Functional trait composition of aquatic plants can serve to disentangle
multiple interacting stressors in lowland streams
⁎, Emma Göthe
, Matthew T. O'Hare
Department of Bioscience, Aarhus University, Vejlsøvej 25, P.O. Box 314, DK-8600 Silkeborg, Denmark
Department of Bioscience, Aarhus University, Ole Worms Allé 1, Building 1135, Room 217, DK-8000 Aarhus C, Denmark
Centre for Ecology and Hydrology, Bush Estate, Penicuik EH26 0QB, United Kingdom
•Functional trait composition of aquatic
plants can distinguish hydromorpholog-
ical degradation from eutrophication in
•A conceptual framework on how
eutrophication and hydromorphological
degradation interact on functional trait
•Weed cutting can set aside light as a
factor controlling trait-abundance pat-
tern in eutrophic lowland streams.
Received 22 September 2015
Received in revised form 5 November 2015
Accepted 5 November 2015
Available online xxxx
Editor: D. Barcelo
Historically, close attention has been paid to negative impacts associated with nutrient loads to streamsand riv-
ers, but today hydromorphological alterations are considered increasingly implicated when lowland streams do
not achieve good ecological status. Here, we explore if trait-abundance patterns of aquatic plants change along
gradients in hydromorphological degradation and eutrophication in lowland stream sites located in Denmark.
Speciﬁcally, we hypothesised that: i) changes in trait-abundance patterns occur along gradients in
hydromorphologicaldegradation and ii) trait-abundancepatterns can serve to disentangle effects of eutrophica-
tion and hydromorphological degradation in lowland streams reﬂecting that the mechanisms behind changes
differ. We used monitoring data from a total of 147 stream reaches with combined data on aquatic plant species
abundance, catchmentland use, hydromorphological alterations (i.e. planform, cross section, weed cutting) and
water chemistry parameters. Traitsrelated to life form, dispersal, reproduction and survival together with ecolog-
ical preference valuesfor nutrients and light (Ellenberg N and L) were allocated to41 species representing 79% of
the total species pool.We found clear evidence that habitat degradation (hydromorphologicalalterations and eu-
trophication) mediated selective changes in the trait-abundance patterns of the plant community. Speciﬁctraits
could distinguish hydromorphological degradation (free-ﬂoating, surface; anchored ﬂoating leaves; anchored
heterophylly) from eutrophication (free-ﬂoating, submerged; leaf area). We provide a conceptual framework
for interpretation of how eutrophication and hydromorphological degradation interact and how this is reﬂected
Science of the Total Environment 543 (2016) 230–238
E-mail address: email@example.com (A. Baattrup-Pedersen).
0048-9697/© 2015 Elsevier B.V. All rights reserved.
Contents lists available at ScienceDirect
Science of the Total Environment
journal homepage: www.elsevier.com/locate/scitotenv
in trait-abundance patterns in aquatic plant communities in lowland streams. Our ﬁndings support the merit of
trait-based approaches in biomonitoring as they shed light on mechanisms controlling structural changes under
environmental stress. The ability to disentangle several stressors is particularly important in lowland stream en-
vironments where several stressors act in concert sincethe impact of the most important stressor can betargeted
ﬁrst, which is essential to improve the ecological status.
© 2015 Elsevier B.V. All rights reserved.
Today, anthropogenic pressures related to agriculture are one of the
main drivers of ecological deterioration of stream and river ecosystems,
primarily through emissions of nitrogen and phosphorous, increased
sediment load and hydromorphological alterations (Vörösmarty et al.,
2010). Historically, close attention has been paidto negative impacts as-
sociated with nutrient loads to streams and rivers, but today
hydromorphological alterations are considered increasingly implicated
when lowland streams do not achieve good ecological status (EEA,
2012). Even though the importance of hydromorphological degradation
is accepted as a major stressor, the ability to assess the level of
hydromorphological impact on the biological communities is limited
(e.g. Vaughan et al., 2009; Feld et al., 2014), and there is a clear need
for improving our conceptual understanding of the underlying response
mechanisms. One reason for the current limited knowledge could be
that the high level of spatial and temporal variability characterising
stream and river habitats makes it difﬁcult to assess the hydromorpho-
logical impact at a scale relevant for the biological communities. At the
reach scale, the biota responds to local hydromorphological features
(i.e. the interaction between the ﬂow of water and the channel form),
but, additionally, disturbances occurring at larger spatial scales (stream
network) (Poff, 1997) and even historical disturbances (Harding et al.,
1998) can mask the effect of local factors on species composition
(Poff, 1997; Kail and Wolter, 2013).
The majority of studies investigating the effects of hydromorpho-
logical degradation on biological communities have focused on species
richness and/or multivariate descriptors of species composition (e.g.
Hering et al., 2006; Dahm et al., 2013, but see also Feld et al., 2014,
Elosegi and Sabater, 2013, and references therein). However, the taxo-
nomic composition may differ between regions due to spatial con-
straints on community assemblies, making compositional approaches
vulnerable to scale-dependent processes. Functional community char-
acteristics have been suggested as an alternative or complement to
compositional characteristics. Because the same traits (responding to
similar environmental conditions) can be applied to most species in
the world, functional composition is thought to be less vulnerable to
scale-dependent processes (e.g. Dolédec et al., 2006; Friberg et al.,
2011) than taxonomic composition. Additionally, traits provide a
means to gain insight into the mechanisms mediating the response to
natural and anthropogenic drivers of change (Diaz et al., 2007; Moretti
and Legg, 2009).
Functional trait composition has recently proven useful to assess ef-
fects of eutrophication on aquatic plant communities in European low-
land streams (Baattrup-Pedersen et al., in press). Clear indications exist
that eutrophication promotes species that efﬁciently capture light by
concentrating their photosynthetic active biomass near the water sur-
face and species that utilise light efﬁciently. The mechanism behind
these changes was suggested to be intensiﬁed competition due to en-
hanced aquatic plant growth, bioﬁlm development and more turbid wa-
ters under nutrient-rich conditions. Here, we explore whether
functional trait composition of the aquatic plant community can be
used as a means to assess hydromorphological degradation as well. Cur-
rently, there is no comprehensive theory on how aquatic vegetation re-
sponds to hydromorphological degradation, but from ecological niche
theory we expect that signiﬁcant changes occur (Southwood, 1988).
As aquatic vegetation is known to exhibit preference for and adaptation
to substrates, current velocities and depths (Dawson et al., 1999;
Baattrup-Pedersen and Riis, 1999; Gurnell et al., 2010; Puijalon et al.,
2011), we expect that channelisation (e.g. deepened, widened and
straightened; Brookes and Gregory, 1983; Brookes, 1987; Mattingly
et al., 1993; Verdonschot and Nijboer, 2002; Landwehr and Rhoads,
2003) will restrict the niches available for aquatic plants. In particular
loss of pool–rifﬂe sequences in channelised streams implies that the
habitats that predominate are generally deeper and ﬂow velocities
higher and, at the same time, ﬂow patterns, substrate conditions and
depth characteristics get more uniform (Baattrup-Pedersen and Riis,
1999; Rambaud et al., 2009). Consequently, channelisation likely affects
the presence and distribution of different plant life forms. For example,
the abundance of submerged species is likely to be higher in
channelised streams because these species may bend, thereby mitigat-
ing the increase in drag at higher velocities compared to emergent spe-
cies (Brewer and Parker, 1990; Schutten and Davy, 2000). Furthermore,
species that have their biomass distributed evenly in the water column
experience lower drag than species that have their biomass in high-
velocity areas near the surface (Bal et al., 2011).
Active maintenance of the channelised stream proﬁle by dredging
and mechanical removal of the vegetation by cutting, as performed reg-
ularly in many lowland streams today (for instance Fox and Murphy,
1990; Kaenel and Uehlinger, 1999; Vereecken et al., 2006; Baattrup-
Pedersen et al., 2009; Wiegleb et al., 2014), may also affect the func-
tional traits composition of the aquatic plant community. Dredging is
a very dramatic form of disturbance that can remove all vegetation
and reset the community (Wade and Edwards, 1980; Wade, 1993). In
cases where the channel bed is only superﬁcially scrapped,
overwintering propagules may be left in place, whereas a more
profound removal of the bottom sediments may affect also the
overwintering organs. In regularly dredged streams we may therefore
expect that traits related to species dispersal and establishment may
be of overriding importance for community composition. Effects of
weed cutting has particularly been studied in cases where single species
have been implicated in blocking channels (Dawson, 1989; Dawson,
1976; Pitlo and Dawson, 1990; Kern Hansen and Dawson, 1978),
whereas studies on the effects of cutting on community composition
are scarce and geographically restricted (Baattrup-Pedersen et al.,
2002; Baattrup-Pedersen et al., 2003; Baattrup-Pedersen and Riis,
2004). However, weed cutting can be compared to grazing and
expectedly has a similar selective pressure favouring species that re-
cover fast, i.e. those with intact growth meristems following cutting
(Baattrup-Pedersen et al., 2002; Wood et al., 2012).
To investigate how hydromorphological degradation and eutrophi-
cation affect plant trait composition we have examined how traits
related to morphology, dispersal, survival and the ecological prefer-
ences of thespecies vary along gradients in the impact of these stressors
in lowland streams in Denmark. Speciﬁcally, we hypothesised that:
i) changes in trait-abundance patterns occur along gradients in
hydromorphological degradation and ii) trait-abundance patterns can
be used to identify the main stressor in lowland streams, for instance
high abundance of species that efﬁciently capture or utilise light indi-
cates that the main stressor is related to eutrophication (see Baattrup-
Pedersen et al., in press), whereas these traits will be subordinate in
streams where the main stressor is related to hydromorphological deg-
radation, reﬂecting that weed cutting improves light availability and
therefore the light climate in eutrophic streams.
231A. Baattrup-Pedersen et al. / Science of the Total Environment 543 (2016) 230–238
2.1. Aquatic plants and environmental data
The data used were obtained from the Danish monitoring program
(2004–2007; NOVANA; Friberg et al., 2005). From a total of 147 stream
reaches on which we possessed combined data on aquatic plant cover-
age, catchment and buffer strip land use, hydromorphological alter-
ations (i.e. cross section, planform, weed cutting) and water chemistry
parameters. These 147 sites were all categorised as middle-sized
(types 2 and 3; Baattrup-Pedersen et al., 2004) with a catchment area
larger than 10 km
; they were distributed throughout Denmark and
covered existing gradients in alkalinity and catchment land use. Aquatic
plant data were collected following the protocol described in Pedersen
et al. (2007). In each stream reach, plant recordings were made in
July/August at maximum biomass. Recordings were made in approxi-
mately 150 plots (25 × 25 cm) placed side by side in cross-sectional
transects at a 100 m long stream reach. Depending on the width of the
stream, the number of transects varied from a minimum of 10 to 20 in
small streams. A cover score was allocated to each species present in
the plots using the following abundance scale: 1 = 1–5%, 2 = 6–25%,
3=26–50%, 4 = 51–75%, and 5 = 76–100%. Species abundance at
each stream reach was then calculated as the sum of cover scores to
the maximum score sum (i.e. the number of plots multiplied by the
maximum score of ﬁve; Pedersen et al., 2006).
The catchments upstream of the stream reaches and in a 50 m wide
buffer were delineated using the Analysis Tools in ESRI ArcGIS 9.2. Agri-
cultural land use wasthen determined from a national land coverraster
map (25 m grid) with 12 land cover classes (Nielsen et al., 2000) and
from mandatory annual reports on land use from all farmers to the Dan-
ish Ministry of Agriculture (DFFE, 2008). The latter source contains in-
formation on ﬁeld location and crop type. Only land cover polygons
classiﬁed as arable land were allocated to agriculture.
Hydromorphological features of the stream sites (i.e. cross section
and planform) were recorded at the time of the aquatic plant sampling,
whereas information on weed cutting practice was obtained from the
water authorities. Cross section was categorised as either natural or
channelised using proﬁle characteristics (Table 1). Channelised cross
sections had a trapezoid form with similar depths across the channel
proﬁle, whereas natural cross sections exhibited variability in depth
characteristics. Furthermore, contrary to natural cross sections,
channelised cross sectionswere deeply positioned compared to the sur-
rounding land. To assess sinuosity, stream shape was recorded in the
ﬁeld and data were gathered from aerial photos. The following catego-
ries were used: channelised planform, sinuous planform, straight (nat-
ural) planform and meandering planform (Table 1). Weed cutting was
categorised according to the extent of the cutting of the channel as ei-
ther ‘Nhalf width’, which included reaches cut from their full width to
half the width of the cross section, or ‘net-half width’, which included
reaches cut up to half their width, and ‘none’, which included stream
reaches that were left uncut. The number of stream sites within each
of the three categories is given in Table 1. Water chemistry data used
for the analyses were based on ﬁve yearly samplings conducted within
5 years of the vegetation surveys. Since water chemistry has been rather
constant in Danish streams over the last years we ﬁnd this approach ap-
propriate (Wiberg-Larsen et al., 2013). The water samples were
analysed for phosphate (PO
P), nitrate (NO
N) and ammonium
N) in the laboratory according to European standards.
2.2. Description of traits
A total of 52 submerged and amphibious taxa were observed of
which we were able to allocate traits to 41 species representing 79% of
the total submerged and amphibious species pool. We covered traits
that we believe respond to eutrophication and hydromorphological
degradation (see Introduction). These encompassed morphological
traits including life forms and traits important for species dispersal, re-
production and survival. We also included ecological preference values
(Ellenberg N and L; Table 2). The Ellenberg indicator values (Ellenberg
et al., 1991) offer autecological information on the response of approx-
imately2000 species to a rangeof climatic and edaphic factorsin Central
Europe. EN and EL have recently proven useful in combination with
plant traits to identify the biological mechanism behind changes in
community composition in response to eutrophication (Baattrup-
Pedersen et al., in press). Trait data were extracted from the literature
and online databases (Willby et al., 2000;Table 2). The life forms (LF)
were divided into six categories: free-ﬂoating (surface and submerged),
anchored with ﬂoating and submerged leaves, and anchored amphibi-
ous species with homophyllus emergent leaves and heterophyllus
emergent leaves. Growth morphology was divided into three catego-
ries: single basal, single apical and multi-apical (Table 2). Plant morpho-
logical traits also included a morphology index building on the height
and lateral extension of the canopy and the leaf area of the species. Dis-
persal was characterised by four traits: the abilityto disperse by forming
extensive root–rhizome systems, the ability to reproduce by fragmenta-
tion, the number of seeds and the number of reproductive organs pro-
duced by the species. We also integrated traits related to survival in
terms of overwintering organs such as tubers, turions and rhizomes.
Overviewof number of stream reaches categorised into hydromorphological groups.More
information on how the streams were categorised is given in the methods section.
Variable Category N %
Channel planform PL_Channelised 27 18.4
PL_Meandering 42 28.6
PL_Sinuous 55 37.4
PL_Straight, natural 23 15.6
Cross section CS_Channelised 52 35.4
CS_Natural 95 64.6
Weed cutting intensity Weed Nhalfwidth 68 46.3
Weed bhalfwidth 22 15.0
Weed_none 57 38.8
The 18 functional traitsused in the present study to characterisethe plant species. The se-
lected traits give information on ecological preference (Ellenberg Light and Ellenberg Ni-
trogen), life form, mo rphology (merist em characteristics; leaf area; canopy
characteristics), dispersal (root–rhizome growth; fragmentation; seed production) and
survival (overwintering organs). See text for further explanation.
Short trait name Explanation Category
LE Ellenberg Light Ecological
NE Ellenberg Nitrogen Ecological
Frﬂsr Free ﬂoating, surface Life form
Frﬂsb Free ﬂoating, submerged Life form
Anﬂle Anchored, ﬂoating leaves Life form
Ansule Anchored, submerged leaves Life form
Anemle Anchored, emergent leaves Life form
Anhete Anchored, heterophylly Life form
Meris.ma Meristem multiple apical growth point Morphology
Meris.sb Meristem single basal growth point Morphology
Meris.sa Meristem single apical growth point Morphology
Morph.ind Morphology index = (height + lateral
extension of the canopy) / 2
Leaf.area Leaf area Morphology
Seeds Reproduction by seeds Dispersal
Rhizome Reproduction by rhizomes Dispersal
Frag Reproduction by fragmentation Dispersal
N.rep.org Number of reproductive organs per year and
Overwintering.org Overwintering organs Survival
232 A. Baattrup-Pedersen et al. / Science of the Total Environment 543 (2016) 230–238
The life form traits, and traits covering fragmentation, seeds,
overwintering organs and rhizomes, were based on presence/absence
of the attribute, with a score of 0 for absence, 1 for occasionally but
not generally present attributes and 2 for present attributes. The mor-
phology traits describing the meristem growth point type were based
on presence (1) or absence (0) of the attribute. The number of repro-
ductive organs was classiﬁed into low (b10), medium (10–100), high
(100–1000) and very high (N1000), with values ranging from 1 to 4
based on number per individual per year. Leaf area was classiﬁed ac-
cording to the leaf size categories with values ranging from 1 to 4,
representing small (b1cm
), medium (1–20 cm
), large (20–
) and very large (N100 cm
). The morphology index was also
classiﬁed into categories (2, 3–5, 6–7, 8–9 and 10) with values ranging
from 1 to 5. In some cases species were classiﬁed in-between two cate-
gories regarding the number of reproductive organs, leaf area and mor-
phology index (Willby et al., 2000). In these cases a classiﬁcation code
in-between was allocated to the particular trait (i.e. 1.5, 2.5, 3.5 and 4.5).
2.3. Data analysis
The relationships between species abundance and trait variables
were examined using multivariate ordination techniques. Stream typol-
ogy did not signiﬁcantly explain variation in trait composition (RDA
ANOVA; p N0.05) and we therefore analysed sites from the two typolo-
gies together. We used RLQ and fourth-corner analyses to assess the co-
variation between environmental variables (hydromorphological and
nutrient variables) and traits. RLQ analysis is an extension of co-inertia
analysis that provides an overview of the multivariate associations by
searching for a combination of traits (Table Q) and environmental vari-
ables (Table R) with maximum covariance, which is weighted by the
abundance of the species in the plots (Table L) (Dolédec et al., 1996).
First, correspondence analysis (CA) was applied to Table L, principle
component analysis was applied to Table Q and Hill and Smith analysis
(Hill and Smith, 1976) was applied to Table R as it contained a mix of
qualitative (hydromorphological variables) and quantitative (water
chemistry variables) variables. RLQ analysis was then applied, combin-
ing the three separate ordinations and identifying the main associations
between Tables R and Q, linked by Table L. In the RLQ analysis, the site
scores in Table R constrain the site scores in Table L, and the species
scores in Table Q constrain the species scores in Table L. The axis that
maximises the covariance in Table L is then selected, resulting in a com-
promise between thebest joint combination of site scores by their envi-
ronmental characteristics, the best combination of species scores by
their trait attributes and the simultaneous ordination of sites and scores.
The overall signiﬁcance of the relationship between the environmental
variables (R) and species traits(Q) was assessed with a Monte Carlo test
with 999 permutations on total inertia of the RLQ analyses (Dolédec
et al., 1996).
We also performed a fourth-corner analysis that, similarly to RLQ,
computes a new matrix relating the environmental variables to biolog-
ical traits (Legendre et al., 1997). However, the fourth-corner method
provides an additional signiﬁcance test of all possible bivariate associa-
tions between single traits and environmental variables, which allows
for a more detailed and speciﬁc interpretation of trait-environment as-
sociations. We ﬁrst performed an analysis of variance statistic (i.e. the
global F-statistic) for the categorical environmental variables (cross sec-
tion, river planform and weed cutting) to test whether an overall trait-
environment association existed. In case of a signiﬁcant F, at least one
environmental category differed from the others in terms of species
traits. We then explored the bivariate relationships further. The statis-
tics of the fourth-corner method depend on the type of variables. In
case of two quantitative variables (i.e.when both the trait and the envi-
ronmental variable are quantitative), the Pearson product–moment
correlation (r) coefﬁcient is used. However, in case of quantitative traits
and qualitative environmental variables, Legendre et al. (1997) sug-
gested the use of either homogeneity statistic (d) or Pearson product–
moment correlation coefﬁcient (r). We used the latter option (r) to ob-
tain both the strengthand direction (positive or negative) of association
between the environmental variables and traits. The signiﬁcance of r
and F was obtained by permuting simultaneously the rows of Tables R
and Q (999 runs) following the model proposed by Dolédec et al.
(1996). RLQ and fourth-corner analyses were all performed in R pack-
age ade4 (Dray and Dufour, 2007).
The stream reaches were highly variable regarding total plantcover-
age, present agriculture close to the stream (50 m wide zone) and in the
catchment, and also in the concentration of major nutrients in the
stream water (Table 3). The distribution of plants in the streams was ex-
plained by linking the trait characteristics of the species to the environ-
mental conditions (Monte-Carlo test; p b0.001). The ﬁrst two axes of
the RLQ explained 52% and 22% of the total variance that links the envi-
ronmental characteristicsin Table R with species traits in Table Q (Fig.1;
Table 4). The ﬁrst RLQ axis differentiated channelised reaches from
more meandering reaches (straight natural; meandering) without cut-
tings (Fig. 1a), whereas the second axis differentiated reaches with a
channelised cross section and lower cutting intensity from reaches
with a natural cross section but more intensive cutting. The amounts
of nutrients in the stream water were related to both axis one and
two. Interestingly though, reaches with high levels of PO
P were differ-
entiated from those with high levels of NO
N(Fig. 1a). For the traits, the
ﬁrst RLQ axis differentiated survival (overwintering), productivity (EN)
and dispersal by rhizome growth from some of the speciﬁc life form
characteristics (heterophylly; anchored ﬂoating leaved) and dispersal
by fragmentation, whereas the second axis differentiated meristem
characteristics, i.e. apical and multi-apical meristem growth from basal
meristem growth (Fig. 1b). The differentiation of species according to
We found several signiﬁcant associations between environmental
variables and trait characteristics (Table 5). Generally, we found that
traits within each of the ﬁve categories (ecological preference; life
form; morphology; dispersal; survival) were signiﬁcantly related to
one or more hydromorphological variables (planform, cross section
and weed cutting; Table 5). One of the life form traits (heterophylly)
and one trait related to survival (overwintering) responded to all
types of hydromorphological degradation (i.e. planform, cross section
and weed cutting) (Table 5).
Investigating the speciﬁc hydromorphological stressors, we found
signiﬁcant bivariate associations with several trait characteristics
(Fig. 2). In many instances, the relationships were similar regarding
planform and cross section. The ecological preference for light (LE),
the life form anchored emergent leaves and the life form anchored het-
erophylly were all negatively associated with channelisation, whereas
overwintering organs were positively associated with channelisation
(Fig. 2). Additionally, the morphology index and reproduction by rhi-
zome growth were positively associated with a channelised planform
but not with a channelised cross section. Anchored heterophylly was
negatively associated with weed cutting together with single and
multi-apical growth meristems, whereas both heterophylly and apical
Key characteristics of the study reaches. Nutrient concentrations are based on ﬁve yearly
samples. Percentages of agriculture were derived from GIS in the whole catchment and
in a 50 m wide zone from the stream channel.
Mean SE Min Max
Aquatic plant coverage (%) 32.7 1.7 0.7 84.1
N (mg l
) 0.10 0.01 0.02 0.39
N (mg l
) 3.42 0.17 0.04 8.48
P (mg l
) 0.05 0.00 0.01 0.18
Agriculture buffer (%) 44.3 1.6 0.9 84.1
Agricultural catchment (%) 62.6 1.4 0.6 86.8
233A. Baattrup-Pedersen et al. / Science of the Total Environment 543 (2016) 230–238
growth were positively associated with no cutting. In contrast, single
basal growth meristem was positively associated with weed cutting
but negatively with no cutting. NE and overwintering were also traits
negatively associated with no cutting.
Several traits were also signiﬁcantly related to the level of eutrophi-
cation (Fig. 2), but the response varied for the different types of nutri-
ents. The ecological preference for light (LE) decreased with increasing
levels of NH
N but increased with increasing levels of NO
no signiﬁcant relationship was found between LE and PO
P(Fig. 2). In-
stead, a signiﬁcant relationship was found between the ecological pref-
erence for nutrients (NE) and PO
P(Fig. 2). Several life form
characteristics were also related to the concentration of nutrients in
the stream water. For example, anchored species with submerged
leaves increased with increasing concentrations of NO
whereas anchored species with emergent leaves increased with
Fig. 1. Results of the ﬁrst two axes of RLQ analysis: (a)coefﬁcients for the environmental variables, (b) coefﬁcients forthe trait, (c) eigenvalues and scoresof species and (d) eigenvalues
with the ﬁrst two axes in black. The ‘d’values give thegrid size for scale comparison across thethree ﬁgures. Codes forvariables are given in Tables 1 and 2 and for species in Suppo rting
information Table S1.
Summaryof the RLQ analysis: eigenvalues and percentage of totalco-inertia accounted for
by ﬁrst three RLQ axes, covariance refers to the covariance between the two new sets of
factorialscores projected onto the ﬁrst three RLQ axes (square root of eigenvalue); corre-
lation refers to the correlationbetween the two new setsof factorial scoresprojected onto
the ﬁrst three RLQ axes; cumulative inertia refers to the variance of each set of factorial
scores computed in the RLQ analysis, both for the environment and for the traits.
Axis 1 Axis 2
Eigenvalues 0.39 0.17
% of total co-inertia 52% 22%
Covariance 0.62 0.41
Correlation 0.33 0.21
Cumulative inertia (environment) 1.58 2.80
Cumulative inertia (traits) 2.25 5.26
234 A. Baattrup-Pedersen et al. / Science of the Total Environment 543 (2016) 230–238
increasing concentrations of NO
P, but declined with increas-
ing concentrations of NH
N(Fig. 3). We also found that several morpho-
logical traits (i.e. position of growth meristems and canopy
characteristics) were signiﬁcantly related to the concentration of nutri-
ents (Fig. 2). In particular, apical growth meristems were negatively as-
sociated with increasing levels of PO
P, whereas single basal growth
meristems were positively associated with increasing levels of PO
(Fig. 2). The overwintering capacity increased signiﬁcantlywith increas-
ing concentrations of PO
N, but declined with increasing con-
centrations of NO
We found a highly signiﬁcant relationship between aquatic plant
trait composition and important environmental stressors affecting low-
land stream habitats. This ﬁnding strongly indicates that habitat degra-
dation (hydromorphological alterations and eutrophication) mediates
selective changes in the mean functional trait composition of the
community. In accordance with our ﬁrst hypothesis, we found that
several life forms and growth characteristics were affected by
hydromorphological degradation. In particular, the trait heterophylly
responded in a consistent manner to our measures of hydromor-
phological degradation as the abundance of heterophyllous species
was negatively associated with a channelised planform, a channelised
cross section and intensive weed cutting. The consistency in the re-
sponse of this trait was also reﬂected in positive associations between
heterophylly and a natural planform (meanderingor naturally straight),
a natural cross section and absence of weed cutting.
We suggest that the lower fraction of heterophyllous species in
channelised and weed cut stream reaches is related to less diverse and
less variable habitats, speciﬁcally lack of appropriate depositional habi-
tats. Thus, heterophyllous species are particularly abundant in streams
with heterogeneous environmental conditions (Levins, 1963;
Bradshaw, 1965; Cook and Johnson, 1968), and heterophylly can be
seen as a means to maximise resource uptake by producing submerged
leaves under high ﬂow velocities in winter and early spring and then
later on during summer as the sediment builds up and water depth de-
clines, by producing ﬂoating or aerial type leaves (Allsopp, 1965;
Sculthorpe, 1967). In natural sinuous and meandering streams, sedi-
ment is eroded along the outside of meander bends and deposited fur-
ther downstream on the inside of the bends where the shear stress is
lower (Pedersen et al., 2006), and depositional areas may build up dur-
ing summer as ﬂow velocities decline. In channelised reaches, on the
other hand, habitatsare quite uniform regarding depth and velocity pat-
terns. Furthermore, dredging and weed cutting add to homogenise in-
stream habitat conditions by preventing deposition zones from devel-
oping fully through ﬂuvial geomorphological processes during low
ﬂow in summer.
Other life form traits were also uniquely affected by hydromorpho-
logical degradation (free-ﬂoating surface, anchored ﬂoating leaves),
which may be due to differences in ﬂow tolerances among life forms.
Today, the dependency between plant traits (e.g. morphology, ﬂexibil-
ity and size) and ﬂow characteristics in streams (e.g. velocity, turbu-
lence and unidirectional versus oscillating ﬂow) is poorly elucidated
(Sand-Jensen, 2003; Sand-Jensen, 2008; Bal et al., 2011). However, it
has been shown that with increasing velocities, plants with evenly dis-
tributed biomass along the water column will experience lower drag
values as they have less biomass in high-velocity areas near the surface
compared to species that concentrate their biomass near the surface
(Bal et al., 2011). For example, a species like Sparganium ssp. with linear
ﬂexible leaves is less susceptible to high water velocities compared to,
for instance, Potamogeton pectinatus that grows from multi-apical mer-
istems and concentrates its biomass in the upper waters (Sand-Jensen
et al., 1989). Therefore, the ﬁnding that free-ﬂoating species were neg-
atively associated with channelisation may reﬂect the fact that this life
form is associated with low current velocity habitats, for example back-
waters that are rare in channelised streams, whereas the opposite holds
true for more natural reaches, especially so if no weed cutting takes
The overwintering capacity (i.e. species with extensive formation of
vegetative propagules such as turbers, turions and rhizomes) also
responded to all types of hydromorphological degradation (channelised
planform, channelised cross section, high weed cutting intensity). A
common characteristic for vegetative propagules is that they remain
dormant during the coldest seasons (Sculthorpe, 1967). Species with
extensive formation of propagules may therefore better survive
unfavourable conditions during winter. Species with a high
Pseudo F-statistics of the fourth-corner analysis for the categorical hydromorphological
variables (cross section, river shape and weed cutting).
Stat. Planform Cross section Weed cutting
Value Prob. Value Prob. Value Prob.
LE F 2.75 0.005 ⁎⁎ 4.06 0.014 ⁎0.59 0.377
NE F 1.11 0.162 0.85 0.24 2.42 0.032 ⁎
Frﬂsr F 0.94 0.233 0.66 0.291 0.52 0.438
Frﬂsb F 0.45 0.54 1.29 0.158 0.38 0.51
Anﬂle F 0.10 0.92 2.55 0.052 1.52 0.097
Ansule F 1.08 0.182 0.54 0.375 0.62 0.359
Anemle F 1.96 0.027 ⁎⁎ 9.10 0.002 ⁎⁎ 0.89 0.252
Anhete F 4.51 0.001 ⁎⁎⁎ 3.25 0.027 ⁎⁎ 7.14 0.002 ⁎⁎
Meris.ma F 1.34 0.091 3.30 0.024 ⁎⁎ 1.42 0.108
Meris.sb F 1.42 0.081 0.52 0.371 3.47 0.007 ⁎⁎
Meris.sa F 1.17 0.126 3.21 0.028 ⁎⁎ 3.32 0.005 ⁎⁎
Morph.ind F 1.11 0.158 0.41 0.424 1.45 0.092
Leaf.area F 0.17 0.851 1.27 0.169 0.13 0.81
Seeds F 0.53 0.45 1.26 0.172 1.01 0.196
Rhizome F 2.21 0.011 ⁎0.29 0.494 1.17 0.162
Frag F 2.91 0.006 ⁎⁎ 0.13 0.655 1.15 0.156
N.rep.org F 1.16 0.134 2.14 0.073 1.32 0.147
F 6.12 0.002 ⁎⁎ 4.80 0.013 ⁎⁎ 9.64 0.001 ⁎⁎⁎
Fig. 2. Results of the fourth-corner analysis. The table shows all possible bivariate associa-
tions between the environmental variables (hydromorphological categories and nutrient
variables) and aquatic plant traits. Signiﬁcant (p b0.05) positive associations are repre-
sented by blackcells and signiﬁcantnegative associations by greycells. Non-signiﬁcant as-
sociationsare white. Codesfor traits and environmental variables are explained in Tables1
and 2. Note that NO
N were all log transformed prior to analyses.
235A. Baattrup-Pedersen et al. / Science of the Total Environment 543 (2016) 230–238
overwintering capacity in the present study were a highly diverse
group comprising several submerged species (e.g. Potamogeton spp.,
Myriophyllum spp., Ceratophyllum spp., Elodea canadensis), free-
ﬂoating species (e.g. Lemna spp., Spirodela polyrhiza,Utricularia spp.)
and species producing both submerged and emergent leaves (e.g.
Sparganium spp., Sagittaria sagittifolia,Myositis palustris). A majority of
the species were, however, submerged, reﬂecting that propagules may
also be used for propagation and dispersal in streams (Sculthorpe,
The coupling between hydromorphological degradation and the
abundance of species witha high overwintering capacity can be associ-
ated with a generally higher resilience of the plant community in an-
thropogenically disturbed habitats. Additionally, weed cutting and
dredging may provide a competitive advantage for species with a high
overwintering capacity. Weed cutting may extend the growing season
by improving the amount of light that reaches the stream bottom
(Dawson,1976; Ham et al., 1981). Dawson (1976) observed that cutting
during summer increased the summer biomass the following year by
giving rise to a higher overwintering biomass from which re-growth
could take place. Propagule-forming species may have a particular ad-
vantage in these streams, especially in regions like Denmark with die-
back during winter. Furthermore, species producing overwintering or-
gans may also survive weed cuttings performed in late autumn since
overwintering propagules are likely to be left intact near the stream
As expected, we found that traits associated with growth meristem
characteristics were inﬂuenced by weed cutting (basal, single and
multi-apical growth meristems). A clear and consistent pattern was
found with higher abundance of species growing from basal meristems
in stream reaches with high weed cutting intensity and lower abun-
dance of these species in stream reaches without weed cutting. At the
same time, we found the opposite pattern for species growing from api-
cal meristems. Previous studies have shown that weed cutting canhave
a severe inﬂuence on community structure (Baattrup-Pedersenand Riis,
1999; Baattrup-Pedersen et al., 2003, 2004; Pedersen et al., 2006), and
here we provide evidence that tolerance towards cutting can be con-
trolled by growth meristem characteristics. This ﬁnding seems intui-
tively logical since the position of the growth meristem determines
the potential for re-growth following cutting. That is, species with
basal meristems have an intact growth point after cutting and may
therefore start re-growth immediately after the intervention, whereas
species with apical growth meristems likely exhibit delayed re-growth.
Sparganium spp. has previously been identiﬁed as a species that is
highly tolerant to intensive weed cutting (Baattrup-Pedersen et al.,
2003). A likely reason is that the leaf-producing meristems are located
just above thestream bottom and therefore remain intact following cut-
ting. Additionally, this species also has extensive rhizomes that may
provide a competitive advantage in disturbed environments
(Baattrup-Pedersen et al., 2003; Wiegleb et al., 2014). Using continuous
multi-year data, Wiegleb et al. (2014) noticed an increase in Sparganium
emersum along with other rhizomatic species in response to increasing
anthropogenic disturbance in German rivers when combining both
hydromorphological stressors (e.g. cutting, dredging and construction
work) and stressors associated with water quality such as malfunction
of sewage plants and intensiﬁcation of agricultural land use. Interest-
ingly, S. emersum is also widely distributed and abundant in the least-
disturbed lowland streams in Europe (Baattrup-Pedersen et al., 2008)
but still less abundant than in anthropogenically disturbed streams
(Riis et al., 2000; Baattrup-Pedersen et al., 2002; Pedersen et al., 2006;
Birk and Wilby, 2010; Steffen et al., 2013). The success of this species
in non-impacted streams may be linked to its low light compensation
point and low photosaturation level (Sand-Jensen et al., 1989), which
explains why it is often observed in the sub-layer in multi-layer com-
munities (Baattrup-Pedersen et al., 2002).
In accordance with our second hypothesis, we also found that some
traits could distinguish hydromorphological degradation (free-ﬂoating,
Fig. 3. A conceptual model integrating ﬁndings obtained in this study on how hydromorphological degradation affects trait-abundance patterns in aquatic plant communities in lowland
streams with a recently published study on how eutrophication affects trait-abundance patterns in European lowland streams (Baattrup-Pedersenetal.,2015). Dotted arrow indicates
how weed cutting can interact with eutrophication in affecting trait-abundance patterns. Weed cutting improves light availability, implying that efﬁcient light capture or utilisation are
set aside as traits of importance in eutrophic streams. Instead traits associated with high levels of productivity become important, such as preference for high nutrient levels (Ellenberg
N) and the ability to form dense standing crops with extensive lateral spread (Morphology index). Note that the overwintering capacity of the plant community was positively affected
N but negatively affected by NO
N(notspeciﬁed in the ﬁgure).
236 A. Baattrup-Pedersen et al. / Science of the Total Environment 543 (2016) 230–238
surface; anchored ﬂoating leaves; anchored heterophylly) from eutro-
phication (free-ﬂoating submerged; leaf area). We have summarised
our main ﬁndings in Fig. 3 together with those found in a recently pub-
lished study on how eutrophication affects trait-abundance patterns of
aquatic plants in European lowland streams (Baattrup-Pedersen et al.,
in press). Together these results provide a conceptual framework for in-
terpretation of how multiple stressors interact and affect trait-
abundance patterns in lowland streams. Lowland streams that are
mainly affected by eutrophication have a high abundance of species
growing from apical meristems and species that can efﬁciently utilise
light (Baattrup-Pedersen et al., in press), because species possessing
these traits have a competitive advantage due to light limitation under
eutrophic conditions. The apparently opposite response found here in
streams with high phosphate levels likely indicates that weed cutting
interacts with nutrient availability in affecting trait-abundancepatterns
as depictedin ﬁg. 3. Thus, as opposedto streams with high nitrate levels,
we found that high phosphate levels were positively associated with
weed cutting intensity and, consequently, streams with high phosphate
levels also experienced regular cuttings. Therefore, weed cutting proba-
bly sets aside light as a factor controlling species composition under
phosphate-rich conditions likely because plenty of light reach the
stream bottom following biomass removal, and at the same time shad-
ing from epiphytic algae becomes less important since it takes time be-
fore mats develop on the new leaves that develop after cutting. This
interpretation of our results also implies that the positive response ob-
served between phosphate levels in the stream water and the abun-
dance of species with basal growth meristems is without causality, but
merely reﬂects that species growing from basal meristems dominate
in regularly cut reaches.
Interestingly, we also observed a direct and positive response of
traits associated with species productivity (NE and a high morphology
index; Birk et al., 2006; Dudley et al., 2013) and phosphate levels in
the stream water, indicating that phosphate played a direct role in
trait-abundance patterns as depicted in Fig. 3. Again, this ﬁnding may
highlight that constraints associated with low light availability in eutro-
phic streams (Hilton et al., 2006; Baattrup-Pedersen et al., in press)are
relieved if these streams are regularly cut. Biomass removal through
cutting improves the light climate and may enable productive species
that are inefﬁcient in light capture or utilisation to compete successfully
provided that they are resilient towards weed cutting. According to
Grime (1988), this should be reﬂected in an increase in the proportion
of species with an ecological preference for high nutrient levels (high
Ellenberg N) as well as an increase in tall species forming dense stand-
ing crops with extensive lateral spread as depicted in Fig. 3.
We found clear evidence that habitat degradation in the studied
lowland streams (hydromorphological alterations and eutrophication)
mediated selective changes in the functional trait composition of the
aquatic plant community. The abundance of heterophyllous species
showed a unique and negative response to hydromorphological degra-
dation (channelised planform, a channelised cross section and higher
weed cutting intensity), probably reﬂecting that channelisation
and weed cutting both contribute to homogenise in-stream habitats,
leaving restricted space for deposition zones suitable forheterophyllous
species. We also found indications that eutrophication and
hydromorphological degradation interacted in their effects on the trait
composition of the community. We provide a conceptual framework
for interpretation of how eutrophication and hydromorphological deg-
radation interact on trait-abundance patterns (Fig. 3) that might be ap-
plicable for plant-dominated lowland streams in other regions as well.
We propose that weed cutting can set aside light as a factor controlling
species composition under nutrient-rich conditions in lowland streams
and that high productive species will increase in abundance. Our ﬁnd-
ings support the merit of trait-based approaches in biomonitoring as
these can throw light on mechanisms controlling structural changes
under environmental stress (Fig. 3;Baattrup-Pedersen et al., in press).
The ability to disentangle several stressors is particularly important in
lowland stream environments where several stressors are acting in con-
cert since it can enable managers to target the impact of the most im-
portant stressor ﬁrst, this being essential to improve the ecological
Supplementary data to this article can be found online at http://dx.
The study was supported by the European Union 7th Framework
Project REFORM under contract no. 282656 and MARS under contract
no. 603378. We thank Anne Mette Poulsen for manuscript editing and
Tinna Christensen for ﬁgure layout.
Allsopp, A., 1965. Land and water forms: physiological aspects. Handbuch der
Pﬂanzenphysiologie 15, 1236–1255.
Baattrup-Pedersen, A., Riis, T., 1999. Plant diversity and composition in relation to sub-
stratum characteristics in regulated and unregulated Danish streams. Freshw. Biol.
Baattrup-Pedersen, A., Larsen, S.E., Riis, T., 2002. Long-term effects of stream management
on plant communities in two Danish lowland streams. Hydrobiologia 48, 33–45.
Baattrup-Pedersen, A., Larsen, S.E., Riis, T., 2003. Composition and richness of plant com-
munitiesin small Danish streams—inﬂuence of environmental factorsand weed cut-
ting. Hydrobiologia 495, 171–179.
Baattrup-Pedersen, A., Riis, T., 2004. Impacts of different weed cutting practices on plant
species diversity and composition in a Danish stream. River Res. Appl. 20, 103–114.
Baattrup-Pedersen, A., Friberg, N., Pedersen, M.L., Skriver, J., Kronvang, B., Larsen, S.E.,
2004. Anvendelse af Vandrammedirektivet i danske vandløb. Danmarks
Miljøundersøgelser, Aarhus Universitet. Faglig rapport fra DMU. vol. 499 (In Danish).
Baattrup-Pedersen, A., Springe, G., Riis, T., Larsen, S.E., Sand-Jensen, K., Kjellerup Larsen,
L.M., 2008. The search for reference conditions for stream vegetation in northern
Europe. Freshw. Biol. 53, 1890–1901.
Baattrup-Pedersen, A., Kristensen, E.A., Jorgensen, J., Skriver, J., Kronvang, B., Andersen,
H.E., Hoffman, C.C., Kjellerup Larsen,L.M., 2009. Can a priori deﬁned reference criteria
be used to select reference sites in Danish streams? Implications for implementing
the Water Framework Directive. J. Environ. Monit. 11, 344–352.
Baattrup-Pedersen, A., Göthe, E., Larsen, S.E., O'Hare, M., Birk, S., Riis, T., Friberg, N., 2015.
Plant trait characteristics vary with size and eutrophication in European lowland
streams. J. Appl. Ecol. http://dx.doi.org/10.1111/1365-2664.12509l (in press).
Bal, K.D., Bouma, T.J., Buis, K., Struyf, E., Jonas, S., Backz, H., Meire, P., 2011. Trade-off be-
tween drag reduction and light interception ofaquatic plants: comparing ﬁve aquatic
plants with contrasting morphology. Funct. Ecol. 25, 1197–1205.
Birk,S.,Korte,T.,Hering,D.,2006.2006. Intercalibration of assessment methods for
aquatic plants in lowland streams: direct comparison and analysis of common met-
rics. Hydrobiologia 566, 417–430.
Birk, S., Wilby, N., 2010. Towards harmonization of ecological quality classiﬁcation: estab-
lishing common groundsin European plant assessment for rivers. Hydrobiologia 652,
Bradshaw, A.D., 1965. Evolutionary signiﬁcance of phenotypic plasticity in plants. Adv.
Genet. 13, 115–155.
Brewer, C.A., Parker, M., 1990. Adaptations of macrophytes to life in moving water: up-
slope limits and mechanical properties of stems. Hydrobiologia 194, 133–142.
Brookes, A., 1987. The distribution and management of channelized streams in Denmark.
Regul. Riv. 1, 3–16.
Brookes,A., Gregory, J., 1983.An assessment of river channelisation inEngland and Wales.
Sci. Total Environ. 27, 97–111.
Cook, S.A., Johnson, M.P., 1968. Adaptation to heterogeneous environments I. Variation in
heterophylly in Ranunculus ﬂammula L. Evolution 22, 496–516.
Dahm, V., Hering, D., Nemitz, D.,Graf, W., Schmidt-Kloiber, A.,Leitner, P., Melcher, A., Feld,
K., 2013. Effects of physico-chemistry, land use and hydromorphologyon three river-
ine organism groups: a comparative analysis with monitoring data from Germany
and Austria. Hydrobiologia 704, 389–415.
Dawson, F.H., 1976. The annual production of the aquatic aquatic plant, Ranunculus
penicillatus var. calcareus (R.W. Butcher) C.D.K. Cook. Aquat. Bot. 2, 51–73.
Dawson, F.H., 1989. Ecology and management of water plants in lowland streams. Rep.
Freshwat. Biol. Ass. 57, 43–60.
Dawson, F.H., Raven, P.J., Gravelle, M.J., 1999. Distribution of the morphological groups of
aquatic plants for rivers in the UK. Hydrobiologia 415, 123–130.
DFFE. Enkeltbetaling; 2008. http://ferv.fvm.dk/Enkeltbetaling.aspx?ID=20385.
Diaz, S., Lavorel, S., deBello, F., Quétier, F., Grigulisand, K., Robson, T.M., 2007. Incorporat-
ing plant functional diversity effects in ecosystem service assessments. PNAS 104,
Dolédec, S., Chessel, D., ter Braak, C.J.F., Champely, S., 1996. Matching species traits to en-
vironmental variables: a new three-table ordination method. Environ. Ecol. Stat. 3,
237A. Baattrup-Pedersen et al. / Science of the Total Environment 543 (2016) 230–238
Dolédec, S., Phillips, N., Scarsbrook, M., Riley, R.H., Townsend, C.R., 2006. Comparison of
structural and functional approaches to determining landuse effects on grassland
stream invertebrate communities. J. North Am. Benthol. Soc. 25, 44–60.
Dray, S., Dufour, A.B.,2007. The ade4 package:implementing the duality diagramfor ecol-
ogists. J. Stat. Softw. 22, 1–20.
Dudley, B., Dunbar, M., Penning, E., Kolada, A., Hellsten, S., Oggioni, A., Bertrin, V., Ecke, F.,
Søndergaard, M., 2013. Measurements of uncertainty in aquatic plant metrics usedto
assess European lake water quality. Hydrobiologia 704, 179–191.
EEA, 2012. Climate change impacts and vulnerability in Europe 2012. EEA Report no 12.
European Environment Agency.
Ellenberg, H., Weber, H.E., Düll, R., Wirth, V., Werner, W., Paulissen, D., 1991. Zeigerwerte
von Pﬂanzen in Mitteleuropa. Scr. Geol. 18, 1–248.
Elosegi, A., Sabater, S., 2013. Effects of hydromorphological impacts on river ecosystem
functioning: a review and suggestions for assessing ecological impacts. Hydrobiologia
Feld, C.K., deBello, F., Dolédec, S., 2014. Biodiversity of traits and species both show weak
responses to hydromorphological alteration in lowland river macroinvertebrates.
Freshw. Biol. 59, 233–248.
Fox, A.M., Murphy, K.J., 1990. The efﬁcacy and ecological impacts of herbicide and cutting
regime on the submerged plant communities of four British rivers. I. Comparison of
management efﬁcacies. J. Appl. Ecol. 27, 520–540.
Friberg,N., Baattrup-Pedersen, A., Pedersen, M.L., Skriver,J., 2005. The new Danish Stream
Monitoring Programme (NOVANA) - preparing monitoring activities for the Water
Framework Directive era. Environ. Monit. Assess. 111, 27–42.
Friberg, N., Bonada, N., Bradley, D.C., Dunbar, M.J., Edwards, F.K., Grey, J., et al., 2011. Bio-
monitoring of human impacts in freshwater ecosystems: the good, the bad and the
ugly. Adv. Ecol. Res. 44, 1–68.
Grime, J.P.,1988. The C–S–R model of primary plant strategies —origins, implications and
tests. In: Gottlieb, J.D., Jain, K.S. (Eds.), Plant Evolutionary Biology. Chapman and Hall,
Gurnell, A.M., O'Hare, J.M., O'Hare, M.T., Dunbar, M.J., Scarlett, P.M., 2010. An exploration
of associations betweenassemblages of aquatic plant morphotypes and channel geo-
morphological properties within British rivers. Geomorphology 116, 135–144.
Ham, S.F., Wright, J.F., Berrie, A.D., 1981. Growth and recession of aquatic aquatic plants
on an unshaded section of the River Lambourn, England, from 1971 to 1976. Freshw.
Biol. 11, 381–390.
Harding, J.S., Benﬁeld, E.F., Bolstad, P.V., Helfman, G.S., Jones, E.B.D., 1998. Stream biodi-
versity: the ghost of land use past. Proc. Nat. Acad. Sci. USA 95, 14843–14847.
Hering, D., Johnson, R.K., Kramm, S., Schmutz, S., Szoszkiewicz, K., Verdonschot, P.F.M.,
2006. Assessment of European streams with diatoms, aquatic plants, macroinverte-
brates and ﬁsh: a comparative metric-based analysis of organism response to stress.
Freshw. Biol. 51, 1757–1785.
Hill, M.O., Smith, A.J.E., 1976. Principal component analysis of taxonomic data with multi-
state discrete characters. Taxon 25, 249–255.
Hilton, J., O'Hare, M., Bowes, M.J., Jones, I., 2006. How green is my river? A new paradigm
of eutrophication in rivers. Sci. Total Environ. 365, 66–83.
Kaenel, B., Uehlinger, R., 1999. Aquatic plant manageme nt: ecological effects in two
streams of the Swiss Plateau. Hydrobiologia 145, 257–263.
Kail, J., Wolter, C., 2013. Pressures at larger spatial scales strongly inﬂuence theecological
status of heavily modiﬁed river water bodies in Germany. Sci. Total Environ. 454,
Kern Hansen, U., Dawson, F.H., 1978. The standing crop of aquatic plants of lowland
streams in Denmark and their inter-relationships of water nutrients in plant, sedi-
ment and water. EWRS 5th Symposium on Aquatic Weeds, pp. 143–150.
Landwehr, K., Rhoads, B.L., 2003. Depositional response of a headwater stream to chan-
nelization East Central Illinois USA. River Res. Appl. 19, 77–100.
Legendre, P., Galzin, R., Harmelin-Vivien, M.L., 1997. Relating behaviour to habitat: solu-
tions to the fourth-corner problem. Ecology 78, 547–562.
Levins, R., 1963. Theory of ﬁtness in a heterogeneous environment. II. Developmental
ﬂexibility and niche selection. Am. Nat. 47, 75–90.
Mattingly, R.L., Herricks, E.E.,Johnston, D.M., 1993. Channelization and levee construction
in Illinois: review and implications for management. Environ. Manag. 17, 781–795.
Moretti, M., Legg, C., 2009. Combining plant and animal traits to assess community func-
tional responses to disturbance. Ecography 32, 299–309.
Nielsen,K., Stjernholm, M., Olsen, B.Ø, Müller-Wohlfeil, D.I., Madsen,I.L., et al., 2000. Areal
Informations Systemet - AIS. Ministry of Environment, National Environmental Re-
search Institute, Copenhagen (in Danish).
Pedersen, T.C.M., Baattrup-Pedersen, A., Madsen, T.V., 2006. Effects of stream restoration
and management on plant communities in lowland strea ms. Freshw. Biol. 51,
Pedersen, M.L., Baattrup-Pedersen, A., Wiberg-Larsen, P., 2007. Økologisk overvågning i
vandløb og på vandløbsnære arealer under NOVANA 2004-2009 (Teknisk anvisning
fra DMU; 21). Danmarks Miljøundersøgelser, Aarhus Universitet In Danish.
Pitlo, R.H., Dawson, F.H., 1990. Flow-resistence of aquatic weeds. In: Pieterse, A.H. ,
Murphy, K.J. (Eds.), Aquatic Weeds. The Ecology and Management of Nuisance
Aquatic Vegetation. Oxford University Press, Oxford.
Poff, N.L., 1997. Landscape ﬁlters and species traits: towards mechanistic understanding
and prediction in stream ecology. J. North Am. Benthol. Soc. 16, 391–409.
Puijalon, S., Bouma, T.J., Douady, C.J., Van Groenenda el, J., Anten, N.P.R., Martel, E.,
Bornette, G., 2011. Plant resistance to mechanical stress: evidence of an avoidance–
2tolerance trade-off. New Phytol. 191, 1141–1149.
Rambaud,M., Combroux, I., Haury, J., Moret, J., Machon, N., Zavodna, et al.,2009. Relation-
ships between channelization structures, environmental characteristics, and plant
communities in four French streams in the Seine–Normandy catch ment. J. North
Am. Benthol. Soc. 28, 596–610.
Riis, T., Sand-Jensen, K., Vestergaard, O., 2000. Plant communities in lowland streams:
species composition and environmental factors. Aquat. Bot. 66, 255–272.
Sand-Jensen, K., Jeppesen, E., Nielsen, K., van der Bijl, L., Hjermind, L., Nielsen, L.W., et al.,
1989. Growth of aquatic plants and ecosystem consequences in a lowland Danish
stream. Freshw. Biol. 22, 15–32.
Sand-Jensen, K., 2003.Drag an d reconﬁguration of freshwater aquaticplants. Freshw. Biol.
Sand-Jensen, K., 2008. Drag forces on common plant species in temperate streams: con-
sequences of morphology, velocity and biomass. Hydrobiologia 610, 307–319.
Schutten, J., Davy, A.J., 2000. Predicting the hydraulic forces on submerged macrophytes
from current velocity, biomass and morphology. Oecologia 123, 445–452.
Sculthorpe, C.D., 1967. The Biology of Aquati c Vascular Plants. Edward Arnold Ltd.,
Southwood, T.R.E., 1988. Tactics, strategies and templates. Oikos 52, 3–18.
Steffen, K., Becker, T., Herr, W., Leuschner, C., 2013. Diversity loss in the plant vegetation
of northwest German streams and rivers between the 1950s and 2010. Hydrobiologia
Vaughan,I.P., Diamond, M., Gurnell, A.M., Hall, K.A., Jenkins,A., Milner, N.J., et al., 2009. In-
tegrating ecology with hydromorphology: a priority for river science and manage-
ment. Aquat. Conserv. Mar. Freshwat. Ecosyst. 19, 113–125.
Verdonschot, P.F.M., Nijboer, R.C., 2002. Towards a decision support system for stream
restoration in the Net herlands: an overview of restoration projects and future
needs. Hydrobiologia 478, 131–148.
Vereecken, H., Baetens, J., Viaene, P., Mostaert, F., Meire, P., 2006. Ecological management
of aquatic plants: effects in lowland streams. Hydrobiologia 570, 205–210.
Vörösmarty, C.J., McIntyre, P.B., Gessner, M.O., Dudgeon, D., Prusevich, A., Greeen, P.,et al.,
2010. Global threats to human water security and river biodiversity. Nature 467,
Wade, P.M., 1993. The inﬂuence of vegetation pre-dredging on the post-dredging com-
munity. J. Aquat. Plant Manag. 31, 141–144.
Wade, P.M., Edwards, R.W., 1980. The effect of channel maintenance on the aquatic
aquatic plants of the drainage channels of the Monmouthshire Levels, South Wales,
Wiberg-Larsen P, Windolf J, Bøgestrand J, Larsen SE, Thodsen H, OvesenNB et al. Vandløb
2013. NOVANA. Aar hus Universite t, DCE —Nationalt Center for Miljø og Energi.
Videnskabelig rapport fra DCE —Nationalt Center for Miljø og Energi nr. 121; 2015.
http://dce2.au.dk/pub/SR121.pdf (in Danish).
Wiegleb,G., Bröring, U., Filetti, M., Brux, H., Herr,W., 2014. Long-termdynamics of aquatic
plant dominance and growth-for m types in two north-west German lowland
streams. Freshw. Biol. 59, 1012–1025.
Willby, N.J., Abernethy, V.J., D emars, B.O.L., 2000. Attribute-based clas siﬁcation of
European hydrophytes and its relationship to habitat utilization. Freshw. Biol. 43,
Wood, K.A., Stillman, R.A., Clarke, R.T., Daunt, F., O'Hare, M.T., 2012. The impact of water-
fowl herbivory on plant standing crop: a meta-analysis. Hydrobiologia 686, 157–167.
238 A. Baattrup-Pedersen et al. / Science of the Total Environment 543 (2016) 230–238