Guse Et Al 2015 Eco-Hydrologic Model Cascades

Science of the Total Environment 533 (2015) 542556
Contents lists available at ScienceDirect
Science of the Total Environment

journal homepage: www.elsevier.com/locate/scitotenv
Eco-hydrologic model cascades: Simulating land use and climate change

impacts on hydrology, hydraulics and habitats for sh
and macroinvertebrates
Bjrn Guse a,, Jochem Kail b,c, Johannes Radinger b, Maria Schrder c, Jens Kiesel a, Daniel Hering c,
Christian Wolter b, Nicola Fohrer a
a
b
c
Christian-Albrechts-University of Kiel, Institute of Natural Resource Conservation, Department of Hydrology and Water Resources Management, Kiel, Germany
Leibniz-Institute of Freshwater Ecology and Inland Fisheries, Berlin, Germany
University of Duisburg-Essen, Faculty of Biology, Department of Aquatic Ecology, Essen, Germany
H I G H L I G H T S
A recently developed model cascade is applied in different scenario simulations.

Impact of climate and land use change on hydrology and biota is assessed.
Climate change affects hydrology, hydraulics and shes.
Abundance of macroinvertebrates is inuenced by land use and climate change.
Direction and intensity of impact on habitat suitability is highly species-dependent.
a r t i c l e
i n f o
Article history:
Received 24 November 2014
Received in revised form 18 May 2015
Accepted 18 May 2015
Available online xxxx
Editor: D. Barcelo
a b s t r a c t
Climate and land use changes affect the hydro- and biosphere at different spatial scales. These changes alter
hydrological processes at the catchment scale, which impact hydrodynamics and habitat conditions for biota at
the river reach scale. In order to investigate the impact of large-scale changes on biota, a cascade of models at
different scales is required. Using scenario simulations, the impact of climate and land use change can be
compared along the model cascade.
Such a cascade of consecutively coupled models was applied in this study. Discharge and water quality are
predicted with a hydrological model at the catchment scale. The hydraulic ow conditions are predicted by
hydrodynamic models. The habitat suitability under these hydraulic and water quality conditions is assessed
based on habitat models for sh and macroinvertebrates. This modelling cascade was applied to predict and
compare the impacts of climate- and land use changes at different scales to nally assess their effects on sh
and macroinvertebrates.
Model simulations revealed that magnitude and direction of change differed along the modelling cascade. Whilst
the hydrological model predicted a relevant decrease of discharge due to climate change, the hydraulic conditions changed less.
Generally, the habitat suitability for sh decreased but this was strongly species-specic and suitability even
increased for some species. In contrast to climate change, the effect of land use change on discharge was negligible. However, land use change had a stronger impact on the modelled nitrate concentrations affecting the
abundances of macroinvertebrates.
The scenario simulations for the two organism groups illustrated that direction and intensity of changes in
habitat suitability are highly species-dependent. Thus, a joined model analysis of different organism groups
combined with the results of hydrological and hydrodynamic models is recommended to assess the impact of
climate and land use changes on river ecosystems.
2015 Elsevier B.V. All rights reserved.
1. Introduction
Corresponding author.
E-mail address: bguse@hydrology.uni-kiel.de (B. Guse).
http://dx.doi.org/10.1016/j.scitotenv.2015.05.078
0048-9697/ 2015 Elsevier B.V. All rights reserved.
Climate and land use changes are intrinsic drivers of hydrological processes (Blschl et al., 2007; Juckem et al., 2008; Li et al., 2009). These
processes are summarized in the river discharge which reects various
kinds of changes in the catchment (Ceola et al., 2014). Several model
B. Guse et al. / Science of the Total Environment 533 (2015) 542556
studies analysed the individual or combined effects of climate or land

use change on hydrology (e.g., Niehoff et al., 2002; Hesse et al., 2008;
Juckem et al., 2008; Tu, 2009; Li et al., 2009; Setegn et al., 2011;
Santos et al., 2014), sediment or nutrients (e.g., Huang et al., 2009; Tu,
2009; Li et al., 2011; Martinkova et al., 2011; Bieger et al., 2013) at the
catchment scale.
Catchment-wide hydrological changes also impact conditions at
smaller scales ranging from in-stream hydraulics in river segments
to habitat conditions of biota at the reach scale (Kiesel et al., 2009,
2013). Aquatic biota is strongly impacted by large-scale environmental pressures as revealed recently by several empirical studies
(Dahm et al., 2013; Marzin et al., 2013). It was emphasised that
catchment-scale pressures such as land-use or climate are more relevant in controlling occurrences and abundances of macroinvertebrate
and sh communities in comparison to pressures at smaller spatial
scales (Roth et al., 1996; Allan and Johnson, 1997; Hughes et al., 2008;
Stephenson and Morin, 2009; Sundermann et al., 2013).
The importance of pressures acting at different spatial scales and
the vulnerability to changes of the abiotic conditions differ between
organism groups such as sh and macroinvertebrates (Hering et al.,
2006; Pander and Geist, 2013; Mller et al., 2014). Abundance and
diversity of sh species are highly affected by habitat degradation and
habitat loss at local scale (Hering et al., 2006; Kail and Wolter, 2013),
species-specic dispersal and fragmentation (Radinger and Wolter,
2014, 2015), and by alterations of natural ow velocity patterns and
natural ow regimes (Logez et al., 2013; McManamay and Frimpong,
2014).
Distribution and composition of macroinvertebrates on the catchment scale are primarly affected by water quality and land use change
at the catchment scale (e.g., Feld and Hering, 2007; Sundermann et al.,
2013; Valle Junior et al., 2015). Increased nitrogen loads lead to
an increase in biomass and productivity of primary producers. The
resulting accumulation of organic material degrades water quality and
leads to a loss of habitats resulting in shifts in community compositions
(Camargo and Alonso, 2006).
Empirical linkages, however, do not explain the underlying processes ranging from hydrological and hydraulic processes to the occurrence
and dispersal of biota (Beechie et al., 2011). Thus, process-based models
are required for reliable predictions under changing conditions. These
process-based models are commonly applied to hydrology or hydraulics
but rarely to biota.
To obtain a fundamental and holistic understanding of the functioning of habitat-species interactions, all dominant processes from hydrology to biota have to be represented at their relevant scales and analysed
together in an integrated abioticbiotic model cascade. A model cascade
is a sequence of consecutively combined models where one model
provides the input for another model as presented by Kiesel et al.
(2009). Whilst process-based model cascades spanning from global
change to biological processes are commonly applied in lakes (Mooij
et al., 2010), the use of such models in rivers is still underdeveloped
(Kiesel et al., 2009).
Kiesel et al. (2009) and Kiesel et al. (2013) developed a rst prototype, which couples an eco-hydrological and 1D hydrodynamic model
to predict discharge at the catchment scale and the resulting water
level at the downstream end of a study reach. They predicted the
respective ow and substrate conditions using a 2D morphodynamic
model. The suitability of these habitats for macroinvertebrates was
assessed based on habitat models (Kiesel et al., 2015). This modelling
cascade was recently further developed by including morphological
and dispersal models by Kail et al. (2015). In this model framework,
sh were considered besides macroinvertebrates to test the applicability of such an integrated modelling framework for different organism
groups. This modelling framework consists of separate models for the
major processes and depicts the different habitat characteristics of sh
and macroinvertebrates. The principal applicability of their modelling
framework was demonstrated in Kail et al. (2015).
543
Applying this modelling framework in scenario simulations enables

the assessment of the effects of changing climate and land use on abiotic
conditions and related biota. Scenario simulations were performed
to estimate potential effects of changes as a basis for restoration of freshwater habitats. The effects of climate and land use change on biota act
through intertwined and complex dependencies. These connections
can be disentangled by using the modelling framework.
Therefore, this study used the modelling framework presented by
Kail et al. (2015) to investigate consecutive impacts of land use and
climate change on hydrology, hydrodynamic and biota. Two different
process chains were analysed for considering the specic habitat preferences of sh and macroinvertebrates, respectively. It is therefore one of
the rst attempts to simulate the abundance and occurrence of riverine
species, resulting from a chain of processes from the catchment hydrology down to local habitat suitability under climate and land use change.
The main objectives of this study are:
1. to simulate the impact of climate and land use changes along the
whole modelling framework from catchment scale hydrology to
reach scale hydrodynamics on habitat suitability for sh and macroinvertebrates.
2. to compare the relevance of predicted climate and land use changes
for future eco-hydrology (river discharge, nutrient concentration),
hydrodynamics (ow velocity, water depth) and selected sh and
macroinvertebrates species.
2. Material and methods

2.1. Design of the model cascade
The single parts of the applied integrated modelling framework
developed by Kail et al. (2015) are presented briey step-by-step in
the order of decreasing spatial scale (Table 1) and the model cascade
is then described in detail in Section 2.3.
Due to the different environmental requirements of the two selected
organism groups, two process chains were modelled. The rst process
chain analyses the impact of hydraulic habitat conditions on sh, whilst
the second process chain analyses the effects of water quality, i.e., nitrate concentrations, for macroinvertebrates (Fig. 1).
Both process chains start with the eco-hydrological model SWAT
(Arnold et al., 1998) (a in Fig. 1), which provides discharge and water
quality time series at the catchment scale. The modelled time series as
simulated with the SWAT model in the climate and land use scenarios
are the initial input to the scenario simulations in both model process
chains.
In the sh process chain, the 1-D hydraulic model HEC-RAS (USACE,
2002) (b in Fig. 1) was used to develop a rating curve for the downstream end of the study reach to predict the water level for given
discharges. Discharge and water level are necessary input parameters
for the 2-D hydrodynamic model iRIC-FaSTMECH (c in Fig. 1), which
predicts depth-averaged ow velocity and water depth for all raster
cells of a two-dimensional computational grid. The suitability of these
hydraulic habitat conditions in each raster cell was assessed using a
sh habitat model (d in Fig. 1). The estimated values were weighted
by the respective grid cell area and summarized as Weighted Usable
Area [WUA] (Bovee and Cochnauer, 1977; Payne, 2003), a single value
of habitat suitability for the entire study reach.
In the macroinvertebrate process chain, the SWAT model provides
nitrate time series. This nitrate concentration was included in a further
development of the recently published habitat evaluation tool (HET, e in
Fig. 1) (Kiesel et al., 2015), which models the presence and abundance
of macroinvertebrates for a dened spatial unit on the reach scale.
Thus, changes in the nitrate concentration affect the abundances of
macroinvertebrates.
544
Table 1
Used models and their scales of application. The columns Input and Output are focused on the connections between the model in the cascade and only include selected inputs and outputs
of a model.
Scale
Model type
Model name
Input
Output
Catchment
Eco-hydrological model
SWAT
Daily time series of discharge and nitrate
River segment
1-D hydraulic model
HEC-RAS
River reach
2-D hydrodynamic model
iRIC-FaSTMECH
River reach
Fish habitat model
GRASS GIS r.fuzzy.system
River reach
Macroinvertebrate habitat model
Habitat evaluation tool
Daily time series of climate variables

Spatial distribution of crop shares
Rating curve at lower end
River geometry
Discharge
Rating curve
River bathymetry
Flow velocity
Water depth
Fuzzy rules
Fuzzy membership
Nitrate concentration
Rating curve for cross-sections

Flow velocity
Water depth
Habitat suitability maps
Weighted usable area
Species abundances
2.2. Study areas and spatial scales
2.3. Model chain and model parametrisation
The Treene is a mesoscale sand-bed river in the northern German

lowlands with a size of 481 km2 at the catchment outlet Treia (Fig. 2).
The catchment is dominated by groundwater ow and characterised
by low hydraulic gradients (Kiesel et al., 2010; Pfannerstill et al.,
2014a). Elevations range from 2 to 80 m. Agriculture is the dominant
land use in the Treene catchment (70%), whilst only small parts of the
catchment are covered by forest (8%) or urban areas (8%) (Fig. 2A).
Whilst the percentages of forest and urban areas as well as the agricultural areas only changed marginally during the last decade, the cultivation of the agricultural crops was highly dynamic (Marquardt, 2008;
Geertz, 2012). The most remarkable change in the agricultural crops
was an increase of corn elds due to the rising demand for energy
crops (Geertz, 2012; Guse et al., 2014b).
In addition to the catchment scale, two other scales (river segment
and river reach) were considered as shown in Fig. 2. The 1D hydraulic
model was used at the river segment scale, and the 2D-hydrodynamic
and habitat models at the reach-scale. The near-natural study reach
260 m in length was located in the lower part of the catchment (subbasin 85 in Fig. 2B). It was considered near-natural since it is heavily
meandering, has a mean bankfull width (10.9 m) and depth (1.4 m)
typical of a natural river with highly cohesive river banks (80100%
silt/clay content) and a bankfull discharge of 6.0 m3/s (Kail et al.,
2015). It differs from fully natural reaches since bed-material is pure
sand with a D50 of 0.16 mm, whilst gravel-patches would be present
in the natural state, and the reach is bordered by pasture instead of naturally occurring oodplain forest (Pottgiesser and Sommerhuser,
2008). Mean discharge is 3.0 m3/s and average slope 0.045.
2.3.1. Eco-hydrological model SWAT at the catchment scale

The hydrological conditions in the Treene catchment were modelled
using the eco-hydrological model SWAT (Arnold et al., 1998; Arnold and
Fohrer, 2005) (a in Fig. 1). The SWAT model is a continuous conceptual
process-based model including water balance and nutrient cycles. The
smallest spatially explicit units in this semi-distributed model are
subbasins. Water balance for each subbasin is calculated at the land
phase under consideration of daily values for precipitation, evapotranspiration, runoff components and soil water storage. The subbasins are
connected in the water phase in which water is routed to the next
downstream subbasin. Spatial explicit outputs are provided for the
subbasin outlet. We refer to the theoretical documentation of the
SWAT model for further details (Neitsch et al., 2011).
For the catchment modelling, a digital elevation model with a
resolution of 25 m (LVERMA S-H, 1995), a land use map (Automatisierte
Liegendschaftskarte (digitalised property map) (LVERMA S-H, 2004))
and a soil map (BGR, 1999) were used. Measured climate data provided for four climate stations (Fig. 2A) by the German Weather Service was available from 1950 to 2012. The STAR model provided by the
Potsdam Institute for Climate Impact Research (PIK) delivered climate
change scenario data from 2013 to 2060 (Orlowsky et al., 2008) for
the same four stations. Discharge data from six spatially distributed hydrological stations including the catchment outlet Treia were used for
calibration (Fig. 2A). Daily measurements of nitrate concentrations
from the hydrological station Treia from 10/2010 to 09/2012 were
analysed in the laboratory and used to calibrate and validate the
SWAT model.
f
a
Fig. 1. Two process chains of the model cascade as used in the climate and land use change scenario simulations. Letters are assigned to the models and scenarios in the order of their use in
this study. MIV = macroinvertebrates.
545
Fig. 2. Different scales in the Treene catchment. A) Land use in the catchment including available discharge and climate stations. B) River segment with mapped cross-sections and the
study river reach (subbasin 85). C) Administrative borders and major cities in Germany with the Treene catchment. Maps A and B are linked by the hydrological station Sollerup and
the river network of HEC-RAS which are shown in both maps.
The SWAT model was set up for the Treene catchment resulting
in 108 subbasins (see also Guse et al. (2014b)). A separate subbasin
(Nr. 85 in Fig. 2B) was dened for the study river reach to provide
model outputs (discharge time series) for the hydrodynamic models.
Five typical crop rotations of the Treene catchment including the
major crops corn and wheat were implemented in the SWAT model,
which is also important for simulation of land use scenarios as presented in Guse et al. (2014a).
The model was run for the recent period from 19972012 including
a warm-up period of four years. Thus, the calibration period for
discharge was set from 20012005 and the validation period from
20062012. The SWAT model outputs further used are discharge for
the sh model chain and nitrate time series for the macroinvertebrate
chain for selected subbasins.
2.3.2. 1-D hydraulic model HEC-RAS
The SWAT model provides discharge time series as output but no
information on the water level. Since the water level at the downstream
end of the study river reach is required for the 2-D hydrodynamic
model, the 1-D hydraulic model HEC-RAS (USACE, 2002) (b in Fig. 1)
was used to develop a rating curve (discharge and corresponding
water levels). The HEC-RAS model requires discharge values and, in
the case of subcritical conditions, water level values at the downstream
end of the river segment as input data.
In this study, the HEC-RAS model was set-up for a 3 km river segment starting upstream from the hydrological station Sollerup to the
downstream end of the study river reach (see Fig. 2B). In June 2011,
42 cross-sections were mapped with a spacing of 50 to 100 m
representing the meandering river network of the river segment adequately. These cross-sections were imported to the HEC-RAS model
using HEC-GeoRAS (USACE, 2011). Measured ow velocities were
used to calibrate the roughness coefcient. A rating curve for the station
Sollerup was used as the lower boundary condition. Since there is

one relevant tributary to the Treene River within the modelled
river segment, additional discharge data was used from a station
(Sollerupmuehle), situated close to the conuence of that tributary
(Fig. 2).
2.3.3. 2-D hydrodynamic modelling
The 2D hydrodynamic model FaSTMECH (c in Fig. 1), which is part of
the modelling software iRIC (Nelson et al., 2010), was used to predict
the hydraulic habitat conditions (e.g., ow velocity, water depth). Similar to other 2D hydrodynamic models, FaSTMECH numerically solves
the basic mass conservation equation and two horizontal components
of momentum conservation to calculate two horizontal ow velocity
vectors and water depth.
In March 2011, during mean ow conditions, river channel bathymetry and ow velocity were mapped in the study reach at 54 crosssections with a spacing of max. 5 m (Fig. 2B). The mapped conditions
were used to set up the 2D hydrodynamic model in FaSTMECH with a
model-grid resolution of 0.25 m2. FaSTMECH provides ow velocity
and water depth for each raster cell as consecutive input for the sh
habitat model.
2.3.4. Fish habitat modelling
The GRASS GIS tool r.fuzzy.system (Jasiewicz, 2011) (d in Fig. 1) was
applied to assess and quantify the effects of changes in ow velocity and
water depth on the suitability of habitat conditions for sh in the study
reach based on species-specic habitat suitability rules. Habitat suitability rules are commonly based on empirical observations or expert
opinion (Bovee et al., 1998).
The applied fuzzy logic habitat model does not divide the habitat variables (e.g., ow velocity) into distinct classes (e.g., low, medium, high),
but utilise sliding thresholds instead, so called fuzzy-memberships
546
classes, to dene habitat suitability (Ahmadi-Nedushan et al., 2006;

Mouton et al., 2007). The inputs required by r.fuzzy.system are
(i) raster maps providing the hydraulic predictor parameters ow velocity and water depth, (ii) a le that denes the fuzzy memberships for the
input and output parameters (habitat suitability) and (iii) a rule le that
describes the relationship of all possible parameter combinations and
the resulting species- and/or life stage-specic habitat suitability. Fuzzy
rules were set by experts based on large eld data sets and ecological information obtained from the literature. Five fuzzy-memberships were
dened for both ow velocity and water depth (very low, low, medium,
high, very high). The species-habitat relationships considering all possible pairwise combinations of ow velocity fuzzy memberships classes
and water depth fuzzy memberships classes (n = 5 5 = 25), were developed for each sh species separately. For each combination of habitat
conditions, one out of four habitat suitability fuzzy-memberships classes
(no, low, medium, and high) was assigned. A nal model-internal
defuzzication (centroid method) step transformed these fuzzy output
memberships into a quantitative value of habitat suitability ranging
from zero (unsuitable) to one (highly suitable).
Habitat suitability was calculated for each cell of a regular GIS raster
of 25 25 cm based on the output of the 2D hydrodynamic computational grid. The suitabilities (ranging from zero to one) of all cells were
weighted by their area and summed up providing the Weighted Usable
Area (WUA, Bovee and Cochnauer (1977), Payne (2003)), a single value
of species-specic suitability for the entire study reach. The WUA ([m2])
was estimated using Eq. (1). Ai describes the area of a single model cell
and SIi the habitat suitability index of that respective cell.
WUA
n
X
Ai SIi
i1
In this study, the r.fuzzy.system was applied for three sh species:

Spined loach (Cobitis taenia), Eurasian minnow (Phoxinus phoxinus),
and Dace (Leuciscus leuciscus). These species have been selected,
because they are typical for the study area (Spratte and Hartmann,
1998; Schaarschmidt et al., 2005), but differ in their hydraulic habitat
requirements. Spined loach is a small-bodied (maximum length
14 cm), bottom-dwelling, riverine sh species which preferably colonises larger slow owing lowland rivers. It prefers medium depths
and low to medium ow velocities (Copp and Vilizzi, 2004). Eurasian
minnow is a small-bodied (maximum length 14 cm), riverine sh
species too, but preferably colonises smaller, faster owing rivers and
therein low to medium depths and low ow velocities (Mastrorillo
et al., 1996). Its habitat suitability is typically inversely related to
depth and ow velocity (Mastrorillo et al., 1997). Dace is a largebodied (maximum length N35 cm), riverine sh species preferring
deeper pools and therein especially the faster owing front parts of
pools. Generally habitat suitability of dace is more dependent on
P. phoxinus
1.5
water depth (preferring deep water) than on ow velocity, but tolerates

also higher ow velocities (Lamouroux et al., 1999; Vlach et al., 2005).
Habitat requirements of the selected species have been extracted
from several eld studies (Freyhof, 1998; Grift, 2001; Bischoff, 2002)
and the literature (Bless, 1992; Mann, 1996; Mastrorillo et al., 1997;
Grift et al., 2003; Scharf et al., 2011). Habitat suitability rules based on
water depth and ow velocity for all three sh species were set as illustrated in Fig. 3.
2.3.5. Macroinvertebrate habitat modelling
The recently developed Habitat Evaluation Tool (HET) (Kiesel et al.,
2015) (e in Fig. 1) was used to simulate the abundance of selected
macroinvertebrate species in the model reach. Its predictions are mainly
based on species-specic empirical habitat (here: substrates) abundance relationships. In this study, habitat conditions were described
using bottom substratum composition for two reasons. First, substrate
has been widely identied as one of the most important variables
explaining the presence and abundance of benthic macroinvertebrates
on a micro scale (Beisel et al., 1998). Macroinvertebrates are typically
associated with certain substrates due to various morphological and
physiological adaptations (Pardo and Armitage, 1997) as analysed in
previous studies (e.g., Schrder et al., 2013), since they provide important requisites such as shelter, food and reproduction.
Second, habitat-specic, i.e., substrate-specic, macroinvertebrate
samples are necessary to compile a generating dataset for the HET
model (for more details see Kiesel et al. (2015)).
As detected in a macroinvertebrate sampling campaign in March
2011, the bottom substrate of the study reach mainly consisted of
pure sand and the whole Treene River is dominated by sandy substrate
with small amounts of gravel. Therefore, discharge changes due to
climate and land use change and resulting sediment transport were
not expected to result in major changes of substrate composition, and
hence not modelled (see discussion in Kail et al. (2015)).
In contrast, nutrient loads were expected to markedly change in
climate and land-use scenarios and affect macroinvertebrates through
changes in oxygen concentrations. Therefore, the abundance predicted
by HET based on the present substrate composition was further developed to account for the effect of changed nutrient loads in the scenarios.
Nitrate-abundance relationships were derived for all species with more
than thirty occurrences in the samples from lowland streams available
in a national-wide German database (Umweltbundesamt, UBA).
Based on substrate and nitrate concentrations, the presence and
abundance of eight species were simulated. These species comprise
six taxonomic groups: Odonata (Calopteryx splendens), Diptera
(Chironomini Gen. sp., Prodiamesa olivcacea, Tanypodinae Gen. sp.),
Hirudinea (Erpobdella octoculata), Crustacea (Gammarus pulex), Bivalvia
(Pisidium sp.) and Ephemeroptera (Sialis lutaria). Whilst C. splendens,
E. octoculata and S. lutaria are classied as predatory species, G. pulex
primarily feeds on coarse organic material (Schmidt-Kloiber and
C. taenia
L. leuciscus
1.0
Water depth [m]
0.8
1.0
0.6
0.4
0.5
0.2
0.0
0.0
0.0
0.5
1.0
0.0
0.5
1.0
0.0
0.5
1.0
Flow velocity [m/s]

Fig. 3. Habitat suitability for three sh species in relation to water depth and ow velocity.
Hering, 2011). Pisidium sp., on the other hand, is an active lter feeder
predominantly living in soft, ne mineralic substrates such as sand or
mud, whereas all Diptera species are gatherers and feed on detritus
deposited on the river bottom (Schmedtje and Colling, 1996).
The approach how to consider nitrate loads in HET is exemplarily
described for Chironomini Gen. sp. in the following (Table 2). First,
species abundance was calculated based on the samples taken in the
study reach (here 67 individuals for Chironomini Gen sp.). Second, samples of the UBA database, in which the respective species was occurring,
were assigned to one of three classes according to the nitrate concentrations of the samples. Class boundaries were set at the 25% and the 75%
percentile. Thus, the three classes included samples with nitrate
concentrations b 25%, 2575% and N75% percentiles, respectively (see
example for Chironomini Gen sp. in Table 2). For each class, the 75% percentile of the abundance values of the respective samples was calculated as a proxy for the highest abundance occurring at the given range of
nitrate concentration.
Next, the measured nitrate concentration in the study reach as
observed during the eld campaign in March 2011 (3.58 mg/l) was
assigned to the corresponding nutrient class (e.g., class II for
Chironomini Gen sp.). Here, we assumed similar nitrate values for the
model reach and the measurement station Treia.
In order to estimate the species abundances during model simulations, species-specic weighting factors were obtained by allocating
the modelled nitrate value to one of the three nutrient classes for each
species. To obtain this, the mean abundance of the assigned nutrient
class was divided by the mean abundance of the nutrient class present
for a species at the day of sampling (see Table 2). The weighting factor
is one if the modelled nitrate loads are within the same nutrient class
as the value that has been measured during the sampling campaign in
2011 (class II for Chironomini Gen sp.). If the modelled values lie outside
the nutrient class of the 2011 measurements, the weighting factor leads
to an increase or a decrease of the average abundance of the modelled
species depending whether the nitrate concentrations are lower or
higher in the specic class. For Chironomini Gen sp., an increase in
nitrate leads to a reduction in the abundance, whilst a decrease in the
nitrate concentration results in a higher abundance.
To match scenario simulations and macroinvertebrate data, the
simulation results of the SWAT model were adapted to the data availability of the macroinvertebrates. To enable a temporal comparison
between the modelled nutrient loads and the nutrient conditions
during the macroinvertebrate sampling campaign, mean modelled
values were calculated for the rst week of March (01.07.03) for the
years 2021 to 2030. Thus, changes in macroinvertebrate abundance
were predicted if the modelled nutrient concentrations fall outside the
boundaries of the class containing the nutrient concentration as
observed during macroinvertebrate sampling (March) in at least one
year during the simulation period.
547
2.4.1. Climate change scenario

An increase in temperature and variations in the precipitation
regime due to changing climate conditions inuence the hydrological
cycle with seasonal variations (Bronstert et al., 2007; Huang et al.,
2010; Setegn et al., 2011; Singh et al., 2011). Future climate scenario
data from the STAtistical Regional model (STAR) (Orlowsky et al.,
2008) were used as climate input data for the SWAT model (f in
Fig. 1). STAR is a statistical Regional Climate Model (RCM), which is
based on a resampling approach and driven by a linear increase of the
mean annual temperature. The STAR model assumes that the typical
characteristics of the future climate resemble those of the recent data
also in the case of increasing temperatures (Orlowsky et al., 2008).
The STAR model provides different scenarios with temperature
increases between 0 K and 3 K. To analyse the maximum differences,
the two extreme STAR scenarios were used in this study with no temperature increase (0 K) (baseline scenario) and with a linear increase
in the temperature of up to 3 K until 2060. Fig. 4 shows higher temperatures in the climate change (3 K)-scenario and similar mean annual
precipitation in both scenarios. STAR includes 100 realisations of each
scenario allowing an analysis of the variability within one scenario.
Since the simulation period ends in 2030, the average temperature in
this period is only 1 K higher in the climate change scenario compared
to the 0 K-scenario.
2.4.2. Land use change scenario
Three land use scenarios were developed with a data-based approach by analysing recent changes in the Treene catchment (Geertz,
2012) and based on data on the share of different agricultural crops
from 2003, 2007 and 2010 at the community level provided by the
regional statistical administration (Statistik Nord) (g in Fig. 1).
In the food scenario, rising food prices on the global market were
assumed resulting in an increase in the share of corn and wheat
(Fig. 5). The low corn-scenario was related to a replacement of corn
by rape and pastures. Subsequently, the best practise scenario was
related to environmental conditions and was focused on a structural
diversity in agricultural crops. This scenario is thus characterised by
the most balanced crop shares.
In order to consider the input requirements of the SWAT model, the
information on agricultural crops was transformed from the community
level to the subbasin level of the SWAT model under consideration of
the spatial crop distribution. The crop shares which were available
for each community were implemented into the SWAT model as ve
different sequences of crops, i.e., crop rotations. Thus, the contributions
of the ve crop rotations in each subbasin represents the spatial
distribution of the seven crop shares according to the statistical data.
The simulation in the SWAT model was started with the land use
situation in 2003. The shares of the crop rotations were updated
2.4. Scenario description

Climate change and land use scenarios were simulated to assess the
impact of changing future conditions on hydrology, hydraulics and biota
at the catchment scale. In order to allow a reliable comparison between
both drivers, a simulation period from 2021 to 2030 was selected.
Table 2
Example of the macroinvertebrate calculation in HET for Chironomini Gen sp.
Nitrate classes
II
III
Thresholds in nitrate concentration [mg/l]

Abundances [ind/0.25 m2]
Weighting factor []
01.39
104.0
1.48
1.404.14
70.3
1
N4.14
62.4
0.89
Fig. 4. Boxplots of mean annual precipitation and mean daily temperature for the two
STAR scenarios (baseline, change) for the period from 2021 to 2030. Each boxplot is
based on the 100 simulations of the scenario.
Share in %
548
60
50
40
30
20
10
0
60
50
40
30
20
10
0
Corn
Rape
60
50
40
30
20
10
0
60
50
40
30
20
10
0
Wheat
Pasture
Rye
60
50
40
30
20
10
0
60
50
40
30
20
10
0
60
50
40
30
20
10
0
Barley
Forage plants
Baseline
Food
Low corn
Best practice
Fig. 5. Relative contribution of different crops within the three future land use scenarios (2030) and the baseline scenario (2010) in the Treene catchment.
dynamically 2007 and 2010 based on recent data so that the prevailing
crop distributions in these years were considered in the model simulations. Thus, the contribution of each agricultural crop was modied during the model simulation. For the future period, the shares of the crop
rotations were updated continuously in every year from 2013 to 2030
in order to implement a realistic temporal dynamic in the agricultural
crops.
To obtain a model response to land use change only, the baseline
scenario with no temperature increase (0 K-scenario) was used for the
land use scenarios. Thus, the same climatological conditions as for the
recent period were assumed.
2.5. Scenario simulations along the model cascade
Climate and land use change scenarios were developed at the catchment scale and simulated with the SWAT model and evaluated for the
period from 20212030 to use the same period for both scenarios. The
scenario simulations were adapted to the characteristics of the models
of the two model chains as presented in Fig. 1.
In the simulation of the sh process chain (Fig. 1), high (Q25) and
low (Q75) discharges were considered besides median discharge
(Q50) due to the ecological importance of these ow conditions
(Richter et al., 1996; Poff et al., 1997). The three discharge magnitudes
were used to calculate hydraulic variables and the resulting weighted
usable area (WUA) for the sh species. The model results were provided
in monthly resolution to consider seasonal differences. All 100 simulation runs of both climate change scenarios and the three single simulations of each of the three land use scenarios were used in this process
chain. The results of the 100 simulations were averaged in a monthly
resolution for the simulation period.
In contrast, for modelling the effect of changes in nutrient concentrations on macroinvertebrates in the climate and land use scenarios, one
representative simulation out of the 100 climate change simulation
runs was selected. Among the 100 simulations of the baseline (0 K)
and climate change scenario, a STAR data set with an average mean
annual precipitation was used as proposed by Hesse et al. (2008). The
STAR simulation run from the 0 K-scenario with an average mean
annual precipitation was used as baseline in the three land use change
scenarios.
Averaging the nutrient concentrations of the 100 climate change
simulation runs would have resulted in similar absolute values in nutrient concentrations in March of each year within the simulation period.
However, in the HET model, changes in macroinvertebrate abundance
can only occur if the modelled nutrient concentrations fall outside the
class boundaries in at least one year during the simulation period. In
the case of an averaging of all 100 simulations the class boundaries are
not exceeded and thus the predicted abundances are identical. To
increase comparability with the single runs of the land use change
scenarios, a single representative climate change simulation run was
selected.
3. Results
3.1. Model simulations for recent conditions
The comparison of measured and modelled time series for discharge
and nitrate showed a good representation of hydrological and water
quality conditions by the SWAT model. In the case of discharge, the
NashSutcliffe Efciency Index (Nash and Sutcliffe, 1970) was 0.72
and 0.79 for the calibration period, as well as, 0.71 and 0.80 for the
validation period for the hydrological station Sollerup at the lower end
of the river segment and the catchment outlet Treia, respectively. The
Percent Bias (PBIAS) was between 8.2% to 3.2% and 1.4% to 0.2%
for the calibration and validation period, respectively (see also Guse
et al. (2014a)). The measured nitrate time series was satisfyingly
reproduced by the SWAT model with a NashSutcliffe Efciency Index
of 0.62 and 0.74 and a PBIAS of 1% and 8% for the calibration and
validation period, respectively. Thus, the SWAT model results for
discharge and nitrate could be used for the next models within the
model cascade.
The HEC-RAS model was calibrated to a roughness of Mannings n =
0.03 s/m1/3 which is a plausible value according to Chow (1959). The
nal output was a rating curve for the upstream end of the river
segment, corresponding to the downstream end of the study reach
that served as input for the 2-D hydrodynamic model.
For the 2-D hydrodynamic model, the best t between mapped and
modelled water levels (mean difference 6 mm, max. 15 mm) was
obtained using the initial roughness value back-calculated from
measured velocities (Manning n = 0.058). This can be considered as a
good t given the mapping accuracy of 5 mm, the maximum error of
discharge (1.7%) being well below the acceptable error of 3% given in
the FaSTMECH manual.
3.2. Baseline scenario simulation with constant land use conditions
The results of the baseline scenario simulation with no temperature
increase and constant land use during the modelling period
(20212030) were summarized for all models in Fig. 6. The mean
monthly discharge was highest in winter (6 m3/s) and decreases in
late summer and autumn to about 1 m3/s in September. Similarly,
nitrate showed high concentrations in winter and spring (in maximum
5 mg/l in April) and lower values in summer (2.5 mg/l).
Flow velocity and water depth showed the same pattern with
maximum values in January and December. Low ow velocities were
predicted from July to October. The mean monthly values of ow

velocity ranged from 0.2 m/s in September and October to 0.45 m/s in
January. Mean monthly ow depth varied between 0.45 m in October
and 1.1 m in January.
The resulting weighted usable area (WUA) showed two different
patterns for the three sh species. For C. taenia and L. leuciscus,
the WUA was high in January (about 2000 m2) and decreased until
September (8001200 m2) before it increased again. In contrast, for
P. phoxinus, the WUA increased with decreasing discharge in the rst
half of the year, remained rather stable from June to October (about
1200 m2) and decreased in NovemberDecember.
Among the predicted macroinvertebrate species G. pulex dominated
as a rather ubiquitous generalist species (abundance of 350 ind/
0.25 m2), whilst other species were predicted being less abundant,
e.g., S. lutaria and E. octoculara.
3.3. Model simulations of climate and land use change
The results of the climate and land use scenario simulations were
presented successively for the models in order of their occurrence
in the model cascade as numbered from a to e in Fig. 1. All model simulation results were presented as relative differences to the baseline
scenario which had the same climate and land use conditions as in the
recent modelling.
3.3.1. Hydrological model simulations
In the scenario simulations, discharge only changed to a relevant
extent in the climate scenario, whilst the nitrate concentrations are
affected both in the climate and land use scenarios. In the climate
scenario, the discharge for the hydrological station Sollerup was lower
compared to the baseline scenario for the three selected discharge
magnitudes (Q25, Q50, Q75) from March to December (Fig. 7). The largest differences occurred in autumn (25% in October). With decreasing
discharge magnitude, the reduction of the discharge occurred later in
the year.
Nitrate concentrations were smaller in the climate scenario, with
the largest relative decrease in late summer and autumn (August to October). The seasonal patterns were similar for all three discharge magnitudes but the largest decrease of nitrate concentration occurred during
low ow conditions (Q75) in September (50%).
Hydrology
549
Water quality
5
Nitrate [mg/l]
Discharge [m3/s]
5
4
3
2
4
3
2
1
1
0
0
1
10 11 12
Month
Flow velocity
10 11 12
10 11 12
Water depth
1.0
0.4
Water depth [m]
Flow velocity [m/s]
Month
0.3
0.2
0.1
0.0
0.8
0.6
0.4
0.2
0.0
10 11 12
Month
Fishes
2500
C. taenis
Months
Macroinvertebrates
L. leuciscus
P. phoxinus
Tanypodinae Gen. sp.

S. lutaria
2000
WUA [m]
P. olivacea
1500
Pisidium sp.
G. pulex
1000
E. octoculata
500
Chironomini Gen. sp.

C. splendens
0
1
Month
10 11 12
50
100 150 200 250 300 350
Abundances [ind/0.25m]
Fig. 6. Results of the baseline scenario (20212030) for medium ow conditions (Q50). Discharge and nitrate are shown for the hydrological station Sollerup. The hydrodynamic variables
and the sh habitat suitability (WUA) as well as the abundances of macroinvertebrates are shown for the study reach.
550
10
0
10
20
30
40
50
10
5
0
5
10
15
20
25
10
0
10
20
30
40
50
10
5
0
5
10
15
20
25
10
0
10
20
30
40
50
10
11
12
Q50
10
5
0
5
10
15
20
25
Q25
Nitrate
Q75
Rel. Diff. to Baseline [%]
Discharge
Months
10
11
12
Months
Climate
Food
Low corn
Best practise
Fig. 7. Differences in mean monthly estimates of discharge and nitrate for three discharge magnitudes (high: Q25, medium: Q50 and low ows: Q75) between the future climate and the
three land use scenarios (Food, Low corn, Best practise) compared to the baseline scenario for the hydrological station Sollerup for the period 20212030.
The model simulations of the three future land use scenarios showed
no relevant change in discharge (b 5%, Fig. 7), but substantial differences
in the nitrate concentrations. Nitrate was predicted to slightly increase
in the food scenario (mean monthly values + 4%) and to decrease
in the low-corn ( 19%) and the best practise scenario ( 31%).
The highest differences to the baseline scenario were predicted for the
best practise scenario at low ow conditions (Q75) in August with a
decrease by 43%.
3.3.2. Hydrodynamic model simulations

The monthly patterns of the hydraulic variables followed the impact
of climate and land use change on discharge. However, there were
partly differences between the effects on the discharge and hydraulic
variables. The values of the hydraulic variables were similar in baseline
and climate change scenarios in the rst half of the year with relative
differences less than 5% (Fig. 8). An increase of ow velocity and
water depth was only predicted for January and February. In the second
10 11 12
Q50
Q25
Water depth
10
5
0
5
10
15
20
25
10
5
0
5
10
15
20
25
10
5
0
5
10
15
20
25
Q75
Flow velocity
10
5
0
5
10
15
20
25
10
5
0
5
10
15
20
25
10
5
0
5
10
15
20
25
Months
Climate
10 11 12
Months
Food
Low corn
Best practise
Fig. 8. Percent differences between climate and the three land use scenarios (Food, Low corn, Best practise) in relation to their baseline scenario for ow velocity and water depth for three
different discharge magnitudes (high: Q25, medium: Q50 and low ows: Q75) for the simulation period from 2021 to 2030.
half of the year, differences were markedly higher. For high ows (Q25),
the two hydraulic variables were lower in the climate scenario with
the maximum differences from August to October (about 10%). At
medium ows (Q50), the maximum differences were predicted to
occur in October ( 12% for ow velocity). A more complex pattern
was observed for low ows (Q75). The largest decrease in ow velocity
occurred in October with the largest relative differences of all three
discharge magnitudes of N20%. However, in October, the water depth
is higher in the climate scenario. The highest reductions in water
depth were predicted in the climate scenario in September ( 20%)
and November ( 10%). In this context, it has to be considered that
absolute water depth is lowest in September with an average depth of
0.4 m under baseline conditions (Fig. 6).
Overall, the impact of the land use change scenarios on the hydraulic variables was low (Fig. 8). In the food scenario, an increase
in ow velocity and water depth was detected from August to November. The largest increase in ow velocity occurred for low ows
(Q75) (5%) in October. The opposite trend was estimated in the
low corn scenario. Here, ow velocity and water depth decreased
in autumn. Similarly, the maximum relative difference was observed
for the Q75 in October ( 3%). An increase in the hydraulic variables
was only observed in January and February. The lowest differences
were detected for the best practise scenario whose results resemble
the baseline scenario.
was mostly increasing throughout the year and up to 10% change in

October. At low and medium ows the WUA changed seasonally
between increase and decrease, with both, highest decrease and
increase predicted at Q50 in summer (JulySeptember) and autumn
(OctoberNovember), respectively.
The different land use scenarios had substantially lower effects on
WUA compared to climate change. For C. taenia and L. leuciscus, the
relative difference was always less than 5% for the three land use scenarios. Although at low levels WUA was predicted to increase following the
food scenario and to decrease in the other two in particular from August
to October. In contrast, the highest loss of WUA for P. phoxinus was
observed in the food scenario at high ows in March (Fig. 9) when the
impacts of land use change on WUA well exceeded those of climate
change.
3.3.4. Macroinvertebrate habitat modelling
The abundances of the eight macroinvertebrates species were affected both in the climate and land use scenarios. The response of macroinvertebrates to nitrate concentrations was strongly species dependent.
Fig. 10 shows that the ve species in the rst and second column
showed different responses to the variations in nitrate concentration
(see rst panel in Fig. 10) in future scenarios. The impact of changing
conditions on macroinvertebrates showed a distinct pattern: Whilst
the mean abundance of most species (e.g., G. pulex, Pisidium sp.,
P. olivacea) decreased with decreasing nitrate concentrations, those
of Tanypodinae Gen. sp. tended to increase.
In contrast, three species E. octoculata, C. splendens and S. lutaria
(third column of Fig. 10) were not correlated to nitrate concentration.
Here, the modelled abundances of the species were partly increasing
and partly decreasing without a clear pattern. A particular decrease
was simulated in the climate scenarios for C. splendens ( 12%) and
S. lutaria (15%).
Concerning the land use scenarios, the differences in species
abundances were higher in both, the low corn and best practise scenarios, compared to the food scenario. Even though the direction of change
was species-specic, the magnitude of change was smaller in the food
scenario. For most species, all three land use scenarios showed the
3.3.3. Fish habitat modelling

Overall, sh was stronger affected by climate change than by land
use change with differences in the impacts on the weighted usable
area (WUA) of the three species (Fig. 9). In regard to the climate change
scenario, the results of the sh habitat suitability simulations were
similar for C. taenia and L. leuciscus and differed for P. phoxinus.
C. taenia and L. leuciscus were predicted to exhibit an increase in WUA
in January and February and then a decrease until November with the
highest reduction of WUA in September (Fig. 9). The only exception
was an increase of WUA at Q75 in October.
In contrast, the predicted impact of climate change on habitat
suitability of P. phoxinus was generally lower. At high ows, the WUA
P. phoxinus
10
5
0
5
10
15
20
25
10
5
0
5
10
15
20
25
10
5
0
5
10
15
20
25
10
5
0
5
10
15
20
25
10
5
0
5
10
15
20
25
10
5
0
5
10
15
20
25
10
5
0
5
10
15
20
25
10 11 12
Months
10 11 12
Months
Climate
Food
Q50
10
5
0
5
10
15
20
25
Q25
L. leuciscus
10
5
0
5
10
15
20
25
Q75
C. taenia
551
10 11 12
Months
Low corn
Best practise
Fig. 9. Comparison of relative differences of monthly WUA for three sh species for the three discharge magnitudes (high: Q25, medium: Q50 and low ows: Q75) between the climate and
the three land use (Food, Low corn, Best practise) change scenarios and the baseline scenarios.
552
20
20
20
0
20
0
20
S. lutaria
20
20
P. olivacea
Climate
20
20
20
C. splendens
20
20
0
20
20
20
Pisidium sp.
Tanypodinae Gen. sp.
Chironomini Gen. sp.
20
E. octoculata
20
G. pulex
20
Nitrate
Food
Low corn
Best
Fig. 10. Relative differences of macroinvertebrate species abundances of the future scenario to the baseline scenario for the period 20212030 for rst seven days in March. The food, low
corn and best practise scenarios are compared with the land use baseline scenario. The climate scenario is compared with constant climate (0 K-scenario). In addition, the upper left panel
shows the corresponding changes in nitrate for the same period.
same direction of change, i.e., either an increase or decrease in

abundance, whilst the magnitude of change was highly speciesspecic. In contrast, the direction of change for the specic species
differed between climate and land use scenarios. The highest variation
of modelled abundances was observed for Tanypodinae Gen. sp. in the
low corn (10%), best practise (30%) and climate scenario (10%). The
low corn and best practise scenarios yielded the highest increase in
species abundances.
In contrast, the land use scenarios had only minor effects on

discharge since in this study the major components of the hydrological
cycle were not markedly affected by the changes in crop cultivations.
Land use scenarios affected nitrate concentrations. Depending on the
specically assumed changes in crops, nitrate loads were predicted
to increase in the food scenario and to decrease in the low corn and
best practise scenario. These changes in nitrate loads were caused by
different amounts of fertilizer applied or soil coverages during the
year between the crop rotations.
4. Discussion
4.1. Impacts of climate and land use change on discharge and nitrate
4.2. Impacts of climate and land use change on ow velocity and

water depth
The higher temperature in the climate scenario led to an increase in

evapotranspiration resulting in a reduction of the percolation into soil
and groundwater. Since groundwater is the dominant runoff component in this lowland catchment (Kiesel et al., 2010; Pfannerstill et al.,
2014a, 2014b), the total runoff was reduced in the climate change
scenario. Lower discharges due to higher evapotranspiration were
predicted for all three ow conditions (Q25, Q50, Q75), since there
was an overall lower amount of available water to feed the river. A
high impact of climate change was predicted on discharge with major
changes for low discharge conditions in late summer and autumn.
Lower discharges at the end of the summer and autumn were also detected in other climate change studies in different regions in Germany
(Hattermann et al., 2004; Huang et al., 2010). Due to the dominance
of groundwater contribution in the Treene catchment, river ow was
especially reduced in these months when the water storages were
depleted before the recharge in winter. Climate change thus is expected
to reduce the discharge especially in the driest phase of the year which
leads to more extreme hydrological conditions. Periods of low ows
were longer in the climate change scenario.
Nitrate loads also changed due to the impact of climate change on
discharge. Different precipitation regimes inuenced the runoff components such as surface runoff and thus, the input of nitrate into the rivers
(Kundzewicz et al., 2008).
In general, the modelled hydraulic conditions reected the discharge

changes predicted by the eco-hydrological model, with all three
discharge magnitudes (Q25, Q50, Q75) showing the same seasonal
pattern. However, the hydraulic changes cannot be directly deduced
from the hydrological conditions. The percentage change of the two
hydraulic variables was lower compared to discharge changes. Moreover, hydraulic variables did not change proportional to discharge. For
example, whilst the largest reduction in discharge during low ows
(Q75) caused by climate change was predicted to occur in October,
the largest decrease of ow depth was predicted for September. This
was probably caused by the asymmetric river cross-sections with a
deep pool in the outer bend and a shallow bar in the inner bend.
The bar is inundated at higher ows and these shallow areas result in
a relatively small mean depth. At lower discharges, the shallow bar
areas get dry, and mean depth is calculated based on a much
narrower cross-section mainly consisting of the rather deep pool
area, resulting in an increase of mean depth. Variations in the impact
of climate change on hydraulic variables under different morphological conditions at river cross-sections were also emphasised by Hauer
et al. (2013). The strongest decrease in ow velocity and water depth
in the climate scenario was estimated in autumn, where both hydraulic variables reach its minimum value. Thus, the lowest values
were further reduced.
Hydraulic variables thus have distinct patterns and even though the
general trend is similar to the discharge changes, it is necessary to
model the hydraulic habitat conditions to assess the effect of climate
and land use changes on abiotic habitat conditions.
4.3. Impacts of climate and land use change on shes
The predicted changes in WUA for shes corresponded to the different combinations of shallower or deeper and lower or faster owing
water based on changes in discharge under future climate and land
use conditions. All three modelled riverine sh species were affected
from the predicted decrease in discharge and the corresponding reduction of ow velocity (Bless, 1992; Mann, 1996; Grift et al., 2003).
Furthermore, L. leuciscus (Mann, 1996) and C. taenia were affected by
the decrease of deep pool areas (Bohlen, 2003), that is better tolerated
by P. phoxinus (Bless, 1992; Mann, 1996; Mastrorillo et al., 1997). The
latter was reected in the opposing response of P. phoxinus to climate
change still gaining WUA at decreasing discharge whilst the other two
species already started losing WUA. For P. phoxinus WUA typically
increased until water depth and ow velocity were too low and
exceeded lower tolerance thresholds.
Consequently, the climate change scenarios predicted a discharge
reduction for summer that was below a critical threshold and resulted
in a general decline of available aquatic habitats and thus, in a loss of
WUA even for P. phoxinus in August and September. Thus, habitat suitability is expected to decrease due to the changing climate conditions.
Furthermore, the model simulations clearly demonstrated a higher
impact of climate change on the WUA for sh than land use change. It
should be noted that the process chain for sh focused on the effect of
discharge changes on habitat suitability solely. Land use scenarios had
a minor effect on discharge and hydraulic habitat conditions but markedly affected water quality (nitrate concentrations). However, the
predicted nitrate changes were well in the tolerance range of the
modelled sh species (Wolter et al., 2003; Schinegger et al., 2013) and
thus not considered a potential limiting factor for sh (Wolter et al.,
2003).
Correspondingly, changing habitat suitability for sh with decreasing discharge and the resulting changes of the hydraulic habitat
conditions were also observed by Hauer et al. (2013). The impact of
changing habitat conditions is very complex and the effects of climate
and land use change on the suitability for sh is highly speciesspecic. The same changes in hydraulic variables can have a different
impact on sh due to the varying habitat preferences and life history
traits.
Whilst the simulation results for sh seem plausible, they could not
be validated since species- and habitat-specic sh samples were not
available for the study reach.
Moreover, it must be noted that this assessment was solely based on
the depth and ow preferences of the species, whilst riverine sh are
often markedly limited by the availability of gravel substrates for
spawning (Mann, 1996; Wolter et al., 2003; Schinegger et al., 2013),
which was not considered in this study since the bottom substrate of
the study reach was pure sand and only small gravel patches were
found in the Treene River.
4.4. Impacts of climate and land use changes on macroinvertebrates
The relationship of species abundances and nitrate concentration
was species-dependent. The macroinvertebrate species were classied
into two groups according to the feeding preferences. The rst group
comprised species with feeding preferences supposedly related to the
availability of organic matter. Among these species were shredders
(e.g., G. pulex) and deposit feeders (e.g., Chironomini Gen. sp.,
P. olivacea, Tanypodinae Gen sp.). The abundances of these species
were related to the nitrate concentration in the scenario simulations,
since higher nitrate concentrations might lead to an increase in primary
553
production (Burgin and Hamilton, 2007; Camargo and Alonso, 2006).

Although Chironomini Gen. sp. and Tanypodinae Gen sp. were also
detritus feeders, their abundances decreased with increasing nutrient
levels. This may be related to the fact that Chironomini Gen. sp. and
Tanypodinae Gen sp. are broad taxonomic levels, potentially comprising
sensitive species which may be diminished in numbers under high
nutrient loads. On the other hand, E. octoculata, C. splendens and
S. lutaria were classied as predatory species, primarily feeding on
smaller macroinvertebrates. Abundances of these species were not
related to varying nitrate concentrations as they are not dependent on
organic food sources and may thus be more strongly inuenced by
other habitat factors (Schmedtje and Colling, 1996; Schmidt-Kloiber
and Hering, 2011). Thus, the differences in abundances largely varied
in their direction between the species. To understand the impact of
the scenario simulations on the species, it has to be highlighted that
abundance values only changed if the boundaries between the three
nitrate classes were crossed in at least one out of the ten years of the
evaluation period (20212030).
In this study, we focused on substrate to describe the habitats of
macroinvertebrates for an ecological and a more technical reason.
First, we considered bottom substrate rather than hydraulic variables
(e.g., water depth) as a superior variable in describing bottomdwelling macroinvertebrates, which is in contrast to sh that might
inhabit the entire water column. Only turbulences directly above the
stream bottom are likely to inuence the macroinvertebrates assemblage directly, e.g., through the dislocation of specimens (Reid and
Thoms, 2008). Extreme hydrological events (e.g., drought, oods) may
signicantly alter the macroinvertebrate community and affect species
richness and abundances (Stubbington et al., 2015). However, drastic
changes in the discharge regime were not predicted by the abiotic
models within the present study. Thus, the changes in discharge as
calculated in the different scenarios were not likely to signicantly
inuence the macroinvertebrate community as opposed to the sh
community. Considering other habitat parameters besides substrate
composition would most probably improve model performance, and
could be included. Second, habitat-specic samples are needed for
model calibration. However, ecologically relevant data for different
hydraulic habitat conditions due to changes in discharge (e.g., low, medium, high ow) are normally not provided by biological sampling
campaigns since they are limited by the prevailing discharge conditions
during eld sampling.
A validation for this model could not be performed since this would
require substrate samples with associated nitrate measurements from
the Treene or a comparable river type, which were not available. In
the case of more data, it might be possible to also include more nitrate
classes to obtain a more continuous pattern. Overall, the model results
rather showed a tendency for selected species without reporting actual
species abundances.
4.5. Applicability of the modelling approach in river management and
transferability of the results
As all models, the modelling cascade is a simplied representation of
the real world. This model cascade does consider some important but by
far not all variables affecting the presence and abundance of sh and
macroinvertebrates. Moreover, such complex modelling cascades that
couple different models are still in their infancy. Their application is
data and labour intensive, and the transferability of the empirical relationships used for the biological models (e.g., HET) is limited to similar
catchments. Therefore, the modelling cascade is presently not suited
to exactly predict the presence and abundance of sh and macroinvertebrates in a large number of catchments. However, additional variables
(e.g., water temperature) can be included in future application in the
modelling cascade. Its application in some case-study catchments has
already the potential to help identifying the main factors that limit
sh and macroinvertebrates in principle as well as predicting general
554
trends, i.e., evoked by changes in climate and land use. Such case-study
applications could already give important insights into many elds of
river ecology and management, for example by comparing the (cost) effectiveness of restoration measures between different spatial scales in
the light of climate and land use change. Moreover, data availability increases supported by the monitoring programme of the EU Water
Framework Directive (WFD). Furthermore, ecohydrological models
are applied to an increasing number of catchments focusing on different
aspects (e.g., ood protection and WFD implementation). Therefore,
such modelling cascades have the potential to be widely applicable in
future.
5. Conclusion
In this study, different models were coupled to predict the effects of
climate and land-use changes on discharge, water quality (nitrate
loads), hydraulics, and the resulting habitat suitability in a nearnatural study reach for selected sh and macroinvertebrate species.
The simulation results emphasise the usefulness of process-based
models to predict the impact of large-scale stressors on river biota.
The model cascade includes as many factors as possible to reproduce
the abiotic-biotic process chain. At the same time the model cascade
was maintained technically feasible. Within the model cascade, separate
process chains were distinguished for sh focusing on hydraulic habitat
changes, and for macroinvertebrates focussing on water quality.
Simulations of land use and climate change scenarios along
the model cascade illustrate similar trends but differences in the
response of hydrology and hydraulics. Changes in discharge cannot be
linearly transformed into changes of hydraulic variables and habitat
suitability.
The scenario simulations emphasised highly species-specic responses that are markedly affected by the complexity of the speciesspecic habitat-sh interactions. Furthermore, a reduction in nitrate
might result in an increase as well as in a decrease in the abundances
of macroinvertebrate species, depending on the species-specic
sensitivity.
These results reveal the complexity of the underlying cause-effect
chains, which require a rm insight in hydrological and hydraulic
processes as well as in habitat preferences of the targeted organism
groups and species.
The scenario simulations have illustrated the capability and exibility of the model cascade proposed by Kail et al. (2015). The model cascade is constructed to be generally applicable to different study areas
and various research questions.
Acknowledgements
We thank Schleswig-Holstein Authority for National Parks, Coastal
& Ocean Protection (LKN-SH) for the discharge and water level data,
the German Weather Service (DWD) for the climate data and the
Potsdam Institute for Climate Impact Research (PIK) for the STAR data.
Furthermore, we thank the Wasser- und Bodenverband and the Land
Schleswig-Holstein (LAND-SH) for the Digitales Anlagenverzeichnis
(DAV).
This study is part of the IWRM-net project IMPACT and has been
funded by the German Federal Ministry for Education and Research
(BMBF) (grant number 02WM1134 (IGB), 02WM1135 (UDE),
02WM1136 (CAU)). We thank all other partners from the IMPACT
project for the fruitful discussions.
We thank Dalibor Sulc from the Technical University of Prague
for setting up the HEC-RAS model and Thea-Lina Mller and the laboratory crew of the Institute of Natural Resources Conservation of the
Christian-Albrechts-University of Kiel for carrying out the water quality
analysis.
We would like to thank the community of the open source software
R, which was used for the majority of the analysis.
References
Ahmadi-Nedushan, B., St-Hilaire, A., Berube, M., Robichaud, E., Thiemonge, N., Bobee, B.,
2006. A review of statistical methods for the evaluation of aquatic habitat suitability
for instream ow assessment. River Res. Appl. 22, 503523.
Allan, J.D., Johnson, L.B., 1997. Catchment-scale analysis of aquatic ecosystems. Freshw.
Biol. 37, 107111.
Arnold, J., Fohrer, N., 2005. SWAT2000: current capabilities and research opportunities in
applied watershed modelling. Hydrol. Process. 19, 563572.
Arnold, J., Srinivasan, R., Muttiah, R., Williams, J., 1998. Large area hydrologic modeling and assessment part I: model development. J. Am. Water Resour. Assoc. 34,
7389.
Beechie, T.J., Sear, D.A., Olden, J.D., Pess, G.R., Bufngton, J.M., Moir, H., Roni, P., Pollock,
M.M., 2011. Process-based principles for restoring river ecosystems. Bioscience 60
(3), 209222.
Beisel, J.-N., Usseglio-Polatera, P., Thomas, S., Moreteau, J.-C., 1998. Stream community
structure in relation to spatial variation: the inuence of mesohabitat characteristics.
Hydrobiologia 389 (1), 7388.
BGR, 1999. Bodenbersichtskarte im Massstab 1:200000, Verbreitung der
Bodengesellschaften. Bundesanstalt fr Geowissenschaften und Rohstoffe, Hannover.
Bieger, K., Hrmann, G., Fohrer, N., 2013. The impact of land use change in the Xiangxi
catchment (China) on water balance and sediment transport. Reg. Environ. Chang.
http://dx.doi.org/10.1007/s10113-013-0429-3.
Bischoff, A., 2002. Juvenile Fish Recruitment in the Large Lowland River Oder: Assessing
the Role of Physical Factors and Habitat Availability. Shaker Verlag, Aachen.
Bless, R., 1992. Einsichten in die kologie der Elritze Phoxinus phoxinus (L.). Schriftenreihe
fr Landschaftspege und Naturschutz 35 pp. 168.
Blschl, G., Ardoin-Bardin, S., Bonell, M., Dorninger, M., Goodrich, D., Gutknecht, D.,
Matamoros, D., Merz, B., Shand, P., Szolgay, J., 2007. At what scales do climate
variability and land cover change impact on ooding and low ows? Hydrol. Process.
21, 12411247.
Bohlen, J., 2003. Untersuchungen zur Autkologie des Steinbeissers, Cobitis taenia (Phd
thesis). Humboldt-University, Berlin.
Bovee, K.D., Cochnauer, T., 1977. Development and evaluation of weighted criteria,
probability-of-use curves for instream ow assessments: sheries. Instream Flow Information Paper 3, FWS/OBS-77/63. Fish and Wildlife Service, U.S. Department of the
Interior, Washington, D.C. (38 pp. Available: https://www.fort.usgs.gov/publication/
353. Accessed 01. September 2014).
Bovee, K., Lamb, B., Bartholow, J., Stalnaker, C., Taylor, J., Henriksen, J., 1998. Stream
habitat analysis using the in-stream ow incremental methodology. U.S. Geological
Survey, Biological Resources Div., Information and Technology Report, USGS/BRD
19980004.
Bronstert, A., Kolokotronis, V., Schwandt, D., Straub, H., 2007. Comparison and evaluation
of regional climate scenarios for hydrological impact analysis: general scheme and
application example. Int. J. Climatol. 27, 15791594.
Burgin, A., Hamilton, S.K., 2007. Have we overemphasized the role of denitrication in
aquatic ecosystems? A review of nitrate removal pathways. Front. Ecol. Environ. 5
(2), 8996.
Camargo, J., Alonso, A., 2006. Ecological and toxicological effects of inorganic nitrogen pollution in aquatic ecosystems: a global assessment. Environ. Int. 32,
831849.
Ceola, S., Bertuzzo, E., Singer, G., Battin, T.J., Montanari, A., Rinaldo, A., 2014. Hydrologic
controls on basin-scale distribution of benthic invertebrates. Water Resour. Res. 50.
http://dx.doi.org/10.1002/2013WR015112.
Chow, V.-T., 1959. Open-Channel Hydraulics. McGraw-Hill.
Copp, G.H., Vilizzi, L., 2004. Spatial and ontogenetic variability in the microhabitat use of
stream dwelling spined loach (Cobitis taenia) and stone loach (Barbatula barbatula).
J. Appl. Ichthyol. 6, 440451.
Dahm, V., Hering, D., Nemitz, D., Graf, W., Schmidt-Kloiber, A., Leitner, P., Melcher, A., Feld,
C., 2013. Effects of physico-chemistry, land use and hydromorphology on three riverine organism groups: a comparative analysis with monitoring data from Germany
and Austria. Hydrobiologia 704, 389415.
Feld, C.K., Hering, D., 2007. Community structure or function: effects of environmental
stress on benthic macroinvertebrates at different spatial scales. Freshw. Biol. 52,
13801399.
Freyhof, J., 1998. Strukturierende Faktoren fr die Fischgemeinschaft der Sieg. Cuvillier
Verlag, Gttingen.
Geertz, J., 2012. Land Use Scenarios for the River Treene Catchment in Northern Germany
(Bachelor thesis). Brandenburgische Technische Universitt Cottbus.
Grift, R., 2001. How Fish Benet from Floodplain Restoration along the Lower River Rhine
(PhD thesis). Wageningen University.
Grift, R., Buijse, A., Van Densen, W., Machiels, M., Kranenbarg, J., Breteler, J., Backx, J., 2003.
Suitable habitats for 0-group sh in rehabilitated oodplains along the lower river
rhine. River Res. Appl. 19, 353374.
Guse, B., Pfannerstill, M., Geertz, J., Fohrer, N., 2014a. Einzugsgebietsmodellierung des
Einusses von dynamischen nderungen in der rumlichen Verteilung von
Ackerfrchten auf die Wasserqualitt in drei Landnutzungsszenarien. Forum Hydrol.
Wasserbewirtsch. 34 (14), 103110.
Guse, B., Reusser, D.E., Fohrer, N., 2014b. How to improve the representation of hydrological processes in SWAT for a lowland catchment temporal analysis of parameter
sensitivity and model performance. Hydrol. Process. 28, 26512670.
Hattermann, F., Krysanova, V., Wechsung, F., Wattenbach, M., 2004. Integrating groundwater dynamics in regional hydrological modelling. Environ. Model Softw. 19, 10391051.
Hauer, C., Unfer, G., Holzmann, H., Schmutz, S., Habersack, H., 2013. The impact of
discharge change on physical instream habitats and its response to river morphology.
Clim. Chang. 116, 827850.

Hering, D., Johnson, R.K., Kramm, S., Schmutz, S., Szoszkiewicz, K., Verdonschot, P.F.M.,
2006. Assessment of european streams with diatoms, macrophytes, macroinvertebrates and sh: a comparative metric-based analysis of organism response to stress.
Freshw. Biol. 51, 17571785.
Hesse, C., Krysanova, V., Pazolt, J., Hattermann, F., 2008. Eco-hydrological modelling in a
highly regulated lowland catchment to nd measures for improving water quality.
Ecol. Model. 218 (12), 135148.
Huang, S., Hesse, C., Krysanova, V., Hattermann, F., 2009. From meso- to macro-scale
dynamic water quality modelling for the assessment of land use change scenarios.
Ecol. Model. 220 (19), 25432558.
Huang, S., Krysanova, V., sterle, H., Hattermann, F.F., 2010. Simulation of spatiotemporal
dynamics of water uxes in Germany under climate change. Hydrol. Process. 24,
32893306.
Hughes, S.J., Ferreira, T., Cortes, R.V., 2008. Hierarchical spatial patterns and drivers of
change in bentic macroinvertebrate communities in an intermittent mediterranean
river. Aquat. Conserv. Mar. Freshwat. Ecosyst. 18, 742760.
Jasiewicz, J., 2011. A new grass gis fuzzy inference system for massive data analysis.
Comput. Geosci. 37, 15251531.
Juckem, P.F., Hunt, R.J., Anderson, M.P., Robertson, D.M., 2008. Effects of climate and land
management change on streamow in the driftless area of Wisconsin. J. Hydrol. 355,
123130.
Kail, J., Wolter, C., 2013. Pressures at larger spatial scales strongly inuence the ecological
status of heavily modied river water bodies in germany. Sci. Total Environ. 454455,
4050.
Kail, J., Guse, B., Radinger, J., Schrder, M., Kiesel, J., Kleinhans, M., Schuurman, F., Fohrer,
N., Hering, D., Wolter, C., 2015. Coupling models in a modelling framework to assess
the effect of pressures on abiotic habitat conditions and biota of rivers. PLoS One
(accepted).
Kiesel, J., Hering, D., Schmalz, B., Fohrer, N., 2009. A transdisciplinary approach for
modelling macroinvertebrate habitats in lowland streams. IAHS-AISH Publ. 328,
2433.
Kiesel, J., Fohrer, N., Schmalz, B., White, M.J., 2010. Incorporating landscape depressions
and tile drainages of lowland catchments into spatially distributed hydrologic modeling. Hydrol. Process. 24, 14721486.
Kiesel, J., Schmalz, B., Brown, G.L., Fohrer, N., 2013. Application of a hydrological-hydraulic
modelling cascade in lowlands for investigating water and sediment uxes in catchment, channel and reach. J. Hydrol. Hydromech. 61, 334346.
Kiesel, J., Schrder, M., Hering, D., Schmalz, B., Hrmann, G., Jhnig, S., Fohrer, N., 2015. A
new model linking macroinvertebrate assemblages to habitat composition in rivers:
development, sensitivity and univariate application. Fundam. Appl. Limnol. 186
(12), 117133.
Kundzewicz, Z.W., Mata, L.J., Arnell, N.W., Dll, P., Jimenez, B., Miller, K., Oki, T., Sen, Z.,
Shiklomanov, I., 2008. The implications of projected climate change for freshwater
resources and their management. Hydrol. Sci. J. 53 (1), 310.
Lamouroux, N., Capra, H., Pouilly, M., Souchon, Y., 1999. Fish habitat preferences in large
streams of southern France. Freshw. Biol. 42 (4), 673687.
Li, Z., Liu, W.-Z., Zhang, X.-C., Zheng, F.-L., 2009. Impacts of land use change and climate
variability on hydrology in an agricultural catchment on the loess plateau of China.
J. Hydrol. 377, 3542.
Li, Y., Chen, B.-M., Wang, Z.-G., Peng, S.-L., 2011. Effects of temperature change on water
discharge, and sediment and nutrient loading in the lower pearl river basin based
on SWAT modelling. Hydrol. Sci. J. 56 (1), 6883.
Logez, M., Bady, P., Melcher, A., Pont, D., 2013. A continental-scale analysis of sh assemblage functional structure in European rivers. Ecography 36, 8091.
LVERMA S-H, 1995. Digitales Hhenmodell DHM25 (25 m 25 m). Land Survey Ofce
Schleswig-Holstein, Kiel.
LVERMA S-H, 2004. Automatisierte Liegenschaftskarte (ALK). Vermessungs- und
Katasterverwaltung Schleswig-Holstein.
Mann, R., 1996. Environmental requirements of european non-salmonid sh in rivers.
Hydrobiologia 323, 223235.
Marquardt, J., 2008. GIS-gesttzte Untersuchung des Panzenanbaus fr Biogasanlagen
und dessen potentieller Auswirkungen auf die Beschaffenheit und den Zustand von
Boden und Wasser in Schleswig-Holstein. Diplomarbeit, Geographie, Leibniz
Universitt Hannover.
Martinkova, M., Hesse, C., Krysanova, V., Vetter, T., Hanel, M., 2011. Potential impact of
climate change on nitrate load from the Jizera catchment (Czech Republic). Phys.
Chem. Earth 36, 673683.
Marzin, A., Verdonschot, P., Pont, D., 2013. The relative inuence of catchment, riparian
corridor, and reach-scale anthropogenic pressures on sh and macroinvertebrate
assemblages in French rivers. Hydrobiologia 704, 375389.
Mastrorillo, S., Dauba, F., Belaud, A., 1996. Utilisation des microhabitats par le vairon, le
goujon et la loche franche dans trois rivieres du sud-ouest de la france. Ann.
Limnol.-Int. J. Limnol. 32 (3), 185195.
Mastrorillo, S., Lek, S., Dauba, F., 1997. Predicting the abundance of minnow Phoxinus
phoxinus (Cyprinidae) in the River Ariege (France) using articial neural networks.
Aquat. Living Resour. 10 (3), 169176.
McManamay, R., Frimpong, E., 2014. Hydrologic ltering of sh life history strategies across
the United States: implications for stream ow alteration. Ecol. Appl. 25, 243263.
Mooij, W., Trolle, D., Jeppesen, E., et al., 2010. Challenges and opportunities for integrating
lake ecosystem modelling approaches. Aquat. Ecol. 44, 633667.
Mouton, A., Schneider, M., Depestele, J., Goethals, P., Depauw, N., 2007. Fish habitat
modelling as a tool for river management. Ecol. Eng. 29, 305315.
Mller, M., Pander, J., Geist, J., 2014. The ecological value of stream restoration measures:
an evaluation on ecosystem and target species scales. Ecol. Eng. 62, 129139.
Nash, J.E., Sutcliffe, J.V., 1970. River ow forecasting through conceptual models: part I a
discussion of principles. J. Hydrol. 10, 282290.
555
Neitsch, S., Arnold, J., Kiniry, J., Williams, J., 2011. Soil & water assessment tool theoretical documentation version 2009. Texas Water Resources Institute Technical Report
406.
Nelson, J.M., Shimizu, Y., Takebayashi, H., McDonald, R.R., June, 2010. The international
river interface cooperative (iRIC): Public domain software for river modeling. In Proceedings of the 2nd Joint Federal Interagency Conference.
Niehoff, D., Fritsch, U., Bronstert, A., 2002. Land-use impacts on storm-runoff generation:
scenarios of land-use change and simulation of hydrological response in a meso-scale
catchment in SW-Germany. J. Hydrol. 267, 8093.
Orlowsky, B., Gerstengarbe, F.-W., Werner, P., 2008. A resampling scheme for regional
climate simulations and its performance compared to a dynamical RCM. Theor.
Appl. Climatol. 92, 209223.
Pander, J., Geist, J., 2013. Ecological indicators for stream restoration success. Ecol. Indic.
30, 106118.
Pardo, I., Armitage, P., 1997. Species assemblages as descriptors of mesohabitats.
Hydrobiologia 344 (1), 111128.
Payne, T., 2003. The concept of weighted usable area as relative suitability index. IFIM
Users Workshop, 15 June 2003, Fort Collins, pp. 114.
Pfannerstill, M., Guse, B., Fohrer, N., 2014a. A multi-storage groundwater concept for the
SWAT model to emphasize nonlinear groundwater dynamics in lowland catchments.
Hydrol. Process. 28, 55995612.
Pfannerstill, M., Guse, B., Fohrer, N., 2014b. Smart low ow signature metrics for an
improved overall performance evaluation of hydrological models. J. Hydrol. 510,
447458.
Poff, N., Allan, J., Bain, M., Karr, J., Prestegaard, K., Richter, B., Sparks, R., Stromberg, J., 1997.
The natural ow regime: a new paradigm for riverine conservation and restoration.
Bioscience 47, 769784.
Pottgiesser, T., Sommerhuser, M., 2008. Beschreibung und Bewertung der deutschen
Fliessgewssertypen, Steckbriefe und Anhang. Available:, http://www.wasserblick.
net/servlet/is/18727/?highlight=ie%DFgew%E4ssertypen.2008 (Accessed 1 September 2014).
Radinger, J., Wolter, C., 2015. Disentangling the effects of habitat suitability, dispersal and
fragmentation on the distribution of river shes. Ecol. Appl. 25 (4), 914927.
Radinger, J., Wolter, C., 2014. Patterns and predictors of sh dispersal. Fish and Fish.
15 (3), 456473.
Reid, M.A., Thoms, M.C., 2008. Surface ow types, near-bed hydraulics and the distribution of stream macroinvertebrates. Biogeosciences 5 (4), 10431055.
Richter, B., Baumgartner, J., Powell, J., Braun, D., 1996. A method for assessing hydrologic
alteration within ecosystems. Conserv. Biol. 10, 163174.
Roth, N.E., Allan, J.D., Erickson, D.L., 1996. Landscape inuences on stream biotic integrity
assessed at multiple spatial scales. Landsc. Ecol. 11, 141156.
Santos, R.M.B., Sanches Fernandes, L.F., Moura, J.P., Pereira, M.G., Pacheco, F.A.L., 2014. The
impact of climate change, human interference, scale and modeling uncertainties on
the estimation of aquifer properties and river ow components. J. Hydrol. 519,
12971314.
Schaarschmidt, T., Arzbach, H., Bock, R., Borkmann, I., Braemick, U., Brunke, M., Lemcke, R.,
Kaemmereit, M., Meyer, L., Tappenbeck, L., 2005. Die Fischfauna der kleinen
Fliessgewsser Nord- und Nordostdeutschlands - Leitbildentwicklung und
typgerechte Anpassung des Bewertungsschemas nach EU-Wasserrahmenrichtlinie.
Abschlussbericht. LAWA- Projekt im Rahmen des Lndernanzprogramms Wasser
und Boden im Auftrag des Umweltministeriums Mecklenburg-Vorpommernp. 330 S.
Scharf, J., Braemick, U., Fredrich, F., Rothe, U., Schuhr, H., Tautenhahn, M., Wolter, C., Zahn,
S., 2011. Fische in Brandenburg - Aktuelle Kartierung und Beschreibung der
mrkischen Fischfauna. Institut fr Binnenscherei e.V. Potsdam-Sacrow (Hrsg.).
Schinegger, R., Trautwein, C., Schmutz, S., 2013. Pressure-specic and multiple pressure response of sh assemblages in european running waters. Limnologica 43,
348361.
Schmedtje, U., Colling, M., 1996. kologische Typisierung der aquatischen Makrofauna.
Informationsberichte des Bayerischen Landesamtes fr Wasserwirtschaft, Mnchen.
Schmidt-Kloiber, A., Hering, D., 2011. The Taxa and Autecology Database For Freshwater
Organisms, Version 5.0. www.freshwaterecology.info (accessed on 26.01.2015).
Schrder, M., Kiesel, J., Schauttmann, A., Jhnig, S.C., Lorenz, A.W., Kramm, S., Keizer-Vlek,
H., Rolauffs, P., Graf, W., Leitner, P., Hering, D., 2013. Substratum associations of
benthic invertebrates in lowland and moutain streams. Ecol. Indic. 30, 178189.
Setegn, S.G., Rayner, D., Melesse, A.M., Dargahi, B., Srinivasan, R., 2011. Impact of climate
change on the hydroclimatology of Lake Tana Basin, Ethiopia. Water Resour. Res. 47.
http://dx.doi.org/10.1029/2010WR009248.
Singh, R., Wagener, T., van Werkhoven, K., Mann, M., Crane, R., 2011. A trading-space-fortime approach to probabilistic continuous streamow predictions in a changing
climate. Hydrol. Earth Syst. Sci. 15, 35913603.
Spratte, S., Hartmann, U., 1998. Fischartenkataster. Ssswassersche und Neunaugen in
Schleswig-Holstein. Ministerium fr lndliche Rume, Landwirtschaft, Ernhrung
und Tourismus des Landes Schleswig-Holstein, Kiel.
Stephenson, J.M., Morin, A., 2009. Covariation of stream community structure and
biomass of algae, invertebrates and sh with forest cover at multiple spatial scales.
Freshw. Biol. 54, 21392154.
Stubbington, R., Boulton, A.J., Sally, L., Wood, P.J., 2015. Changes in invertebrate assemblage composition in benthic and hyporheic zones during a severe supraseasonal
drought. Freshw. Sci. 34 (1), 344354.
Sundermann, A., Gerhardt, M., Kappes, H., Haase, P., 2013. Stressor prioritisation in
riverine ecosystems: which environmental factors shape benthic invertebrate assemblage metrics? Ecol. Indic. 27, 8396.
Tu, J., 2009. Combined impact of climate and land use changes on streamow and water
quality in eastern Massachusetts, USA. J. Hydrol. 379, 268283.
USACE, 2002. River Analysis System (HEC-RAS), User's Manual. Hydrologic Engineering
Center (HEC), US Army Corps of Engineers.
556
USACE, 2011. HEC-GeoRAS, GIS Tools for Support of HECRAS Using ArcGIS. Users Manual,
Version 4.3.93. US Army Corps of Engineers, Hydraulic Engineering Centre.
Valle Junior, R., Varandas, S., Pacheco, F., Pereira, V., Santos, C., Cortes, R., Fernandes,
L., 2015. Impacts of land use conicts on riverine ecosystems. Land Use Policy
43, 4862.
Vlach, P., Dusek, J., Svatora, M., Moravec, P., 2005. Fish assemblage structure, habitat and
microhabitat preference of ve sh species in a small stream. Folia Zool.-Praha 54 (4),
421.
Wolter, C., Arlinghaus, R., Grosch, U.A., Vilcinskas, A., 2003. Fische & Fischerei in Berlin. Z.
Fischkunde Suppl. 2, 1156.

Guse Et Al 2015 Eco-Hydrologic Model Cascades

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Guse Et Al 2015 Eco-Hydrologic Model Cascades

Uploaded by

Copyright:

Available Formats

Science of the Total Environment 533 (2015) 542556

Contents lists available at ScienceDirect

Science of the Total Environment

Eco-hydrologic model cascades: Simulating land use and climate change

A recently developed model cascade is applied in different scenario simulations.

B. Guse et al. / Science of the Total Environment 533 (2015) 542556

studies analysed the individual or combined effects of climate or land

Applying this modelling framework in scenario simulations enables

2. Material and methods

B. Guse et al. / Science of the Total Environment 533 (2015) 542556

Daily time series of discharge and nitrate

1-D hydraulic model

2-D hydrodynamic model

Fish habitat model

GRASS GIS r.fuzzy.system

Macroinvertebrate habitat model

Habitat evaluation tool

Daily time series of climate variables

Rating curve for cross-sections

2.2. Study areas and spatial scales

2.3. Model chain and model parametrisation

The Treene is a mesoscale sand-bed river in the northern German

2.3.1. Eco-hydrological model SWAT at the catchment scale

B. Guse et al. / Science of the Total Environment 533 (2015) 542556

Sollerup was used as the lower boundary condition. Since there is

B. Guse et al. / Science of the Total Environment 533 (2015) 542556

classes, to dene habitat suitability (Ahmadi-Nedushan et al., 2006;

In this study, the r.fuzzy.system was applied for three sh species:

water depth (preferring deep water) than on ow velocity, but tolerates

Water depth [m]

Flow velocity [m/s]

B. Guse et al. / Science of the Total Environment 533 (2015) 542556

2.4.1. Climate change scenario

2.4. Scenario description

Thresholds in nitrate concentration [mg/l]

B. Guse et al. / Science of the Total Environment 533 (2015) 542556

B. Guse et al. / Science of the Total Environment 533 (2015) 542556

predicted from July to October. The mean monthly values of ow

Water depth [m]

Flow velocity [m/s]

Tanypodinae Gen. sp.

Chironomini Gen. sp.

100 150 200 250 300 350

B. Guse et al. / Science of the Total Environment 533 (2015) 542556

Rel. Diff. to Baseline [%]

3.3.2. Hydrodynamic model simulations

Rel. Diff. to Baseline [%]

B. Guse et al. / Science of the Total Environment 533 (2015) 542556

was mostly increasing throughout the year and up to 10% change in

3.3.3. Fish habitat modelling

Rel. Diff. to Baseline [%]

B. Guse et al. / Science of the Total Environment 533 (2015) 542556

Tanypodinae Gen. sp.

Rel. Diff. to Baseline [%]

Chironomini Gen. sp.

same direction of change, i.e., either an increase or decrease in

In contrast, the land use scenarios had only minor effects on

4.2. Impacts of climate and land use change on ow velocity and

The higher temperature in the climate scenario led to an increase in

In general, the modelled hydraulic conditions reected the discharge

B. Guse et al. / Science of the Total Environment 533 (2015) 542556

production (Burgin and Hamilton, 2007; Camargo and Alonso, 2006).

B. Guse et al. / Science of the Total Environment 533 (2015) 542556