined dramat-
ically according to the extended loss of ecological
functioning of the river/floodplain system. Rehabili-
tation of the ADR started in the nineties and covers
both impounded and free flowing river sections.
The present paper represents a meta-analysis of
available data on river rehabilitation projects per-
formed at the ADR. The aim of this paper is to provide
an overview on different rehabilitation types and their
efficiency evaluated by the indicator fish. The follow-
ing hypotheses were: (1) Rehabilitation has a positive
influence on the number of fish species and fish
community structure depending on the morphological
rehabilitation type, length of rehabilitation, and time
lag after construction. (2) The fish community signif-
icantly evolves toward the Leitbild after rehabilitation.
Materials and methods
Source of information, extracted parameters,
and classifications
In total, six monitoring projects were analysed (Rez-
ner, 2001; Zauner et al., 2001,2008; Ginzler, 2002;
Schabus & Reckendorfer, 2006; Keckeis et al., 2007)
describing 19 different sites with rehabilitation pro-
jects (Fig. 2). According to the designations of the
BAW Leitbildkatalog 2011 which refer the reference
conditions, site 1–4 are located in the adapted Leitbild
Danube section ‘‘Oberes Donautal Passau–Aschach’’,
site 5–7 in ‘‘Wachau Melk–Krems’’, and site 8–19 in
‘‘Tullner–Wiener Becken, Krems–Hainburg’’. Due to
the channelization of the Austrian Danube in the
nineteenth century the Danube has a uniform profile
with a river width of 270–300 m in studies areas.
Sampled habitats are located along the shoreline with a
maximum water depth of 2–3 m.
Rehabilitation projects were grouped in three
categories. The rehabilitation type ‘‘Instream Habitat
Enhancement’’ (IHE) refers to structures designed to
increase the variability of habitat parameters (e.g.,
width, depth, velocity) within the river channel (gravel
banks: site 4, 7, 8, and 19, hook groynes: site 2 and 3).
The rehabilitation type ‘‘backwater Enhancement’’
(bE) was used for habitat enhancement implemented
within impoundments (small lentic backwaters: site
12 and 15, small backwaters with low flow velocity:
site 9–11, 13–14, and 16). The rehabilitation type
‘‘extended Enhancement’’ (eE) refers to measures
providing the possibility of dynamic bank develop-
ment (removal of rip rap along the shoreline: site 1, 17,
and 18) including erosion and deposition processes
and/or created/connected side arms (site 6) or oxbow
lakes (site 5) (Fig. 3).
Ecological guilds appear to be good indicators of
the ecological integrity and functioning of river–
floodplain systems (Aarts et al., 2004). Zauner &
Eberstaller (1999) developed a classification scheme
for the Austrian fish fauna that categorizes species
into groups with similar ecological requirements.
50 Hydrobiologia (2014) 729:49–60
123
Author's personal copy

Fig. 1 Original anabranched section of the Danube River in Vienna (Schweickhardt 1830–1846)
Fig. 2 Study area with site IDs. Impounded sections in gray, free flowing sections in black
Hydrobiologia (2014) 729:49–60 51
123
Author's personal copy

Fig. 3 Categories of
rehabilitation types: aIHE
with gravel banks. bbE in
dammed river section in
Vienna. Backwater with
stagnant water. ceE
provides side erosion due to
removal of rip rap and
connects side arms to the
main channel
52 Hydrobiologia (2014) 729:49–60
123
Author's personal copy

Therefore, fish species were grouped into three habitat
guilds, i.e., rheophilic, eurytopic, and limnophilic
according to the EFI?classification. The European
Fish Index is a multimetric index based on a predictive
model that derives reference conditions from abiotic
environmental characteristics of individual sites and
quantifies the deviation between the predicted fish
community (in the ‘‘quasi absence’’ of any human
disturbance) and the observed fish community
(described during a fish sampling occasion). The
metrics used are based on species guilds describing the
main ecological and biological characteristics of the
fish community (http://efi-plus.boku.ac.at/software/
doc/Annexes.pdf). Furthermore, species were classi-
fied according to their Leitbild classification, a clas-
sification referring to the historic abundance, as
defined and used in the Austrian Fish Index—FIA, the
national method for the Water Framework Directive—
WFD (BAW 2011): key species, accompanying spe-
cies, and rare species. We used these metrics to ana-
lyse overall composition of fish assemblages indepen-
dent of habitat requirements, to compare rehabilitation
success with historic conditions and to assess if met-
rics used for the FIA are able to detect rehabilitation
success.
The BACI methodology (before–after–control–
impact) was applied at sites (17–19), before–after
(BA) at sites (1–4) and control–impact (CI) at sites
(5–16). For our analyses we compared B (1–4, 17–19)
or C (5–16) data with data of rehabilitated sites
(Table 1). For data consistency we only used BA data
from studies with BACI methodology. Sampling
efforts were approximately the same before and after
rehabilitation, respectively, at control and rehabilita-
tion sites. Fish sampling methodology was catch-per-
unit-effort. Sampling was done mainly by electrofish-
ing supplemented by riparian trawling and longline
fishing but sampling techniques varied among sites.
Therefore, we used only number of species and the
relative densities of the different habitat guilds and
Leitbild categories as fish metrics. As accurate abun-
dance estimates of the small-sized and bottom-dwell-
ing species bullhead (Cottus gobio L.) are difficult to
obtain, all abundance analyses were done without this
species.
We tested co-linearity of the independent vari-
ables length of rehabilitation [m], time after rehabil-
itation [a], type of rehabilitation (IHE, bE, eE), and
type of monitoring design (CI, BA) with a pair-wise
correlation matrix using Spearman rank correlation.
Non parametric tests with related samples (Wilcoxon
Test) were computed to compare values of number of
species and relative density for habitat guilds and
Leitbild classifications for all sites and each rehabil-
itation type. For further analyses we calculated the
difference of number of species and relative densities
after rehabilitation. Since conventional methods,
such as linear regression analysis, are only suitable
to show main effects and indirect interactions, the
chosen alternative was a Decision Tree procedure
(SPSS PASW
Ò
Decision Trees 18) which is a
graphical non-linear discriminant analysis. The out-
come allows the testing and clear wording of
hypotheses. Classification and Regression Trees
(CRT) splits the data into segments that are as
homogeneous as possible with respect to the depen-
dent variable. A terminal node in which all cases
have the same value for the dependent variable is a
homogeneous, ‘‘pure’’ node. The proportion of
variance explained by the model is:
g2¼1within node variance ðriskÞ
total variance

The total variance is the variance for the dependent
variables before consideration of any independent
variables, which is the variance at the root node (SPSS
PASW
Ò
Decision Trees 18, Chapter 5, p 22).
CRT models were built for two dependent
variables, i.e., change in number of total species
and rheophilic species. The advantage of the CRT
procedure is the ability to figure out the main
effects and the direct interactions. Furthermore, the
procedure processes both categorical and metric
data formats (De’ath & Fabricius, 2000). CRT is
adequate for our data set since the procedure splits
the variables subsequently in two classes instead of
multiple (which is appropriate for bigger data
sets).
Furthermore, non parametric tests (Kruskal–Wallis
Test) were computed to compare rehabilitation types
(IHE, bE, eE) concerning habitat guilds, and Leitbild
classifications. Linear regression models were com-
puted for the length of rehabilitation and time after
rehabilitation concerning habitat guilds, Leitbild clas-
sification, and the parameters number of species and
relative density. The level of significance was
P\0.05. All statistics were calculated using SPSS
PASW.
Hydrobiologia (2014) 729:49–60 53
123
Author's personal copy

Results
General success of rehabilitation
Correlations among independent variables were only
significant between length and time after rehabilitation
(P=0.027, correlation coefficient =-0.507). The
proportion of variance explained by the decision tree
model is 68% for the increase of number of species
after rehabilitation. The incorporated independent
variables were length of rehabilitation, rehabilitation
type, and monitoring design. The first division incor-
porates the variable length of rehabilitation (mean
increase of ten species at a scale [1,000 m and seven
at a scale B1,000 m). Second division incorporates
monitoring design at a scale [1,000 m with a mean
increase of twelve species for CI and six for BA. At a
scale B1,000 m the model incorporates the variable
rehabilitation type and predicts a mean increase of six
species for bE and two species for IHE and eE. The
variable time after rehabilitation was not included in
the model.
In total, median number of species increased about
55%, from 11 before to 17 after rehabilitation. The
maximum increase was an increase of 15 species
found in site 5 and 6 (both from 16 to 31) (Fig. 4).
Median species values of different habitat guilds
increased from five to eight for both, rheophilic
(?26% relative density) and eurytopic (-38% relative
density), and from zero to one for limnophilic species
(Fig. 5). Proportions of relative density of rheophilic
and eurytopic species are more evenly distributed after
rehabilitation respectively at rehabilitated sites com-
pared to control sites (Fig. 5). The relative densities of
rheophilic and eurytopic species are highly dependent
since limnophilic species have very low abundances.
Median species values of Leitbild classification
increased from 4 to 5 for key species (?37% relative
density), from 5 to 7 for accompanying species (-23%
relative density), and from one to three for rare species
(?3% relative density) (Fig. 6). Median number of
exotic species increased from one to two, however,
relative density decreased by 17% (see Appendix in
supplementary materials). The common nase (Chond-
rostroma nasus L.) reached the highest mean increase
of relative density with [20% at all 19 sites.
Table 1 Overview of monitoring study design
Monitoring study Monitoring
design
Number of
treatment sites
Number of
control sites
Number of years
before
Number of
years after
Status
Zauner et al. (2001) BA 4 0 4 6 Impoundment
Zauner et al. (2008) CI 3 1 0 0.5 Free flowing
Rezner (2001) and
Ginzler (2002)
CI 9 3 0 4 Impoundment
Schabus &
Reckendorfer (2006)
BA(CI) 1 (4)
a
1 (2)
a
2 0.5 Free flowing
Keckeis et al. (2007) BA(CI) 2 2 0.5 0.5 Free flowing
a
Cumulative data set in the oxbow-system Orth
Fig. 4 Median (black bar), 25 and 75% percentile (box), and
minimum/maximum (whiskers) of number of species/site at
unrestored (white) and rehabilitated (gray) sites
54 Hydrobiologia (2014) 729:49–60
123
Author's personal copy

Rehabilitation types
The number of species significant increased after
rehabilitation in all rehabilitation types (Wilcoxon
Test, Table 2). The highest number of species
occurred in type eE with 31 species after rehabilita-
tion, followed by IHE with 29 species. The lowest
number of species was observed in bE with 19 species.
Significant differences between unrestored and reha-
bilitated values of number of species and relative
density within rehabilitated types for habitat guilds
and Leitbild classification are shown in Table 2in
bold print (Wilcoxon Test).
Significant differences were only detectable among
rehabilitated types for number of key species (Krus-
kal–Wallis, P=0.021) with the highest increase for
bE (eight sites). The results show that within all habitat
guilds and Leitbild classifications the number of
species increased after rehabilitation in all types.
While the relative rheophilic density increased the one
Fig. 5 Median (black bar), 25 and 75% percentile (box), and
minimum/maximum (whiskers), outliners (open circle) and
extreme values (asterisk) of number of species (left) and relative
density/site (right) at unrestored (white) and rehabilitated sites
(gray) for three habitat guilds (RH rheophilic, EURY eurytopic,
and LIMNO limnophilic)
Fig. 6 Median (black bar), 25 and 75% percentile (box), and
minimum/maximum (whiskers), outliners (open circle) and
extreme values (asterisk) of number of species (left) and relative
density/site (right) at unrestored (white) and rehabilitated sites
(gray) for three Leitbild classifications (KS key species, AS
accompanying species, RS rare species)
Hydrobiologia (2014) 729:49–60 55
123
Author's personal copy

of eurytopic decreased (Fig. 7). No significant differ-
ences were detectable between rehabilitation types.
The highest increase of relative density was reached
by the common nase in IHE and eE, and bleak
(Alburnus alburnus L.) in bE (data not shown).
Effect of length of rehabilitation measure
The length of rehabilitation was highly variable
(average =1.8 km; range =50 m–9.7 km). The
number of rheophilic species (linear regression, P=
0.025, r
2
=0.263, y=0.001x?2.539) and the rel-
ative rheophilic density (linear regression, P=0.042,
r
2
=0.222, y=0.005x?9.037) is significantly pos-
itively correlated with the length of rehabilitation [m].
In contrast, the relative eurytopic density signifi-
cantly decreases with increasing length of rehabil-
itation [m] (linear regression, P=0.043, r
2
=0.22,
y=-0.005x-7.625).
Time effect
The monitoring time varied from a half to 6 years
after rehabilitation (average =3.3 years). Analyses
revealed a significant influence of time on the number
of rare species (linear regression, P=0.039, r
2
=
0.228, y=-0.596x?4.241), relative accompanying
species density (linear regression, P=0.027, r
2
=
0.257, y=-4.579x?6.894), and relative rare spe-
cies density (linear regression, P=0.011, r
2
=0.324,
y=1.039x-1.047). While the number of rare
species and relative accompanying species density
decreased the relative rare species density increased
with time. All other analyses showed no significant
time effect.
Change in number of rheophilic species
after rehabilitation
The proportion of variance explained by the decision
tree model is 75% (g
2
=0.753) for the change in
number of rheophilic species. The incorporated inde-
pendent variables were length of rehabilitation and
rehabilitation type. The mean increase for the number
of rheophilic species after rehabilitation (Fig. 8)is
nine species for a length greater than 3,850 m (Node
2), and five for a length less or equal 3,850 m and
greater than 1,150 m (Node 4). For a length less or
equal 1,150 m the model further includes the
Table 2 pvalues of unrestored versus rehabilitated values for
number of species/relative density, habitat guilds (rheophilic:
RH, eurytopic: EURY, limnophilic: LIMNO), Leitbild classifi-
cation (key species: KS, accompanying species: AS, rare
species: RS), and rehabilitation type (Instream Habitat
Enhancement: IHE, backwater Enhancement: bE, extended
Enhancement: eE) (Wilcoxon test)
n_sp RH EURY LIMNO KS AS RS
All sites (19) .000 .001/.003 .000/.005 .015/.470 .002/.000 .000/.064 .001/.005
IHE (6) .003 .043/.249 .058/.345 .414/.465 .257/.345 .028/.753 .042/.249
bE (8) .000 ––––––
eE (5) .005 .068/.043 .066/.068 .257/.893 .063/.043 .141/.893 .068/.138
Number of sites in brackets. Significant differences are in bold
Fig. 7 Median (black bar), 25 and 75% percentile (box), and
minimum/maximum (whiskers), and outliers (asterisk)of
relative density following three rehabilitation types (IHE
Instream Habitat Enhancement, bE backwater Enhancement,
eE extended Enhancement) and type of habitat guild (white RH
rheophil, light gray EURY eurytop
56 Hydrobiologia (2014) 729:49–60
123
Author's personal copy

independent variable type of rehabilitation and pre-
dicts a mean increase of three species for bE (Node 5).
Discussion
In general, all rehabilitation projects realized at the
ADR showed positive effects for fish diversity and
abundance. The number of species increased by 55%
after rehabilitation. This was a quite consistent pattern
across all sites, only one single site (ID 4) showed a
contrasting trend, i.e., the number of species decreased
by one species. The rehabilitation type at this site was
an ‘‘IHE’’ consisting of a 500 m uniform gravel bank.
The relative abundance of rheophilic species
increased while in consequence, since limnophilic
species in general show very low abundances, eury-
topic species decreased after rehabilitation. Since
species that are highly adapted to riverine conditions
have declined far more than generalist species in
degraded rivers (Aarts et al., 2004) our results
demonstrate that this process can be reversed by river
rehabilitation at least at the local scale.
Our results are in accordance with evaluations of
river–floodplain restoration at the lower River Rhine
by Grift et al. (2001), revealing consistent results
along a gradient of flow and connectivity. They
concluded that the creation of new floodplain waters
with a permanent connection and a constant, moderate
flow probably has the highest potential for supporting
rheophilic fish community in the River Rhine. Fur-
thermore, rehabilitation at the ADR increased the
abundance of key and rare species.
Techniques used to IHE along the ADR, e.g.,
construction of hook groynes and introduction of
gravel bars enhanced fish diversity and abundance and
lead to an increase in Leitbild species. While Roni
et al. (2008) found that firm conclusions for IHE
structures were impossible because of the limited
information provided by the reviewed literature Feld
et al. (2011) found, in accordance with our results, that
fish communities benefit from instream habitat
restoration.
Sites classified as ‘‘bE’’ are located at the Vienna
Danube Island shoreline and are characterized by
urban development, channel straightening and the
construction of a hydroelectric power plant. Due to the
increased habitat diversity, refuges from massive
wave disturbances caused by navigation, and suffi-
cient connection to the main river channel a surpris-
ingly high number of fish species has been attracted by
the constructed habitats. Side arm structures in the
more lenitic central impoundment are mainly impor-
tant for eurytopic species but also host young age
classes of rheophilic species, and provide sufficient
prey fish for predators such as pike (Esox Lucius L.)
and perch (Perca fluviatilis L.) (Chovanec et al.,
2002). Nevertheless, the strong increase in key species
number resulted from the low number of fish species at
the control sites (average number of species at control
sites: bE =9, IHE =13, eE =15); a situation often
Fig. 8 Decision tree for the change in number of rheophilic
species after rehabilitation and the independent variables length
of rehabilitation measure [m], time after restoration [a], type of
rehabilitation (IHE Instream Habitat Enhancement, bE back-
water Enhancement, eE extended Enhancement), and monitor-
ing design (BA before–after, CI control–impact)
Hydrobiologia (2014) 729:49–60 57
123
Author's personal copy

found in heavily degraded, urban rivers (e.g. Wolter &
Vilcinskas, 2000; Paul & Meyer, 2001; Miltner et al.,
2004). Regrettably, the created habitats lose their
function during floods—when the water level of the
impoundment is lowered for flood protection—caus-
ing complete dewatering of the habitats.
The best improvement is achieved by ‘‘eE’’ with the
highest number of fish species after rehabilitation and a
significant increase of both rheophilic and key species
density. eE includes oxbow lakes, side arms and
dynamic river banks and the highest effect of a 60%
increase in rheophilic abundance is achieved in side
arms with permanent flow. Also rare species showed an
increase but only for sites where side arms or oxbow
lakes were connected (ID 5 and 6, data see Appendix in
supplementary materials). Those sites further host the
highest number of Leitbild species after rehabilitation.
Grift et al. (2001) found in the lower River Rhine that
compared to the groyne fields and floodplain lakes that
were already present before floodplain restoration,
secondary channels, and reconnected oxbow lakes
provided habitats that were significantly better suitable
for 0?fish and formed important spawning and
nursery habitats for rheophilic cyprinids.
We found that limnophilic species increased only
marginally and are still rare. In the main channel, they
may even indicate degradation due to river damming.
They are better indicators for measures in the flood-
plains (Grift et al., 2006). Grift et al. (2001) found that
limnophilic species originally dominating in flood-
plains have disappeared after reconnection. It would
appear that limnophilic species would need rehabili-
tation of larger isolated water bodies (Schomaker &
Wolter, 2011). However, the low level of limnophilic
species is consistent with the Leitbild prior to chan-
nelization when the system was anabranched and
consisted of more than 90% of lotic water bodies
(Hohensinner et al., 2004,2011).
The Austrian Danube has still a high potential for
rehabilitation as only long-distant migratory sturgeons
have become extinct in the past. Recolonisation of
rehabilitated sites is therefore possible for a wide
spectrum of species. Stocking of fish is uncommon in
the Danube and mainly related to flag fish species such
as Danube salmon (Hucho hucho). The fish popula-
tions rely on natural reproduction. Therefore, the
significant increase of species in the rehabilitated sites
clearly reflects the still high potential of the Danube
and is unlikely influenced by stocking.
Native species are increasingly considered to be
impaired or even threatened by invasive species in the
Danube River Basin (Paunovic & Csa
´nyi, 2010;
Bloesch et al., 2011). We documented seven exotic
species (out of 38) at unrestored, and eleven exotic
species (out of 49) at rehabilitated sites (Appendix S2
in supplementary materials). This is supported by
Burgess et al. (2012) who postulated that maintaining
floodplain–mainstem river connectivity can also facil-
itate invasions of non-native species to these habitats.
Although the number of non-native species increased
slightly, the relative abundances decreased signifi-
cantly at rehabilitated sites: especially the species
Neogobius melanostomus L. and Neogobius kessleri
L. which occurred in high densities in the rip rap
sections, but almost disappeared in the newly created
habitats.
As expected, the success of rehabilitation increased
with the length of rehabilitation as shown in the
decision trees for the general number of species and
the number of rheophilic species. This is important,
since it implies that no matter which type of measure is
used, the larger the area treated the result. In other
words, there is no difference among the construction
of gravel bars, hook groynes or other instream
structures and the creation/connection of side channels
or oxbows, until the measure reaches a certain size.
Our results show that improvement of habitat for
rheophilic species is achieved by the construction of
gravel banks and hook groynes at a scale of more than
3.8 km and by bE at a scale less than 1.2 km. The latter
is explained by the high habitat diversity at a
comparable small scale and the low number of species
at control sites. However, as mentioned above, bE do
provide only temporal habitats that are eliminated
during floods.
Co-linearity in the data set was tested between all
independent variables and revealed significant but low
correlations only between time and length after
rehabilitation. The other combinations were evenly
distributed. Observed co-linearity did not affect our
model, as correlating variables are not included in the
regression tree model.
There was no visible time effect in our dataset which
is mainly due to the lack of repeated samples. To better
understand the long-term effects of rehabilitation,
more long-term monitoring programmes are needed.
Although the response of the fish communities was
strong after rehabilitation among all rehabilitation
58 Hydrobiologia (2014) 729:49–60
123
Author's personal copy

types, we have to take into account that in areas where
dynamic regeneration processes are lacking, fixed
structures will lose their function by silting up and
becoming de-coupled from the main river unless
regular and cost expensive maintenance measures are
implemented. The overall goal of rehabilitation should
be to establish conditions that enable self-sustaining,
e.g., dynamic processes that provide all relevant
functions of the river–floodplain system in line with
the Leitbild (Muhar et al., 1995,2000; Buijse et al.,
2005).
Acknowledgments We would like to thank E. Lautsch for his
statistical support and T. Buijse and P. Roni for comments on
previous versions. This article was partly supported by WISER,
Water Bodies in Europe: Integrative Systems to Assess
Ecological Status and Recovery (Contract Number 226273),
BIOFRESH, Biodiversity of Freshwater Ecosystems: Status,
Trends, Pressures, and Conservation Priorities (Contract Number
226874), and REFORM, Restoring Rivers for Effective
Catchment Management (Contract Number 282656).
References
Aarts, B. G. W., F. W. B. Van den Brink & P. H. Nienhuis, 2004.
Habitat loss as the main cause of the slow recovery of fish
faunas of regulated large Rivers in Europe: the transversal
floodplain gradient. River research and Applications 20:
3–23.
BAW, 2011. Institut fu
¨r Gewa
¨ssero
¨kologie, Fischereibiologie
und Seenkunde (Hrsg.). Leitbildkatalog mit adaptierten
Leitbildern fu
¨r Salzach, Inn, Donau, Traun, Enns, March,
Mur, Drau und Rhein, sowie fu
¨r Seeausrinne - aktueller
Stand der Bearbeitung Februar 2011. http://www.baw-igf.
at/downloads.
Bayley, P. B., 1995. Understanding large river–floodplain eco-
systems. BioScience 45: 153–158.
Bloesch, J., C. Sandu & J. Janning, 2011. Integrative water
protection and river basin management policy: the Danube
case. River Syst. 20(1-2): 129–143.
Buijse, A. D., H. Coops, M. Staras, L. H. Jans, G. J. van Geest, R.
E. Grift, B. W. Ibelings, W. Oosterberg & C. J. M. Roozen,
2002. Restoration strategies for river floodplains along
large lowland rivers in Europe. Freshwater Biology 47:
889–907.
Buijse, A. D., F. Klijn, R. S. E. W. Leuven, H. Middelkoop, F.
Schiemer, J. H. Thorp & H. P. Wolfert, 2005. Rehabilita-
tion of large rivers: references, achievements and integra-
tion into river management. Archiv fu
¨r Hydrobiologie
Supplement 155 (Large Rivers 15): 715–738.
Burgess, O. T., W. E. Pine III & S. J. Walsh, 2012. Importance of
Floodplain Connectivity to Fish Populations in the Apal-
achicola River, Florida. River Research Application.
Chovanec, A., F. Schiemer, H. Waidbacher & R. Spolwind,
2002. Rehabilitation of a heavily modified river section of
the Danube in Vienna (Austria): biological assessment of
landscape linkages on different scales. International
Review of Hydrobiology 87: 183–195.
Copp, G. H., 1989. The habitat diversity and fish reproductive
function of floodplain ecosystems. Environmental Biology
of Fishes 26: 1–26.
De’ath, G. & K. E. Fabricius, 2000. Classification and regression
trees: a powerful yet simple technique for ecological data
analysis. Ecology 81(11): 3178–3192.
Feld, C. K., S. Birk, D. C. Bradley, D. Hering, J. Kail, A. Marzin,
A. Melcher, D. Nemitz, M. L. Pedersen, F. Pletterbauer, D.
Pont, P. F. Verdonschrot & N. Friberg, 2011. From natural
to degraded rivers and back again: a test of restoration
ecology theory and practice. Advances in Ecological
Research 44: 119–209.
Ginzler, B., 2002. Fischo
¨kologische und morphologische Ver-
ha
¨ltnisse im Einflussbereich des Kraftwerks Wien/
Freudenau unter besonderer Beru
¨cksichtigung der neu
geschaffenen Uferstrukturen am linken Donauufer im
Bereich Donauinsel. Diplomarbeit, Wien, Universita
¨tfu
¨r
Bodenkultur Wien.
Grift, R. E., A. D. Buijse, W. L. T. van Densen & J. G. P. Klein
Breteler. 2001. Restoration of the river–floodplain inter-
action: benefits for the fish community in the River Rhine.
Archiv fu
¨r Hydrobiologie 135(2–4) (Large Rivers
12(2–4)): 173–182.
Grift, R. E., A. D. Buijse & G. J. van Geest, 2006. The status of
limnophilic fish and the need for conservationin floodplains
along the lower Rhine, a large regulated river. Archiv fu
¨r
Hydrobiologie Supplement 158 (Large Rivers 16): 623–648.
Guti, G., 1995. Conservation status of fishes in Hungary.
Opuscula Zoologica Budapest 10: XXVII–XXVIII.
Hohensinner, S., H. Habersack, M. Jungwirth & G. Zauner, 2004.
Reconstruction of the characteristics of a natural alluvial
river–floodplain system and hydromorphological changes
following human modifications: the Danube River
(1812–1991). River Research and Applications 20(1): 5–41.
Hohensinner, S., D. Eberstaller-Fleischhanderl, G. Haidvogl, M.
Herrnegger & M. Weiss, 2008. Die Stadt und der Strom –
Historische Vera
¨nderungen der Wiener Donau-Auen seit
dem 18. Jahrhundert. Abhandlungen der Geologischen
Bundesanstalt 62: 87–93.
Hohensinner, S., M. Jungwirth, S. Muhar & S. Schmutz, 2011.
Spatio-temporal habitat dynamics in a changing Danube
River landscape 1812–2006. River Research and Appli-
cations 27: 939–955.
IUCN, 2000. The 2000 IUCN Red Lists of Threatened Species.
International Union for the Conservation of Nature, Gland.
www.redlist.org.
Junk, W. J., P. B. Bayley & R. E. Sparks, 1989. The flood-pulse
concept in river floodplain systems. Canadian Special
Publication of Fisheries and Aquatic Sciences 106:
110–127.
Keckeis, H., E. Schludermann, V. Bammer & S. Go
¨tsch, 2007.
Projekt Revitalisierung Donauufer. Fischo
¨kologie. Uni-
versita
¨t Wien, Endbericht: 55 pp.
Lelek, A., 1987. The Freshwater Fishes of Europe, 9. Threa-
tened Fishes of Europe. Aula-Verlag, Wiesbaden: 343 pp.
Miltner, R. J., D. White & C. Yoder, 2004. The biotic integrity
of stream in urban and suburbanizing landscapes. Land-
scape and Urban Planning 69: 87–100.
Hydrobiologia (2014) 729:49–60 59
123
Author's personal copy

Muhar, S., S. Schmutz & M. Jungwirth, 1995. River restoration
– goals and perspectives. Hydrobiologia 303: 183–194.
Muhar, S., S. Schwarz, S. Schmutz & M. Jungwirth, 2000.
Identification of rivers with high and good habitat integrity:
methodological approach and applications in Austria.
Hydrobiologia 422(423): 343–358.
PASW
Ò
Decision Trees 18, Copyright 1993–2007.
Paul, M. J. & J. L. Meyer, 2001. Streams in the urban landscape.
Annual Reviews in Ecology and Systematics 32: 333–365.
Paunovic
´, M. & B. Csa
´nyi, 2010. Invasive aquatic species (IAS)
as significant water management issue for the Danube
River Basin. Guidance document on Alien Invasive Spe-
cies within DRB, first draft. 29/05/2010.
Persat, H., J. M. Olivier & J. P. Bravard, 1995. Stream and
riparian management of large braided Mid-European riv-
ers, and consequences for fish. In Armantrout, N. N. (ed.),
Condition of the World’s Aquatic Habitats: Proceedings of
the World Fisheries Congress, Theme 1. Oxford & IBH
Publishing Co, New Delhi: 139–169.
Rezner, C., 2001. Fischo
¨kologische Verha
¨ltnisse im Einfluss-
bereich des Kraftwerkes Freudenau unter besonderer
Beru
¨cksichtigung unterschiedlicher Habitattypen. Diplo-
marbeit, Universita
¨tfu
¨r Bodenkultur Wien, Wien.
Roni, P., K. Hanson & T. Beechie, 2008. Global review of the
physical and biological effectiveness of stream habitat
rehabilitation techniques. North American Journal of
Fisheries Management 28: 856–890.
Schabus, M. & W. Reckendorfer, 2006. Einfluß der Gew-
a
¨sservernetzungsmaßnahmen auf die Adult- und Jung-
fischfauna im Altarmsystem von Orth an der Donau.
Wissenschaftliche Reihe Nationalpark Donauauen. Heft
13(2006): 26 pp.
Schiemer, F., 1995. Revitalisierungsmaßnahmen fu
¨r Augewa
¨s-
ser – Mo
¨glichkeiten und Grenzen. Archiv fu
¨r Hydrobiol-
ogie Supplement 101: 163–178.
Schiemer, F. & T. Spindler, 1989. Endangered fish species of the
Danube River in Austria. Regulated Rivers: Research and
Management 4: 397–407.
Schiemer, F. & H. Waidbacher, 1992. Strategies for conserva-
tion of a Danubian fish fauna. In Boon, P. J., P. Calow & G.
E. Petts (eds), River Conservation and Management.
Wiley, Chichester: 363–382.
Schiemer, F., T. Spindler, H. Wintersberger, A. Schneider & A.
Chovanec, 1991. Fish fry associations: important indica-
tors for the ecological status of large rivers. International
Verein Limnology 24: 2497–2500.
Schiemer, F., C. Baumgartner & K. Tockner, 1999. Restoration
of floodplain rivers: the ‘‘Danube Restoration Project’’.
Regulated Rivers: Research and Management 115:
231–244.
Schiemer, F., M. Jungwirth, S. Muhar & S. Schmutz (eds), 2000.
Fish as indicators for the assessment of ecological integrity
of large rivers. Hydrobiologia 422/423: 271–278.
Schomaker, C. & C. Wolter, 2011. The contribution of long-
term isolated water bodies to floodplain fish diversity.
Freshwater Biology 56: 1469–1480.
Stanford, J. A., J. V. Ward, W. J. Liss, C. A. Frissell, R.
N. Williams, J. A. Lichatowich & C. C. Coutant, 1996. A
general protocol for restoration of regulated rivers. Regu-
lated Rivers: Research and Management 12: 391–413.
Tockner, K. & F. Schiemer, 1997. Ecological aspects of the
restoration strategy for a river–flood-plain system of the
Danube River in Austria. Global Ecological and Biogeo-
graphical Letters 6: 321–329.
Tockner, K., D. Pennetzdorfer, N. Reiner, F. Schiemer & J.
V. Ward, 1999. Hydrological connectivity and the
exchange of organic matter and nutrients in a dynamic
river–floodplain system (Danube, Austria). Freshwater
Biology 41: 521–535.
Ward, J. V., 1998. Riverine landscapes: biodiversity patterns,
disturbance regimes, and aquatic conservation. Biological
Conservation 83: 269–278.
Ward, J. V. & J. A. Stanford, 1995. The serial discontinuity
concept: extending the model to floodplain rivers. Regu-
lated Rivers: Research and Management 10: 159–168.
Wolter, C. & A. Vilcinskas, 2000. Charakterisierung der Fis-
chartendiversita
¨t in Wasserstraßen und urbanen Gewa
¨ss-
ern. Wasser und Boden 52: 14–18.
Wolter, C., A. Bischoff, M. Tautenhahn & A. Vilcinskas, 1999.
Die Fischfauna des unteren Odertals: Arteninventar,
Abundanzen, Bestandsentwicklung und fischo
¨kologische
Bedeutung der Polderfla
¨chen. In Dohle, W., R. Bornkamm
& G. Weigmann (eds), Das Untere Odertal. Limnologie
Aktuell, Stuttgart 9: 369–386.
Zauner, G. & J. Eberstaller, 1999. Klassifizierungsschema der
o
¨sterreichischen Flußfischfauna in bezug auf deren Leb-
ensraumanspru
¨che. O
¨sterreichs Fischerei 52: 198–205.
Zauner, G., P. Pinka & O. Moog, 2001. Pilotstudie Oberes
Donautal – gewa
¨ssero
¨kologische Evaluierung neuges-
chaffener Schotterstrukturen im Stauwurzelbereich des
Kraftwerks Aschach. Im Auftrag des Bundesministeriums
fu
¨r Verkehr, Innovation und Technologie, Republik
O
¨sterreich.
Zauner, G., C. Ratschan & M. Mu
¨hlbauer, 2008. Life Natur
Projekt Wachau. Endbericht Fischo
¨kologie. I. A. Arbe-
itskreis Wachau & Via Donau.
60 Hydrobiologia (2014) 729:49–60
123
Author's personal copy