- Email: [email protected]
A model study on the large-scale effect of macrofauna on the suspended sediment concentration in a shallow shelf sea
Accepted Manuscript A model study on the large-scale effect of macrofauna on the suspended sediment concentration in a shallow shelf sea M.H. Nasermoaddeli, C. Lemmen, F. Kösters, G. Stigge, O. Kerimoglu, H. Burchard, K. Klingbeil, R. Hofmeister, M. Kreus, K.W. Wirtz PII:
S0272-7714(17)30098-7
DOI:
10.1016/j.ecss.2017.11.002
Reference:
YECSS 5659
To appear in:
Estuarine, Coastal and Shelf Science
Received Date: 27 January 2017 Revised Date:
23 October 2017
Accepted Date: 1 November 2017
Please cite this article as: Nasermoaddeli, M.H., Lemmen, C., Kösters, F., Stigge, G., Kerimoglu, O., Burchard, H., Klingbeil, K., Hofmeister, R., Kreus, M., Wirtz, K.W., A model study on the large-scale effect of macrofauna on the suspended sediment concentration in a shallow shelf sea, Estuarine, Coastal and Shelf Science (2017), doi: 10.1016/j.ecss.2017.11.002. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
A model study on the large-scale effect of macrofauna on the
2
suspended sediment concentration in a shallow shelf sea
3
Nasermoaddeli, M. H.a, Lemmen, C.b, Kösters, F.a, Stigge, G.c, Kerimoglu, O.b, Burchard, H.d, Klingbeil,
4
K.d,e, Hofmeister, R.b, Kreus, M. a, Wirtz, K. W.b
5
RI PT
1
a
Federal Waterways Engineering and Research Institute (BAW), Hamburg, Germany
7
b
Helmholtz-Zentrum Geesthacht Zentrum für Material- und Küstenforschung, Institute of Coastal
8
Research, Geesthacht, Germany
9
c
Institute for Applied Ecology (IfAÖ GmbH), Rostock, Germany
10
d
Leibniz Institute for Baltic Sea Research Warnemünde, Dept. for Physical Oceanography and
11
Instrumentation, Rostock, Germany
12
e
M AN U
Dept. of Mathematics, University of Hamburg, Germany
SC
6
13 Abstract
15
The activity of macrofauna on the sea floor is since long known to mediate deposition and erosion of
16
sediment, but so far most studies addressed this effect at a local scale. In the present paper, the
17
contribution of the observed macrofauna distribution (exemplified by a bivalve, the bean-like
18
Fabulina fabula, formerly known as Tellina fabula) on large-scale sediment transport in the southern
19
North Sea is investigated by means of a model study. Macrofauna effects are considered with respect
20
to the critical bed shear stress and erodibility, which are two important factors that control the
21
resuspension rate. Simulation results for a typical winter month revealed for the first time that the
22
suspended sediment concentration (SSC) is increased not only locally but beyond the inhabited
23
zones. This alteration is not confined to near-bed zones but can be observed throughout the entire
24
water column, especially during storm events. These effects are most prominent in the fine silt
25
fraction, coarser and finer fractions are less affected. For a selected storm event in February 2010,
26
we explain the counter-intuitive decrease in near-bed SSC in some areas with a high macrofauna
27
abundance compared to a simulation excluding such macrofauna: A high macrofauna-induced
28
entrainment rate leads to rapid exhaustion of available sediments at the bed in the model and
29
consequently limits the near-bed SSC.
AC C
EP
TE D
14
1
ACCEPTED MANUSCRIPT 1
Key words: macrofauna, suspended sediment, North Sea, Fabulina fabula, coupled modeling
2
1
3
Introduction Suspended sediment dynamics in coastal and shelf seas critically affects major biological
processes and has large economic implications (Puls et al., 1997, Winterwerp and van Kesteren 2004,
5
Gayer et al., 2006, van der Molen et al., 2009). The transport of suspended sediment contributes
6
significantly to material fluxes between the seabed and the water column and between rivers,
7
shallow water systems, and the open sea. It depends on physical factors such as tides and their
8
asymmetry (Burchard et al., 2008; Gräwe et al., 2016), wind and wave forcing (Lettmann et al., 2009),
9
and the morphological characteristics of the seafloor (van Ledden et al., 2004). Sediment transport is
SC
RI PT
4
also affected by biological factors such as microphytobenthos and macrofauna (Jumars and Nowell,
11
1984, Krumbein, 1994, Widdows et al., 1998, Orvain et al., 2006, Briggs et al., 2015, Harris et al.,
12
2016). Benthic organisms exert a considerable influence on sediment dynamics (a three- to sixfold
13
increase in the re-suspension rate as reported by Davis (1993)) since their activity is linked to a large
14
number of biological, geological, and physical factors. For example, macrobenthic animals graze on
15
stabilizing microphytobenthos (Orvain et al., 2014), alter the roughness of the seafloor (Peine et al.,
16
2005), or directly release particles into the water column (Graf and Rosenberg, 1997). In this study, we focus on two effects of macrobenthic communities on suspended sediments:
TE D
17
M AN U
10
changes in (1) the critical bed shear stress for erosion and (2) the erodibility of the bed sediment.
19
These two parameters, together with actual bottom shear stress, largely determine the erosion and
20
resuspension rates (Pleskachevsky et al., 2005 Stanev et al., 2009) and have been identified as
21
sensitive parameters in previous modeling studies (Knaapen et al., 2003, Lumborg et al., 2006, Le Hir
22
et al., 2007, van Prooijen et al., 2011, Orvain et al., 2012).
AC C
EP
18
23
Sediment erodibility of the few upper centimeters of the bed layer is modified (Andersen and
24
Pejrup, 2011), for example, by moving, sheltering and feeding macrofauna. These highly seasonal
25
processes lead to sediment displacement and are subsumed under the term bioturbation (Le Hir et
26
al., 2007). They also alter sediment aggregation (Nowell and Jumars, 1984), porosity and
27
permeability (Graf and Rosenberg, 1997), and the erodibility of the bed layer through
28
biostratification. For example, the vertical transport by upward conveyors (e.g. tube building
29
lugworm Arenicola marina) increases the fine fraction at the water-sediment interface. Conversely, 2
ACCEPTED MANUSCRIPT 1
downward conveyors (e.g. sipunculid worms) decrease the fine fraction (Wheatcroft and Butman,
2
1997).
3
Burrowing and grazing activities of the deposit feeder Fabulina fabula (F. fabula) - formerly known as Tellina fabula - for example, increase the erodibility of surface sediments (Austen et al.,
5
1999, Borsje et al., 2009). Sediment erodibility can be further altered by the aggregation of bulk
6
sediment into fecal pellets or pseudo-feces by deposit feeders. This bio-aggregation increases
7
erodibility since the aggregated material is less cohesive than before consumption (Austen et al.,
8
1999). Fecal pellets are also more prone to erosion due to the reduction in specific surface area, i.e.
9
the ratio of surface area to mass (Andersen and Pejrup, 2011).
SC
RI PT
4
The critical bed shear stress of the uppermost bed layer is also reduced by the generation of a
11
fluff layer, which may be formed by snail tracks, fecal pellets and disintegrated sediment particles
12
due to deposit feeding activities (Orvain et al., 2003). Particles from this layer are brought into
13
suspension far below the critical bed shear stress of the underlying sediment ("fluff layer erosion")
14
(Shimeta et al., 2002).
M AN U
10
Research on suspended sediment dynamics has started to include macrobenthic effects,
16
specifically on erodibility, in both observational and modeling studies. Most of these past studies,
17
however, focused on local effects of benthic fauna (benthos) on sediment transport or were
18
restricted to smaller, e.g. intertidal, scales (Paarlberg et al., 2005; Borsje et al., 2007; Le Hir et al.,
19
2007; Sanford, 2008). To date, there have been very few studies of the large-scale effects of benthic
20
fauna on suspended sediment dynamics (e. g. in Seifert et al. (2009) for the Baltic Sea).
EP
21
TE D
15
In this paper, we quantify local effects of macrobenthos on sediment transport and study how these impact the large-scale distribution of suspended sediments by means of a model study. In
23
particular, the present study aims to yield model based sensitivities in view of the following research
24
questions: (1) To what extent and magnitude do macrobenthic effects on critical bed shear stress and
25
erodibility influence the large-scale near-bed suspended sediment distribution, i.e. sediment
26
concentration patterns, in a shallow shelf-sea (southern North Sea)? (2) Are the effects similar for
27
different sediment classes? (3) Do effects differ for sediment transport near the bed and in the whole
28
water column? (4) Are the impacts of macrofauna limited to events or characteristic for longer time
29
periods?
AC C
22
3
ACCEPTED MANUSCRIPT 1
We address these questions based on a coupled model system taking into account 3D hydrodynamics (tides, baroclinic flow), a simplified wave model and multi-fractioned sediments thus
3
resolving the most relevant processes for coastal and shelf sediment transport. In addition this
4
coupled system includes effects of macrofauna by using the observed spatial distribution of the
5
bivalve F. fabula, which is a dominant species and eponymous characteristic of a macrozoobenthos
6
community of the southern North Sea. We then compare the simulated surface suspended sediment
7
concentration (SSC) with satellite observations and in-situ measurements and analyze the simulation
8
results for a reference scenario including mentioned processes and another scenario that disregards
9
the macrobenthic effects.
11
2.1
12
SC
2
Materials and methods
M AN U
10
RI PT
2
Study Area
The study area comprises the southern North Sea, a shallow shelf system with water depths of up to 50 m, intertidal flats and a number of estuaries at the coastline (Figure 1). Currents are
14
predominantly influenced by the semi-diurnal tides that propagate through the German Bight in an
15
anti-clockwise direction, with amplifications in the estuaries and complex bathymetric interactions
16
(e.g. Becker et al., 1992; Stanev et al., 2014). Substantial quarter-diurnal tidal components exhibiting
17
a significant seasonal signal (Gräwe et al., 2016) are generated (Stanev et al., 2014). The tidal range
18
varies from 2 m near the Dutch coast to about 4 m in the German estuaries. Tidal mixing inhibits
19
thermal and haline stratification in the shallower parts of the German Bight (Schrum, 1997). Coastal
20
upwelling and downwelling due to wind forcing affect the pathways of the discharge from the Elbe,
21
Weser and Ems rivers (Krause et al., 1986). Stratification only occurs seasonally in the deeper parts of
22
the southern North Sea and occasionally near the estuaries (van Leeuwen et al., 2013). At the fronts
23
between the generally vertically well mixed German Bight waters and the stratified North Sea waters,
24
a so-called line of no return (Postma, 1984) defines an area within which sediments are trapped in
25
the German Bight. Between this line of no return and the coast, a thermohaline overturning
26
circulation transports sediments towards the coast such that sediments accumulate in the Wadden
27
Sea (Burchard et al., 2008) and contribute to the formation of estuarine turbidity maxima in the
28
estuaries (e. g. Kappenberg and Grabemann, 2001).
AC C
EP
TE D
13
4
SC
RI PT
ACCEPTED MANUSCRIPT
2 3
M AN U
1
Figure 1: Bathymetry of the southern North Sea as represented on the numerical model grid. The study area of the German Bight is depicted by the black rectangle.
4
Macrobenthic infauna of the German EEZ has classically been categorized as belonging to one of five main benthic infauna communities based on preferred habitat conditions (depth, sediment
6
type), namely, Amphiura filifornis, Tellina (Fabulina) fabula, Nucula nitidosa, Goniadella-Spisula, Spio
7
filicornis (Salzwedel et al., 1985). The categorization has been roughly confirmed by the ICES North
8
Sea Benthos Survey (Duineveld et al., 1991) and the North Sea Benthos Project 2000 (Kröncke et al.,
9
2011) using a grid of sampling points, and by more refined cluster analysis (Neumann et al., 2012).
EP
TE D
5
10
2.2
11
2.2.1 Sedimentological data
AC C
12
Observational data
Sedimentological input data is based on measurements of the Federal Maritime and
13
Hydrographic Agency (BSH). These and other data sources have been compiled and are available as a
14
habitat atlas published by the ‘North Sea Observation and Assessment of Habitats (NOAH)’ project
15
(http://www.noah-project.de).
16 17
Near-surface in-situ SSC time series in the area of West Frisia provided by the Dutch authority Rijkswaterstaat was accessed through OpenEarth (Rijkswaterstaat, 2017). The two cross-shore 5
ACCEPTED MANUSCRIPT 1
transects, Terschelling and Rottumerplate (Figure 3), were selected because of their close proximity
2
to the East Anglian plume and data availability for the period of study.
3
Suspended sediment concentrations close to the water surface can be inferred from Envisat satellite images provided by the European Space Agency (ESA) at a 300 m resolution. The satellite
5
images were pre- and post-processed by ESA and Brockmann Consult using MERIS regional case 2
6
water algorithms (C2R) explained in Doerffer et al. (2006). It should be noted that the calculated
7
concentrations represent the total suspended matter in the near-surface layers, which also includes
8
organic matter not considered in the simulated SSC. The penetration depth of the light emitted from
9
the satellite-borne MERIS sensor depends on the turbidity of the water, for example at a depth of 1 m in the case of high turbidity in our area of interest.
11
2.2.2 Macrofauna data
We accessed the macrozoobenthos database of the Institute for Applied Ecology (IfAÖ GmbH),
M AN U
12
SC
10
RI PT
4
which included 17,710 hauls at 2,347 infauna sampling stations and 2,798 hauls at 1,049 epifauna
14
sampling stations in the German Bight by the time of the data request. Based on an analysis of
15
species abundances, biomasses and presences derived from this database and by means of a
16
literature review, we selected species that are particularly important for sediment transport
17
processes. Due to its bioturbation potential the bivalve F. fabula is one of the most distinctive species
18
in the study area (Borsje et al., 2009), which is also reflected by the naming of the F. fabula
19
community occurring in the German Bight (Rachor and Nehmer, 2003). The distribution of F. fabula
20
in the southern North Sea is based on the data by Creutzberg (1986) as well as the occurrence
21
probability by Reiss et al. (2011), converted to abundance by field data compiled by van Moorsel
22
(2011) and supplemented for the German Bight by data of Dannheim (2014) as shown in Figure 2.
23
The comparison of sedimentological data, in this case the median grain size d50 in mm, with the
24
abundance of F. fabula illustrates that this species not only appears in large numbers in sandy
25
sediments between 0.125 mm and 0.250 mm but also in regions with coarser or finer median grain
26
sizes.
AC C
EP
TE D
13
6
1 2
-
Figure 2: Estimated mean abundance of the bivalve F. fabula in the southern North Sea (color scale, individuals m ²); contours denote median grain size (mm).
4
2.3
Numerical models
TE D
3
5
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
To investigate large-scale macrofauna effects on the suspended sediment transport in the southern North Sea, a generic benthos module (Nasermoaddeli et al., 2014) was coupled to
7
hydrodynamics, sediment transport, and erosion—sedimentation models via the Modular System for
8
Shelves and Coasts (MOSSCO, Lemmen et al., 2017) coupling framework. The MOSSCO framework is
9
built on and extends the Earth System Modeling Framework (ESMF, Hill et al., 2004) and enables
AC C
10
EP
6
flexible and relatively seamless coupling of a data and model components.
11
In our application, MOSSCO includes five model components (hydrodynamics, suspended
12
sediment, erosion—sedimentation, waves and benthos effect) and generic input components that
13
provide data on ocean boundary particle concentration, river loads and macrofauna distribution.
14
2.3.1 Hydrodynamic model
7
ACCEPTED MANUSCRIPT 1
Hydrodynamics are simulated by the General Estuarine Transport Model (GETM; Burchard and Bolding, 2002). GETM is an open source coastal ocean model which has been shown to be of high
3
skill in various studies for the North Sea and Wadden Sea (Stips et al., 2004; Burchard et al., 2008;
4
Lettmann et al., 2009; van Leeuwen et al., 2013; Gräwe et al., 2016). It solves the incompressible
5
Navier-Stokes-Equations and additional prognostic equations for temperature and salinity (Klingbeil
6
and Burchard, 2013). The free surface is calculated efficiently in an explicit mode-splitting algorithm
7
and supports drying-and-flooding of intertidal flats. Quantities are transported with a high-order
8
advection scheme (e. g. Klingbeil et al., 2014). Horizontal sub-grid scale dynamics are parameterized
9
by a Smagorinsky closure (Smagorinsky, 1963). Turbulent vertical viscosities and diffusivities are
RI PT
2
provided by the General Ocean Turbulence Model (GOTM) which offers state-of-the-art turbulence
11
closure from Umlauf and Burchard (2005). For the present study, a dynamic k-ε model with the 2nd-
12
order closure of Canuto et al. (2001) “Model A” was applied. Bottom stresses are calculated on the
13
basis of the law of the wall with a spatially uniform roughness length of z0=1 mm, which was chosen
14
as a typical roughness value for the southern North Sea (e.g. Gräwe et al., 2016). Note that
15
roughness is a calibration factor for the model. The roughness length z0 chosen here relates to a
16
reasonable ripple height of about 30 mm. Spatial variations of the grain related roughness can be
17
neglected as typical sediment grain size diameters for the German Bight are in the order of 0.3 – 0.5
18
mm (medium sand), thus accounting for only 1% of total roughness. Variations in form roughness
19
have been also omitted in order to facilitate the interpretation of the results.
TE D
M AN U
SC
10
In order to include the effect of wind waves, local wave conditions (significant wave height, peak
21
period, wave direction) were parameterized in terms of wind speed, wind direction and water depth
22
according to Breugem and Holthuijsen (2007). These current-only bottom stresses are modified
23
according to Soulsby (1997) to consider the combined wave—current stresses.
24
2.3.2 Sediment transport model
AC C
25
EP
20
Suspended sediment is represented here by three size classes. Sediment mass is transported
26
along the current vector computed by the hydrodynamic model, while diffusion due to turbulence
27
and downward movement with a constant, size-dependent sinking velocity are applied additionally in
28
the vertical. For the transport of the SSC, the same high-order advection scheme is used as for the
29
transport of salinity and temperature in the hydrodynamic model. At open sea and river boundaries,
30
a constant sediment concentration is prescribed as proposed by Gayer et al. (2006). At the water-
8
ACCEPTED MANUSCRIPT 1
sediment interface, the net mass flux for each suspended sediment size class is prescribed as the flux
2
boundary condition for the sediment transport model.
3
The modularity within MOSSCO allows the coupling of a separate model for the calculation of the net sediment flux at the water—sediment interface. For this purpose, the erosion—sedimentation
5
routines of the Deltares Delft3d model were encapsulated and coupled via an ESMF interface to
6
MOSSCO (Nasermoaddeli et al., 2014). The model uses the well-known Partheniades-Krone equation
7
(Partheniades, 1965) extended for a sand-mud mixture to calculate the net sediment flux of cohesive
8
sediment at the water-sediment interface (see Table 1). A similar approach was proposed by van
9
Ledden (2004) or Le Hir et al. (2011) and was applied in Paarlberg et al. (2005).
= ,
12
,
,
∙
∙
∙
= =
,
,
,
,
∙ 1−
,
−1 ,
>
,
<
0,
,
,
≤
Eq. 1 Eq. 2 Eq. 3
,
,
Eq. 4
0,
≥
,
EP
,
11
SC
= ∙
M AN U
Table 1: Erosion and deposition fluxes as implemented in the numerical model according to Partheniades (1965)
TE D
10
RI PT
4
The main parameters taken into account are the erosion flux E (kg m-2 s-1) and the deposition flux D (kg m-2 s-1). In the equations, l is the index of sediment class, g the biological destabilization factor
14
for erodibility (-), M the erodibility factor (kg m-2 s-1), S the Heavy-Side function, ws the settling
15
velocity for each sediment class l, cb the near bed suspended sediment concentration, τ the bed
16
shear stress, τcr,er the critical bed shear stress for erosion, f the biological destabilization factor for
17
critical bed shear stress for erosion, and τcr,depr the critical bed shear stress for deposition. The factors
18
f and g are introduced into the original formula here to account for the macrofauna effects
19
presented in the following section.
AC C
13
9
ACCEPTED MANUSCRIPT The net vertical sediment flux for each class results from the difference between erosion and
2
deposition, both of which are applied to the active single layer of the bed. In this approach, the bed is
3
conceptualized as a well-mixed, non-stratified layer with a given depth. An arbitrary number of
4
sediment size classes can be considered within this highly idealized sediment inventory. The mass
5
fraction of each sediment size class within the well-mixed layer is the ratio of its mass to the total
6
mass of all fractions within the layer. The mass of each sediment size class for a unit area is initialized
7
in the model by a given fraction, thickness, porosity and sediment dry density which is spatially
8
constant. The deposited mass of each sediment size class is added to the rest of the mass of the
9
same class within the layer and mixed instantaneously with the bed material throughout the whole
RI PT
1
layer thickness. The mass fraction of the size class is altered accordingly for the whole depth. The
11
erosion of bed material for each sediment size class is calculated in a similar manner.
12
2.3.3 Benthos model
M AN U
13
SC
10
The inclusion of biotic effects on sediment transport modeling is complicated partly because of the lack of data and knowledge so that it is currently not possible to quantify the effects of
15
macrofauna on sediment transport, especially at the community level. Further complicacy stems
16
from the seasonal variation of biological components, spatial patchiness, and non-linear interactions
17
within the benthic community (Le Hir et al., 2007). In addition, effects on sediment transport
18
properties are highly specific and are thus dependent on the function of the considered benthic
19
species, while collective effects of different species on sediments are difficult to predict due to
20
reciprocal (antagonistic) influences in benthic communities (Kristensen et al., 2013). Currently available models of sediment transport only depict in a very simplified manner the high
EP
21
TE D
14
complexity of effects that macrozoobenthos in their entirety exert on the water-sediment boundary
23
layer. Another issue is the lack of species- or community-specific parameterizations of macrofauna
24
effects on sediment transport parameters. Most sediment transport models that resolve
25
macrozoobenthos have so far only been parameterized for few selected macrofauna species and
26
with few affected sediment transport parameters (e.g. Wood and Widdows, 2002; Knaapen et al.,
27
2003; Paarlberg et al., 2005) or they quantify such effects using process-based differential equations
28
(e.g. François et al., 1997; Orvain et al., 2003; Orvain, 2005; Montserrat Trotsenburg, 2011; Orvain et
29
al., 2012).
AC C
22
10
ACCEPTED MANUSCRIPT Due to the lack of suitable parameterizations for F. fabula, we applied parameterizations which
2
were originally derived for the effect of Macoma balthica on critical bed shear stress and erodibility
3
(Knaapen et al., 2003; Paarlberg et al., 2005). This approach was justified on the basis of a functional
4
classification of a total of 153 macrozoobenthos species from the German Bight according to a trait
5
matrix. Each species was numerically characterized according to seven traits (motility, nutritional
6
habits, habitat structure, sediment transport, size, shape and position in the sediment) as proposed
7
by Darr et al. (2014). Based on the assigned traits, a similarity analysis using a fuzzy coding approach
8
(Chevenet et al., 1994) allowed an aggregated description of macrofauna effects on sediment
9
transport at the community level by clustering species with similar autecology. F. fabula was found to
10
be the species most similar to Macoma balthica, following bivalves Abra alba and Astarte montagui.
11
Our analysis of macrobenthic communities based on the trait matrix by Darr et al. (2014) permitted a
12
generalization of a species specific parametrization. Underlying (laboratory) studies are usually
13
performed using single model species and now can be extrapolated to other dominant species in the
14
region of interest. Given the high number of macrobenthic species in coastal and shelf habitats, we
15
advocate the further development of trait-based approaches, also to permit a reduction in model
16
complexity and parameterization effort.
M AN U
SC
RI PT
1
These two formulations are used as linear modification factors of the corresponding abiotic
18
parameters. It should be noted that both relations rely only on limited data collected by Widdows et
19
al. (2000a, b) in an intertidal basin. The destabilization factors for critical bed shear stress and
20
erodibility are given in Table 2.
Table 2: Factors accounting for destabilization of critical bed shear stress (fd) and erodibility (gd) # $M&
22
&=
+&
− 0.085 ∙ ln$ & + 1
/+ ∙ 0 $/+ + 0 ∙ /12 & 3
AC C
#$
= 0.0016 ∙ ln$
EP
21
TE D
17
Eq. 5 Eq. 6
23
fd is the destabilizing factor for critical bed shear stress, M is the dimensionless abundance
24
(actual abundance of individuals (ind.) divided by the reference abundance of 1 ind. m-2), and gd the
25
destabilizing factor for erodibility. The derivation of gd is based on the assumption that the
26
macrofauna effect reaches a maximum with an increasing abundance M, after which it remains 11
ACCEPTED MANUSCRIPT 1
constant (sigmoid function). The parameters applied here are the maximum biological erosion
2
coefficient γ = 6 x 10-7 ms-1, the erosion coefficient without macrofauna influence I = 4.68 × 10-8 ms-1
3
and the regression coefficients b1 = 0.995 and b2 = 5.08 10-8 ms-1.
4
The aforementioned parameterizations were implemented in MOSSCO’s benthos module, which is a generic object-oriented library. It provides an adaptive structure for linear superposition of
6
macrofauna effects on sediment transport parameters such as critical shear stress, erodibility,
7
roughness, settling velocity, biodeposition and bio-resuspension. Currently, only the first two
8
parameterizations are included in the benthos model. In the following, this parameterization is
9
referred to as the macrofauna effect but the reader should bear in mind that this is based on a rather
RI PT
5
simplistic approach which can only give an indication of the complex interaction between
11
macrofauna, or biology in general, and sediment dynamics.
12
2.4
M AN U
13
Model setup
SC
10
The model was set-up in a way to represent the dynamics of fine sediments in the German Bight.
14
We focus on the topmost layer of easily erodible material, i.e. a layer of stationary suspension (e.g.
15
Hayter and Mehta, 1986). Thus large-scale sand transport and associated morphological changes are
16
not considered here.
The model resolves three silt classes, which differ in settling velocity and transport parameters,
TE D
17 18
similar to Gayer et al. (2006) and in accordance with the measurements of Puls et al. (1995) in the
19
German Bight. Other sediment transport parameters for each silt class are presented in Table 3. The initial fraction of each silt class at the bed and in the water column was assumed to be
21
spatially constant in the computational domain. The silt classes were determined by multiplying silt
22
fractions measured in the shallow water (0-20 m) zone by the ratio of the area of the shallow zone to
23
the total area of the southern North Sea. The silt fractions at open sea boundaries and rivers were
24
selected according to Gayer et al. (2006).
AC C
EP
20
25
The amount of mobile sediments in nature is not only spatially variable but also depends on the
26
vertical structure. In spite of intensive efforts to map surface sediments and their properties for the
27
German Bight (e.g. Arfai et al., 2014) information for the top few centimeters is not available yet.
28
Therefore the initial thickness of the active layer had to be selected and was set at 10 mm. This is in
29
line with comparable studies such as Lumborg et al. (2006) who used 2-5 mm obtained from field 12
ACCEPTED MANUSCRIPT 1
observations for the upper bed layer in Kongsmark, a tidal mudflat in the Danish Wadden Sea.
2
Limiting the erodible layer to a fixed value of 10 mm reflects, in a simplified manner, the limitation of
3
erosion in nature due, for example, to the consolidation or presence of coarse sand in the soil matrix
4
which may prevent the removal of fines by armoring. Critical bed shear stress for erosion (τc,er ) can generally be regarded as a calibration parameter; it
6
was adjusted in order to obtain a comparable structure of simulated SSC with available satellite data
7
and in-situ measurements. We found the calibration values of Gayer et al. (2006) to be a good
8
starting point but adjusted the critical bed shear stress for fine silt to 0.6 N m-2.
The existence of a critical bed shear stress for deposition (τc,dep ) is under debate. Winterwerp
SC
9
RI PT
5
(2007) showed the irrelevance of τc,dep by successfully reproducing experimental results without
11
including this parameter in his model. In contrast, Maa et al. (2008) argue, on the basis of their
12
laboratory experiments, that deposition ceases beyond a specific bed shear stress. We conducted a
13
sensitivity analysis by testing a range of values of τc,dep for each silt class varying between 0.096 N m-2
14
to 2 N m-2, as well as ignoring this parameter. The selected range reflects literature values (i.e. Gayer
15
et al., 2006, and Lumborg et al., 2006). The best fit between simulated and measured SSC was
16
reached by the values of τc,dep given in Table 3. The erodibility, which was constant for the three silt
17
fractions, was also calibrated.
TE D
Table 3: Sediment parameters of the numerical model (SNS setup)
Settling velocity
Initial
Initial
Fraction*
concentration
(m s )
(%)
(g m )
(m)
0.001
8.24
2.5
0.0001
5.46
0.00002
1.21
-1
Coarse silt
Initial
EP
Silt class
AC C
18
M AN U
10
τc,dep -
Erodibility -
-2 -1
(N m ²)
(N m ²)
(kg m s )
0.01
0.78
-
1.0×10-5
2.5
0.01
0.6
1.5
1.0×10-5
0.5
0.01
0.1
1.5
1.0×10-5
-3
thickness
τc,er
(0.031 -0.062 mm) Fine silt
(0.008-0.016 mm) Very fine silt (0.004 – 0.008
13
ACCEPTED MANUSCRIPT mm)
1 2
*fraction in respect of the total mass including all other sediment size classes (sediment classes according to Wentworth, 1922) The model domain covers large parts of the southern North Sea (SNS). We use a curvilinear
4
horizontal grid with 140x100 grid points in the horizontal to align the main axes with the coastline
5
and to increase the resolution in the German Bight up to 1.5 km at the south-east corner and 4.5 km
6
at the north-west corner (see Figure 1 for the model grid). In the vertical, 20 sigma layers are used,
7
resulting in a vertical resolution of 0.5 m at a water depth of 10 m and 2.0 m at a water depth of
8
40 m. This strategy has been shown to work successfully in hindcast simulations with GETM (e. g.
9
Hofmeister et al., 2013).
The numerical time step, as well as a coupling time step between the components, was set at 2
M AN U
10
SC
RI PT
3
minutes. The setup is forced supplied with output data from a regional atmospheric model at the sea
12
surface and the sea surface height at the open boundaries is taken from a barotropic hindcast
13
simulation. A more detailed description of the setup and its validation is given in Kerimoglu et al.
14
(2017). The effect of bottom friction is implemented as temporally and horizontally constant. The
15
bathymetry of the model domain and model grid is shown in Figure 1. Ten major rivers along the
16
coasts on the southern North Sea were considered with respect to their freshwater fluxes and the
17
suspended sediment loads (Radach and Pätsch (2007), Gayer et al. (2006)). We prescribe the
18
suspended sediment concentration at the northern and western open boundaries of the model
19
domain using estimates of total SSC by Heath et al. (2002).
EP
TE D
11
20
The simulation period from January to December 2010 covered both winter as well as summer months (for qualitative assessments) but also matches the availability of validation data for
22
hydrodynamic and partly for suspended sediment. For the fully coupled simulation of sediment
23
transport including macrobenthic effects, the observed distribution of the bivalve F. fabula was
24
implemented as external forcing and implemented in a modular manner. The configuration used for
25
this study can be reproduced by downloading the open source MOSSCO setups repository from
26
https://sf.net/p/mossco/setups , selecting the „sns“ example setup with the „spm“ configuration.
27
3
AC C
21
Results 14
ACCEPTED MANUSCRIPT 1
3.1
Model validation Observed water-levels are reproduced sufficiently well, e.g. the difference in tidal range, as a
3
measure of energy input by the tidal wave, between observed and modelled values is in the order of
4
5 % in the focus domain. Further details are given in the Appendix. The hydrodynamic model also
5
provides a realistic description of horizontal mixing as inferred from domain-scale salinity gradients
6
and vertical stratification from May to September within the deeper (> ~30 m) regions of the German
7
Bight (Kerimoglu et al., 2017).
8 9
RI PT
2
The simulation of waves is less satisfying as observed values are overestimated by the parametric wave model. However, for the scope of this study our rather simple approach provides a sufficient estimate for temporary wave induced excess bed shear stress in addition to the periodic tide induced
11
bed shear stress. Further details are given also in the Appendix.
For the SSC validation, two satellite images recorded on March 1, and April 14, 2010 were
M AN U
12
SC
10
13
selected (Figure 3 a, c). The former is representative of a storm event and the latter of a typical neap
14
tide. Both images display the typically observed coastal gradient (strongly decreasing offshore
15
concentration) as well as the East Anglian plume.
16
Their comparison with simulated near-surface SSC shows a good agreement in terms of reproducing the coastal gradients in the eastern part of the German Bight; but SSC is overestimated
18
at the Elbe and Weser estuaries and underestimated at the Ems estuary, as estuarine dynamics are -
19
due to the relatively coarse model resolution of 0.5 to 1 nautical mile - only resolved as climatological
20
boundaries. It was possible to reproduce the East Anglian plume on March 1 fairly well (Figure 3 a, b).
21
It is, however, missing in the simulation results on April 14 (Figure 3 c, d). It should be noted that the
22
satellite images show the total suspended matter while the simulations only show the suspended
23
sediment concentration.
EP
AC C
24
TE D
17
15
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
1
Figure 3: Snapshots of total suspended matter (TSM) calculated from satellite images on March 1, 2010 (a) and April
TE D
2 3
14, 2010 (c) vs simulated suspended sediment concentration (SSC) near-surface at the same days (b and d)
The sediment transport model was further validated using the time series from satellite images
5
as well as in-situ near-surface SSC at several locations nearshore and offshore of West Frisia. Near-
6
surface simulated SSC with and without macrofauna compare relatively well to satellite derived data
7
(Figure 4). At least the order of magnitude of observations and the simulated near-surface SSC agrees
8
in near shore and offshore regions of West and North Frisia. It was possible to reproduce the
9
temporal variation in SSC well in some periods and locations (for example at Terschelling T50 and
AC C
EP
4
10
Rottumerplate R70, Figure 4 b, c), but SSC was overestimated at 10 km off Terschelling (T10, Figure
11
4 a). The simulated SSC scenario without macrofauna performed better at T10, which could be due to
12
the overestimation of F. fabula abundance by the interpolation at this location.
16
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
1
8 9
simulations (green lines) of concentration of suspended material for coastal transects represented by stations T10 and T50 (10 and 50 km off the West Frisian island of Terschelling), R3 and R70 (3 and 70 km off the island of Rottumerplate outside the Dollart estuary, and S9 and S62 (9 and 62 km off the North Frisian island of Sylt); The simulation was sampled in a 2500 m radius around the station and the satellite (with higher resolution than the simulation) in a 500 m radius.
3.2
TE D
7
Figure 4: Comparison between in-situ measurements (blue diamond), remote sensing observations (red crosses) and
Simulations with macrofauna effects
In the following, the results are presented for a storm event as well as the temporal mean effect of macrofauna on SSC. Simulation results for SSC including macrofauna effects (i.e. using the
EP
2 3 4 5 6
observed spatial distribution of an average abundance of F. fabula) reveal a number of differences
11
between the scenarios with and without macrofauna.
12
3.2.1 Storm event
13
AC C
10
Simulated near-bed SSC with macrofauna differs from the SSC in the scenario without
14
macrofauna, depending on the sediment class, macrofauna abundance, and location. The near-bed
15
SSC of the coarse silt class during a storm event (Figure 5 a) is distributed over a large area in the
16
shallow zone and partly offshore. This patchiness seems to be the imprint of F. fabula inhabitation
17
zones when compared to the contour lines of F. fabula abundance in Figure 5 b. Additionally,
18
macrofauna in the model has led not only to an increase in the near-bed SSC in some inhabited areas 17
ACCEPTED MANUSCRIPT but also to a decrease in other areas. Zones with reduced near-bed SSC correspond mostly to
2
inhabited zones with a higher abundance of F. fabula (shown by thicker contour lines). While few
3
individuals may increase the erosion rate many individuals may induce biodeposition, as found by
4
Friedrichs et al. (2009). Increases in near-bed SSC due to macrofauna effects were more frequent
5
than reductions but the absolute magnitude of 95 percentiles ( 1.65 g m-3) was less than the absolute
6
magnitude of (p5 =-2.56 g m-3) as shown in Figure 5 b, reflecting the macrofauna-induced increase
7
and decrease in the near-bed SSC.
8 9
RI PT
1
A much lower degree of patchiness is observed in the near-bed SSC distribution of the fine silt class (Figure 5 c). Figure 5 d shows a significant macrofauna-related increase in SSC, primarily off the coast of Sylt and also in some parts of the shallow waters along East and West Frisia. These zones
11
correspond to the inhabited regions with F. fabula (shown by contour lines). A closer comparison of
12
macrofauna-enhanced SSC zones with F. fabula abundance contours in Figure 5 d shows that near-
13
bed SSC has been modified in the model results beyond the zones inhabited by macrofauna, altering
14
sediment erosion and deposition properties. This is particularly obvious in the shallow zones such as
15
north of Sylt or offshore of the Ems estuary, which suggests that the contribution of macrofauna to
16
sediment transport may affect large-scale patterns of near-bed SSC of fine silt.
M AN U
17
SC
10
The near-bed SSC of fine silt is also reduced by macrofauna in a number of regions (Figure 5 d). The pattern of near-bed SSC of fine and coarse silt is similar, with the former exhibiting greater
19
magnitudes of change and changes over larger areas. The near-bed SSC of fine silt has been
20
significantly reduced offshore of West Frisia and south-east of Helgoland. Similar to coarse silt
21
macrofauna-induced increase of SSC can be observed to the west and north of Sylt as well as
22
offshore of East Frisia. Seaward of West Frisia, however, macrofauna activities have resulted in a
23
decrease in the near-bed SSC of fine silt and coarse silt. Finally, a significant increase in differences
24
due to macrofauna effects (p95 1.65-7.86 g m-3) can be inferred from Figure 5 b and d.
EP
AC C
25
TE D
18
No patchiness can be detected from near-bed SSC distribution for very fine silt in Figure 5 e. The
26
modified near-bed SSC due to macrofauna for very fine silt, depicted in Figure 5 f, shows a very slight
27
difference (almost negligible) and mostly a reduced concentration in coastal shallow areas,
28
irrespective of the F. fabula distribution. It is inferred that the macrofauna effect does not play any
29
significant role for near-bed SSC for very fine silt during a storm event, in contrast to the other two
30
silt classes. 18
ACCEPTED MANUSCRIPT 1
The concentration of near-bed total silt (sum of all three silt classes) shown in Figure 5 g represents similar gradients and magnitude to those of the fine silt fraction. Patchiness of the near-
3
bed SSC offshore along the coasts and the SSC plume west and south west of Helgoland can still be
4
observed. The contribution of macrofauna to the near-bed SSC (Figure 5 h) is quite similar for near-
5
bed total silt and fine silt class. This implies that the predominant macrofauna effects on the near-
6
bed SSC may be represented by fine silt fractions in the case of a storm event. However, it should be
7
noted that the magnitude of enhancement of near-bed SSC due to macrofauna decreases when
8
integrating the concentration over all silt fractions since in some regions the concentration increased
9
for fine silt, but decreased for coarse silt, such as to the west of Sylt.
AC C
EP
TE D
M AN U
SC
RI PT
2
19
1 2 3 4 5
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Figure 5: Simulated near-bed suspended sediment concentration (SSC) during a storm event in February 2010. The left-hand panels show the near-bed SSC for coarse (a), fine (b), very fine (e) and total (g) SSC; the right-hand panels (b, d, f, h) show the calculated modification due to benthos effects. Contours denote the presence of F. fabula at 10 (thin line), -2
100, and 500 (thick line) individuals m . Summary statistics for n grid cells shown as mean µ and 5-95 percentiles p5,95
20
ACCEPTED MANUSCRIPT 1 2
3.2.2 Storm event, depth-averaged SSC Depth-averaged SSC is especially important for both ecosystem modeling and sediment management and was therefore subject to further investigations. As results for all three silt classes
4
were quite similar to those for near-bed SSC, we present only the total depth-averaged SSC below.
5
Model results of depth averaged SSC shown in Figure 6 are temporal means over a spring-neap-cycle
6
after one month of model spin-up time.
RI PT
3
7
The total depth-averaged SSC in Figure 6 exhibits similar patterns to the total near-bed SSC,
8
except for a weaker patchiness, which is a consequence of the vertical and horizontal mixing of SSC.
9
The main difference seems to be the plume around Helgoland, which lacks in total depth-averaged SSC and thus appears to be specific to the near-bed layer. The plume may hence be due to local
11
resuspension of silt near the bed, usually deposited south-east of Helgoland (Puls et al., 1999). The
12
horizontal distribution of the contribution of macrofauna to the depth-averaged SSC shown in Figure
13
6 b is similar to that of near-bed SSC depicted in Figure 5 h. Thus biological activities of macrofauna
14
as parameterized in the model impact not only the concentration at the sea floor but throughout the
15
water column. Both model results (near bed and depth averaged SSC) show a similar pattern of SSC
16
reduction where macrofauna is present (bluish colors in Figures 5 and 6) and increase adjacent to
17
these regions (reddish colors in Figures 5 and 6). Thus the modelled macrofauna effect extends
18
beyond the inhabitation zone.
AC C
EP
TE D
M AN U
SC
10
21
ACCEPTED MANUSCRIPT
M AN U
SC
RI PT
1
10
spring-neap-cycle (including a storm event) in February 2010 (Jan 31-Feb 13, shown as inset, upper panels a, b) and at the storm event in February 2010 (Feb 3, 2:00 am, location in tidal cycle marked with a red dot, lower panels c, d). The lefthand panels (a, c) show the concentration; the right-hand panels (b, d) show the included contribution of biotic benthic -2
modification. Contours denote the presence of F. fabula at 10 (thin line), 100, and 500 (thick line) individuals m .
EP
9
Figure 6: Simulated depth-averaged total suspended sediment concentration (SSC) temporally averaged over a
Summary statistics for n grid cells shown as mean µ and 5-95 percentiles p5,95.
3.2.3 Temporal mean macrofauna effects
AC C
3 4 5 6 7 8
TE D
2
For a spring-neap tidal cycle during a typical winter month (February 2010) including a storm
11
event, the patchiness of total mean depth-averaged SSC disappears (Figure 6 a), in contrast to the
12
storm event (Figure 6 c). This may indicate strong horizontal mixing of macrofauna-induced changes
13
in SSC during a spring-neap tidal cycle. Furthermore, it can be inferred from Figure 6 b that mean
14
depth-averaged SSC is generally enhanced by macrofauna in most areas of the southern North Sea,
15
also in contrast to the storm event. Specifically, contrary to the storm event, the SSC is increased
16
offshore of West Frisia due to a macrofauna contribution. The only area with slightly decreased SSC is 22
ACCEPTED MANUSCRIPT 1
south-east of Helgoland, i.e. the mouth of the Elbe estuary. The SSC has been increased due to
2
macrofauna mostly along the coastal belt and the effect is highest around the East Frisian Islands and
3
Sylt. The magnitude of increased SSC due to macrofauna effects is still significant (up to 5.24 g m-3),
4
but the maximum reduction of SSC reaches only -1 g m-3. These model results suggest that macrofauna affects the spatio-temporal distribution of SSC
6
beyond the inhabited areas, thus inducing a large-scale impact on SSC in the southern North Sea
7
regardless of physical condition.
8
4
RI PT
5
SC
Discussion
The modelled suspended sediment concentration increases due to macrofauna effects (Davis,
10
1993, Le Hir et al., 2007) beyond the inhabitation zones in the case of the fine silt class but not for
11
coarse silt (Figure 5). This can be explained by the lower settling velocity of this sediment class
12
(0.0001 m s-1 for dispersed grains, note that flocculation is not taken into account in the model)
13
compared to that of coarse silt, allowing transportation beyond their origin of erosion. For example,
14
for a residual tidal velocity (root mean square) of 0.2 m s-1 and a water depth of 10 m in shallow
15
areas, the fine silt can be transported 20 km. Moreover, the macrofauna contribution to fine silt is
16
almost five times greater than for coarse silt, which provides more suspended sediment to be
17
transported much farther, triggering a large-scale impact on SSC. The very fine silt at the bed was
18
rapidly exhausted due to its relatively low initial mass (1/5 of fine or coarse silt) and its very low
19
critical bed shear stress (0.1 N m-2), making this class of sediment in our model runs less sensitive to
20
the biologically mediated sediment properties at the bed (erodibility, critical bed shear stress).
21
Therefore, the difference of very fine silt in near-bed SSC between macrofaunal mediated sediment
22
transport and without macrofauna effect is small. Moreover, the very low settling velocity of these
23
particles (0.0002 m s-1) permitted a large spatial distribution of the slightly modified near-bed SSC by
24
horizontal transport processes beyond the inhabited zones and rendered effects virtually invisible in
25
the model runs. When interpreting the results, it has to be kept in mind that sediment availability,
26
i.e. the thickness of the initial sediment layer, is a calibration parameter in the model. Here we have
27
focused on the topmost layer of sediment only, which accounts for deposition and resuspension of
28
fine sediments but not for the more complex interaction between consolidated bed and active
29
sediment transport layer. In future studies the extension of the simple one-layer bed model towards
30
a multi-layer bed model representing stratigraphy is needed.
AC C
EP
TE D
M AN U
9
23
ACCEPTED MANUSCRIPT 1
The implication of the choice of model parameters becomes especially evident when considering sediment availability. This is essentially set by the initial sediment thickness, here chosen to be 10
3
mm. The unexpected decrease in modelled near-bed SSC in some areas, especially those with high
4
macrofauna abundance (>100 ind. m-2), can now be attributed either to a not sufficient initial
5
sediment thickness in the model or taken as process present in nature as well due to high
6
macrofauna mediated entrainment rates and limited sediment availability (e.g. Stammermann and
7
Piasecki, 2012). In model areas covered or dominated by F. fabula, erodibility increases and the
8
critical bed shear stress decreases. These two processes lead to higher and more frequent
9
resuspension, primarily over these zones. Consequently, the sediment mass is eroded more
RI PT
2
frequently and finally exhausted more quickly than is the case when macrofauna-related effects are
11
not considered. Figure 7 demonstrates this situation (referred to as the reference when the
12
macrofauna effect is included) for the area west of Sylt before and after the storm event.
M AN U
13
SC
10
It was found that, at the same location west of Sylt (Figure 5 d), macrofauna resulted in enhancement of the near-bed SSC of fine silt compared with a reduction for near-bed SSC of coarse
15
silt. The reason can be demonstrated by sediment availability in both cases (with and without
16
macrofauna effects) presented in Figure 7 a. These two figures indicate the importance of sediment
17
availability in the model as a limiting factor. Moreover, it can be observed in Figure 7 a that the
18
sediment mass decreases more rapidly when macrofauna effects on sediment erodibility and critical
19
bed shear stress are included (compare the slope of the two curves of the lower diagram just before
20
the storm event shown with a vertical green line). Sediment mass of coarse silt had been already
21
exhausted prior to the storm event on February 3 so that insufficient sediment was available for
22
erosion during the storm (Figure 7 b). Compared to the case without macrofauna effects, the
23
sediment mass is at least two orders of magnitude greater and the erosion process is not limited by
24
sediment availability. Consequently, the concentration of near-bed SSC during the storm event is
25
higher for the case without any macrofauna effect (light shaded area in the upper Figure 7 b).
AC C
EP
TE D
14
24
RI PT
ACCEPTED MANUSCRIPT
1 2 3 4
the contribution of biotic modification (blue and pink shading, upper panel) in the first two weeks of February 2010.
5
especially after the storm event (green line).
SC
Available mass is shown in the lower panel indicating frequent depletion of sediment available for resuspension,
M AN U
6
Figure 7: Near-bed simulated suspended sediment concentration (SSC, for fine and coarse silt fractions) including
Local in-situ measurements would be needed to further study the interaction of sediment
7
availability and macrofauna effects on sediment transport. The complexity of the interaction of
8
macrofauna and sediment transport was already highlighted by the laboratory flume experiments of
9
Friedrichs et al. (2009). They were able to quantify how the interaction of biogenic structures with the near-bed flow regime affects the sediment flux. At low population densities (roughness densities
11
below 2 %), there was an increase in the erosion fluxes and deposition of suspended material. At
12
densities above 4 %, reduced particle exchange was found and therefore erosion almost stopped
13
inside the test arrays. For roughness densities of about 5 %, it can therefore be assumed that a
14
skimming flow exists where a new benthic boundary layer eventually forms above the biogenic
15
structures. Furthermore filtration activities, especially by bivalves, could affect SSC in nature, even
16
though the extent of this effect is highly dependent on the abundance and the ambient flow
17
conditions (van Duren et al., 2006; Le Hir et al. 2007).
EP
AC C
18
TE D
10
Our results have important implications for ecosystem modeling. Light penetration through
19
the water column is affected by increased turbidity induced by macrofauna at the seabed, which
20
impacts primary production and population dynamics of phytoplankton. This finding indicates the
21
indirect coupling between macrofauna and phytoplankton through SSC. It should, however, be
22
investigated whether such a linkage is limited to storm events or may have long-term effects.
25
ACCEPTED MANUSCRIPT The model experiments which have been presented here partly depend on our choices of model
2
parameters. In the absence of sufficient data for the determination of sediment parameters (critical
3
bed shear stress, erodibility and sediment fractions as well as thickness in a few centimeters of the
4
upper bed layers), these were calibrated by comparing the simulated SSC with measured values or
5
values adopted from literature. The same shortage holds for F. fabula abundance data. This was
6
restricted to the locations of observation and was therefore extrapolated and interpolated for other
7
regions. Seasonal variations in macrofauna abundance were not considered. Furthermore, in the
8
present study, the macrofauna effect of a single dominant species was considered, ignoring the
9
effect of several other species and their non-linear interaction within the community. Most
RI PT
1
importantly the effect of suspension feeders which will introduce a net sink of SSC has not been
11
included so far. The parameterizations adopted for the inclusion of macrofauna effects were based
12
on very limited data for Macoma balthica from literature, which was adopted for F. fabula. These
13
effects can be highly non-linear. As presented in Eq. 6 erodibility increases exponentially between an
14
abundance of 100 to 800 ind. m-2,following a sigmoid function, leading to greatly increased erosion
15
rates in inhabited areas and finally to exhaustion of the available bed material. Another possible
16
limitation of our model is that macrofauna effects were assumed to be constant over time and other
17
macrofauna effects such as those on roughness or biodeposition were not considered. In the case of
18
F. fabula, however, the temporal variability of abundance is known to be limited (Creutzberg, 1986).
19
We did not consider the effect of biological reworking on the migration of bed forms (Borsje et al.,
20
2009) since we focused on suspended sediment transport.
TE D
M AN U
SC
10
Finally, we propose more laboratory experiments addressing the issue of how macrobenthic
22
organisms (also in the form of communities) alter sediment transport properties. This would greatly
23
reduce the uncertainty in critical model formulations. In addition, future work would greatly benefit
24
by the availability of more SSC measurements in the German Bight and ideally data based large-scale
25
SSC estimates in addition to small-scale studies measuring simultaneously hydrodynamic, SSC as well
26
as bed sediment and biological parameters. This may result in new and more accurate and integral
27
(because biological) parameterizations of sediment transport models.
28
5
AC C
EP
21
Conclusions
29
The observed influence of macrofauna on sediments at the seabed and the seabed structure
30
itself has been investigated in terms of large-scale impacts on suspended sediment concentration in 26
ACCEPTED MANUSCRIPT a model study of a shallow shelf sea. The southern North Sea has been taken as an example in order
2
to study the impact of macrofauna in a numerical model by integrating a generic benthos module in a
3
coupled 3D model system. Modification of critical bed shear stress and erodibility constituted the
4
major proxies for macrofauna impacts on sediment. A feature of the present study is the focus on
5
large-scale effects (>100 km), combined with the application of the measured spatial distribution of
6
the bivalve F. fabula as a chosen characteristic species.
RI PT
1
7
The simulation results indicate that macrofauna is not only able to significantly modify SSC locally
8
but also beyond the inhabited zones. As expected, the enhanced SSC due to macrofauna is most
9
pronounced for high energy conditions such as storm events but also persists over a spring-neap tidal cycle. Changes in SSC were not confined to near-bed zones but extended throughout the water
11
column due to vertical mixing.
Furthermore, it was shown that the magnitude and horizontal extent of macrofauna modified
M AN U
12
SC
10
SSC are constrained by sediment availability (in the model) and depend on the sediment properties
14
such as settling velocity and critical bed shear stress for deposition and erosion. Fine silt, being the
15
dominant fraction for suspended sediment in the water column, was found to be very sensitive to
16
the occurrence of macrofauna in the model. While the present study contributes to the improvement
17
of ecosystem and sediment transport modeling by offering new insights into large-scale effects of
18
macrofauna on SSC in the southern North Sea, it should be noted that physical and biological
19
processes in the model are less complex than in nature and rely partly on simplified
20
parameterizations, which opens up various new research issues, both in empirical science and
21
modeling.
22
6
EP
Acknowledgement
AC C
23
TE D
13
This paper is the result of collaboration between the MOSSCO project partners funded by the
24
German Federal Ministry of Education and Research (BMBF) as part of the Coastal Research Agenda
25
for the North Sea and Baltic Sea (grant numbers 03F0668A, 03F0667A, 03F0667B ). Further financial
26
support of K. Klingbeil has been provided by the Collaborative Research Center TRR181 on Energy
27
Transfers in Atmosphere and Ocean, funded by the German Research Foundation. We acknowledge
28
the contribution of J. Dannheim and M. Zeiler of the German Federal Maritime and Hydrographic
29
Agency (BSH) for providing F. fabula abundance data and Wenyan Zhang from the Helmholtz27
ACCEPTED MANUSCRIPT Zentrum Geesthacht (PACES Programme of the Helmholtz-Gemeinschaft) for completing the data set
2
for the southern North Sea. We thank A. Hammrich for contributions to the model validation.
3
Processed satellite data were provided by Brockmann Consult. Moreover, we thank Stefan Foster
4
and two anonymous reviewers as well as Christian Winter as guest editor for their helpful
5
recommendations and suggestions which improved our manuscript significantly.
6
7
7 8
Literature
RI PT
1
Andersen, T., and M. Pejrup (2011), Biological Influences on Sediment Behavior and Transport, in Treatise on Estuarine and Coastal Science, pp. 289–309, Elsevier.
Arfai, J., Jähne, F., Lutz, R., Franke, D., Gaedicke, C. & Kley, J., (2014). Late Palaeozoic to Early
10
Cenozoic geological evolution of the northwestern German North Sea (Entenschnabel): New results
11
and insights. Netherlands Journal of Geosciences -Geologie en Mijnbouw, 93(4): 147-174.
M AN U
12
SC
9
Austen, I., T. J. Andersen, and K. Edelvang (1999), The Influence of Benthic Diatoms and
13
Invertebrates on the Erodibility of an Intertidal Mudflat, the Danish Wadden Sea, Estuarine, Coastal
14
and Shelf Science, 49(1), 99–111, doi:10.1006/ecss.1998.0491.
16 17
Becker, G. A., S. Dick and J. W. Dippner (1992), Hydrography of the German Bight, Marine Ecology Progress Series, 91, 9-18.
TE D
15
Borsje, B. W., S. Hulscher, M. B. de Vries, and G. J. de Boer (2007), Modeling large scale cohesive
18
sediment transport by including biological activity, River, Coastal and Estuarine Morphodynamics:
19
RCEM 2007, 255–262.
Borsje, B. W., Hulscher,Suzanne Jacqueline Marie Hélène, P. M. J. Herman, and M. B. Vries
EP
20
(2009), On the parameterization of biological influences on offshore sand wave dynamics, Ocean
22
Dynamics, 59(5), 659–670, doi:10.1007/s10236-009-0199-0.
AC C
21
23
Bremner, J., S. Rogers, and C. Frid (2006), Methods for describing ecological functioning of
24
marine benthic assemblages using biological traits analysis (BTA), Ecological Indicators, 6(3), 609–
25
622, doi:10.1016/j.ecolind.2005.08.026.
26
Breugem, W. A., and L. H. Holthuijsen (2007), Generalized Shallow Water Wave Growth from
27
Lake George, J. Waterway, Port, Coastal, Ocean Eng., 133(3), 173–182, doi:10.1061/(ASCE)0733-
28
950X(2007)133:3(173). 28
ACCEPTED MANUSCRIPT 1
Briggs, K. B., G. Cartwright, C. T. Friedrichs, and S. Shivarudruppa (2015), Biogenic effects on
2
cohesive sediment erodibility resulting from recurring seasonal hypoxia on the Louisiana shelf,
3
Continental Shelf Research, 93, 17–26, doi:10.1016/j.csr.2014.11.005.
4 5
Burchard, H., and K. Bolding (2002), GETM, A General Estuarine Transport Model: Scientific Documentation, Technical Report EUR 20253 EN. European Commission. Burchard, H., G. Flöser, J. V. Staneva, T. H. Badewien, and R. Riethmüller (2008), Impact of
7
Density Gradients on Net Sediment Transport into the Wadden Sea, J. Phys. Oceanogr., 38(3), 566–
8
587, doi:10.1175/2007JPO3796.1.
9
RI PT
6
Canuto, V. M., A. Howard, Y. Cheng, and M. S. Dubovikov (2001), Ocean Turbulence. Part I: OnePoint Closure Model—Momentum and Heat Vertical Diffusivities, J. Phys. Oceanogr., 31(6), 1413–
11
1426, doi:10.1175/1520-0485(2001)031<1413:OTPIOP>2.0.CO;2.
SC
10
Creutzberg, F. (1986), Distribution patterns of two bivalve species (Nucula turgida, Tellina fabula)
13
along a frontal system in the Southern North Sea, Netherlands Journal of Sea Research, 20(2-3), 305–
14
311, doi:10.1016/0077-7579(86)90052-9.
15
M AN U
12
Dannheim, J. (2014), Bewertungsansätze für Raumordnung und Genehmigungsverfahren im Hinblick auf das benthische System und Habitatstrukturen, Statusbericht und Web-Dienst im
17
GeoSeaPortal des BSH: Forschungsvorhaben im Auftrag des BSH.
18
TE D
16
Darr, A., M. Gogina, and M. L. Zettler (2014), Functional changes in benthic communities along a
19
salinity gradient– a western Baltic case study, Journal of Sea Research, 85, 315–324,
20
doi:10.1016/j.seares.2013.06.003.
Davis, W. R. (1993), The role of bioturbation in sediment resuspension and its interaction with
EP
21
physical shearing, Journal of Experimental Marine Biology and Ecology, 171(2), 187–200,
23
doi:10.1016/0022-0981(93)90003-7.
24 25 26
AC C
22
Doerffer, R., H. Schiller, and M. Peters (2006), Meris regional case 2 water algorithms (c2r), www.brockmann-consult.de/beam/software/plugins/merisc2r-1.1. Duineveld, G., A. Künitzer, U. Niermann, P. de Wilde, and J. S. Gray (1991), The macrobenthos of
27
the north sea, Netherlands Journal of Sea Research, 28(1-2), 53–65, doi:10.1016/0077-
28
7579(91)90004-K.
29
ACCEPTED MANUSCRIPT 1
François, F., J.-C. Poggiale, J.-P. Durbec, and G. Stora (1997), A New Approach for the Modelling
2
of Sediment Reworking Induced by a Macrobenthic Community, Acta Biotheoretica, 45(3/4), 295–
3
319, doi:10.1023/A:1000636109604.
5 6
Friedrichs, M., T. Leipe, F. Peine and G. Graf (2009). "Impact of macrozoobenthic structures on near-bed sediment fluxes." Journal of Marine Systems 75(3-4): 336-347. Gayer, G., S. Dick, A. Pleskachevsky, and W. Rosenthal (2006), Numerical modeling of suspended
RI PT
4
7
matter transport in the North Sea, Ocean Dynamics, 56(1), 62–77, doi:10.1007/s10236-006-0070-5.
8
Graf, G., and R. Rosenberg (1997), Bioresuspension and biodeposition: a review, Journal of
10
Marine Systems, 11(3-4), 269–278, doi:10.1016/S0924-7963(96)00126-1.
SC
9
Gräwe, U., G. Flöser, T. Gerkema, M. Duran-Matute, T. H. Badewien, E. Schulz, and H. Burchard (2016), A numerical model for the entire Wadden Sea: Skill assessment and analysis of
12
hydrodynamics, J. Geophys. Res. Oceans, 121(7), 5231–5251, doi:10.1002/2016JC011655.
13
M AN U
11
Harris, R. J., C. A. Pilditch, B. L. Greenfield, V. Moon, and I. Kröncke (2016), The Influence of
14
Benthic Macrofauna on the Erodibility of Intertidal Sediments with Varying mud Content in Three
15
New Zealand Estuaries, Estuaries and Coasts, 39(3), 815–828, doi:10.1007/s12237-015-0036-2.
17 18
Hayter, E. J. and A. J. Mehta (1986), Modelling cohesive sediment transport in estuarial waters. In: Applied Mathematical Modelling, 10 (4), 294–303, doi: 10.1016/0307-904X(86)90061-2.
TE D
16
Heath, M. R., A. C. Edwards, J. Patsch, and Turrell Scottish Executive Central Research Unit (2002), Modelling the behaviour of nutrients in the coastal waters of Scotland, nowel,
20
http://strathprints.strath.ac.uk/18568/.
22
Hill, C., C. DeLuca, Balaji, M. Suarez, and A. Da Silva (2004), The architecture of the earth system modeling framework, Comput. Sci. Eng., 6(1), 18–28, doi:10.1109/MCISE.2004.1255817.
AC C
21
EP
19
23
Hofmeister, R., K. Bolding, R. D. Hetland, G. Schernewski, H. Siegel, and H. Burchard (2013), The
24
dynamics of cooling water discharge in a shallow, non-tidal embayment, Continental Shelf Research,
25
71, 68–77, doi:10.1016/j.csr.2013.10.006.
26
Jumars, P. A., and A. Nowell (1984), Effects of benthos on sediment transport: difficulties with
27
functional grouping, Continental Shelf Research, 3(2), 115–130, doi:10.1016/0278-4343(84)90002-5.
28
Kappenberg, J., and I. Grabemann (2001), Variability of the Mixing Zones and Estuarine Turbidity
29
Maxima in the Elbe and Weser Estuaries, Estuaries, 24(5), 699, doi:10.2307/1352878. 30
ACCEPTED MANUSCRIPT 1
Kerimoglu, O., R. Hofmeister, J. Maerz, and K. Wirtz (2017), A novel acclimative biogeochemical
2
model and its implementation to the southern North Sea, Biogeosciences Discuss., doi:10.5194/bg-
3
2017-104.
4
Klingbeil, K., and H. Burchard (2013), Implementation of a direct nonhydrostatic pressure gradient discretisation into a layered ocean model, Ocean Modelling, 65, 64–77,
6
doi:10.1016/j.ocemod.2013.02.002.
RI PT
5
7
Klingbeil, K., M. Mohammadi-Aragh, U. Gräwe, and H. Burchard (2014), Quantification of
8
spurious dissipation and mixing – Discrete variance decay in a Finite-Volume framework, Ocean
9
Modelling, 81, 49–64, doi:10.1016/j.ocemod.2014.06.001.
Knaapen, M., H. Holzhauer, S. Hulscher, M. J. Baptist, M. d. Vries, and M. van Ledden (2003), On
11
the modelling of biological effects on morphology in estuaries and seas, edited by A. Sánchez-Arcilla
12
and A. Bateman, pp. 773–783, IHAR.
M AN U
13
SC
10
Krause, G., G. Budeus, D. Gerdes, K. Schaumann, K. Hesse (1986), Frontal systems in the German
14
Bight and their physical and biological effects, in: Nihoul, J.C.J. (Ed.) Marine interfaces
15
ecohydrodynamics: proceedings of the 17th International Liège Colloquium on Ocean
16
Hydrodynamics. Elsevier Oceanography Series, 42, 119-140
Kristensen, E., J. M. Neto, M. Lundkvist, L. Frederiksen, M. Â. Pardal, T. Valdemarsen, and M. R.
TE D
17 18
Flindt (2013), Influence of benthic macroinvertebrates on the erodability of estuarine cohesive
19
sediments: Density- and biomass-specific responses, Estuarine, Coastal and Shelf Science,
20
doi:10.1016/j.ecss.2013.09.020.
Kröncke, I., H. Reiss, J. D. Eggleton, J. Aldridge, M. J. Bergman, S. Cochrane, J. A. Craeymeersch, S.
EP
21
Degraer, N. Desroy, J.-M. Dewarumez, G. C. Duineveld, K. Essink, H. Hillewaert, M. S. Lavaleye, A.
23
Moll, S. Nehring, R. Newell, E. Oug, T. Pohlmann, E. Rachor, M. Robertson, H. Rumohr, M.
24
Schratzberger, R. Smith, E. V. Berghe, J. van Dalfsen, G. van Hoey, M. Vincx, W. Willems, and H. L.
25
Rees (2011), Changes in North Sea macrofauna communities and species distribution between 1986
26
and 2000, Estuarine, Coastal and Shelf Science, 94(1), 1–15, doi:10.1016/j.ecss.2011.04.008.
27
Krumbein, W. E. (1994), Biostabilization of sediments, 526 pp., Carl-von-Ossietzky-Univ.
28
AC C
22
Oldenburg, Oldenburg.
31
ACCEPTED MANUSCRIPT 1
Le Hir, P., Y. Monbet, and F. Orvain (2007), Sediment erodability in sediment transport modelling:
2
Can we account for biota effects?, Continental Shelf Research, 27(8), 1116–1142,
3
doi:10.1016/j.csr.2005.11.016.
5 6
Le Hir, P., Cayocca, F., & Waeles, B. (2011). Dynamics of sand and mud mixtures: a multiprocessbased modelling strategy. Continental Shelf Research, 31(10), S135-S149. Lemmen, C., Hofmeister, R., Klingbeil, K., Nasermoaddeli, M.H., Kerimoglu, O., Burchard, H.,
RI PT
4
7
Kösters, F., Wirtz, K.W. (2017), Modular System for Shelves and Coasts (MOSSCO v1.0) - a flexible and
8
multi-component framework for coupled coastal ocean ecosystem modelling, Geosci. Model Dev.
9
submitted., http://arxiv.org/abs/1706.04224
Lettmann, K. A., J.-O. Wolff, and T. H. Badewien (2009), Modeling the impact of wind and waves
11
on suspended particulate matter fluxes in the East Frisian Wadden Sea (southern North Sea), Ocean
12
Dynamics, 59(2), 239–262, doi:10.1007/s10236-009-0194-5.
M AN U
13
SC
10
Lumborg, U., T. J. Andersen, and M. Pejrup (2006), The effect of Hydrobia ulvae and
14
microphytobenthos on cohesive sediment dynamics on an intertidal mudflat described by means of
15
numerical modelling, Estuarine, Coastal and Shelf Science, 68(1–2), 208–220,
16
doi:10.1016/j.ecss.2005.11.039.
Maa, J. P.-Y., J.-I. Kwon, K.-N. Hwang, and H.-K. Ha (2008), Critical Bed-Shear Stress for Cohesive
TE D
17 18
Sediment Deposition under Steady Flows, J. Hydraul. Eng., 134(12), 1767–1771,
19
doi:10.1061/(ASCE)0733-9429(2008)134:12(1767).
20
Montserrat Trotsenburg, F. (2011), Estuarine ecosystem engineering: biogeomorphology in the estuarine intertidal, Dissertation, Civil Engineering and Geosciences, Hydraulic Engineering,
22
Technische Universiteit Delft, Delft.
Nasermoaddeli, M. H., C. Lemmen, R. Hofmeister, F. Koesters, and K. Klingbeil (2014), The
AC C
23
EP
21
24
benthic geoecology model within the modular system for shelves and coasts (MOSSCO), in HIC2014-
25
11. International Conference on Hydroinformatics: Informatics and the Environment: Data and Model
26
Integration in a Heterogeneous Hydro World, edited by M. Piaseki, City University of New York, New
27
York, USA.
28
Neumann, H., H. Reiss, S. Ehrich, A. Sell, K. Panten, M. Kloppmann, I. Wilhelms, and I. Kröncke
29
(2012), Benthos and demersal fish habitats in the German Exclusive Economic Zone (EEZ) of the
30
North Sea, Helgol Mar Res, doi:10.1007/s10152-012-0334-z. 32
ACCEPTED MANUSCRIPT 1 2
Nowell, A. R. M., and P. A. Jumars (1984), Flow environments of aquatic benthos, Annual Review of Ecology and Systematics, 15, 303–328.
3
Orvain, F. (2005), A model of sediment transport under the influence of surface bioturbation:
4
generalisation to the facultative suspension-feeder Scrobicularia plana, Mar. Ecol. Prog. Ser., 286,
5
43–56, doi:10.3354/meps286043. Orvain, F., K. Guizien, S. Lefebvre, M. Bréret, and C. Dupuy (2014), Relevance of macrozoobenthic
RI PT
6 7
grazers to understand the dynamic behaviour of sediment erodibility and microphytobenthos
8
resuspension in sunny summer conditions, Journal of Sea Research,
9
doi:10.1016/j.seares.2014.03.004.
Orvain, F., P. Le Hir, and P.-G. Sauriau (2003), A model of fluff layer erosion and subsequent bed
SC
10
erosion in the presence of the bioturbator, Hydrobia ulvae, J Mar Res, 61(6), 821–849,
12
doi:10.1357/002224003322981165.
13
M AN U
11
Orvain, F., P. Le Hir, P.-G. Sauriau, and S. Lefebvre (2012), Modelling the effects of macrofauna on sediment transport and bed elevation: Application over a cross-shore mudflat profile and model
15
validation, Estuarine, Coastal and Shelf Science, 108, 64–75, doi:10.1016/j.ecss.2011.12.036.
16
Orvain, F., P.-G. Sauriau, C. Bacher, and M. Prineau (2006), The influence of sediment
17
cohesiveness on bioturbation effects due to Hydrobia ulvae on the initial erosion of intertidal
18
sediments: A study combining flume and model approaches, Journal of Sea Research, 55(1), 54–73,
19
doi:10.1016/j.seares.2005.10.002.
TE D
14
Paarlberg, A. J., M. Knaapen, M. B. de Vries, S. Hulscher, and Z. B. Wang (2005), Biological
21
influences on morphology and bed composition of an intertidal flat, Estuarine, Coastal and Shelf
22
Science, 64(4), 577–590, doi:10.1016/j.ecss.2005.04.008.
24 25 26 27
Partheniades, E. (1965), Erosion and Deposition of Cohesive Soils, Journal of the Hydraulics
AC C
23
EP
20
Division, 91(1), 105–139.
Peine, F., B. Bobertz, and G. Graf (2005), Influence of the blue mussel Mytilus edulis (Linnaeus) on the bottom roughness length (z0) in the south-western Baltic Sea, Baltica, 18(1), 13–22. Pleskachevsky, A., G. Gayer, J. Horstmann, and W. Rosenthal (2005), Synergy of satellite remote
28
sensing and numerical modeling for monitoring of suspended particulate matter, Ocean Dynamics,
29
55(1), 2–9, doi:10.1007/s10236-004-0101-z. 33
ACCEPTED MANUSCRIPT 1 2 3
Postma, H. (1984), Introduction to the symposium on organic matter in the Wadden Sea, Neth. Inst. Sea Res. Pub. Ser., 10, 15–22. Puls, W., J. Beusekom, U. Brockmann, R. Doerffer, U. Hentschke, P. König, D. Murphy, B. Mayer, A. Müller, T. Pohlmann, A. Reimer, R. Schmidt-Nia, and J. Sündermann (1999), SPM concentrations in
5
the German Bight: Comparison between a model simulation and measurements, Deutsche
6
Hydrographische Zeitschrift, 51(2-3), 221–244, doi:10.1007/BF02764175.
7
RI PT
4
Puls, W., H. Kühl, A. Frohse, and P. König (1995), Measurements of the suspended matter settling
8
velocity in the German Bight (North Sea), Deutsche Hydrographische Zeitschrift, 47(4), 259–276,
9
doi:10.1007/BF02737787.
Puls, W., T. Pohlmann, and J. Sündermann (1997), Suspended particulate matter in the Southern
SC
10
North Sea: Application of a numerical model to extend NERC North Sea project data interpretation,
12
Deutsche Hydrographische Zeitschrift, 49(2-3), 307–327, doi:10.1007/BF02764041.
13 14 15
M AN U
11
Rachor, E., and P. Nehmer (2003), Erfassung und Bewertung ökologisch wertvoller Lebensräume in der Nordsee, 175 pp., Bundesamt für Naturschutz, Bremerhaven, hdl:10013/epic.36744.d001. Radach, G., and J. Pätsch (2007), Variability of continental riverine freshwater and nutrient inputs into the North Sea for the years 1977–2000 and its consequences for the assessment of
17
eutrophication, Estuaries and Coasts, 30(1), 66–81, doi:10.1007/BF02782968.
18 19
TE D
16
Reiss, H., S. Cunze, K. König, H. Neumann, and I. Kröncke (2011), Species distribution modelling of marine benthos: a North Sea case study, Mar. Ecol. Prog. Ser., 442, 71–86, doi:10.3354/meps09391. Rijkswaterstaat (2017), Waterbase, http://live.waterbase.nl/ [accessed 19 Jan 2017].
21
Salzwedel, H., E. Rachor, and D. Gerdes (1985), Benthic macrofauna communities in the German
23
Bight, Veröff. Inst. Meeresforsch. Bremerh., 20, 199–267.
AC C
22
EP
20
Sanford, L. P. (2008), Modeling a dynamically varying mixed sediment bed with erosion,
24
deposition, bioturbation, consolidation, and armoring, Computers & Geosciences, 34(10), 1263–1283,
25
doi:10.1016/j.cageo.2008.02.011.
26
Schrum, C. (1997), Thermohaline stratification and instabilities at tidal mixing fronts: Results of
27
an eddy resolving model for the German Bight, Continental Shelf Research, 17(6), 689–716,
28
doi:10.1016/S0278-4343(96)00051-9.
34
ACCEPTED MANUSCRIPT 1
Seifert, T., W. Fennel, and C. Kuhrts (2009), High resolution model studies of transport of
2
sedimentary material in the south-western Baltic, Journal of Marine Systems, 75(3-4), 382–396,
3
doi:10.1016/j.jmarsys.2007.01.017.
4
Shimeta, J., C. L. Amos, S. E. Beaulieu, and O. M. Ashiru (2002), Sequential resuspension of protists by accelerating tidal flow: implications for community structure in the benthic boundary
6
layer, Limnol. Oceangr., 47, 1152–1164.
8 9 10 11
Smagorinsky, J. (1963), General circulation experiments with the primitive equations, Mon. Wea. Rev., 91(3), 99–164, doi:10.1175/1520-0493(1963)091<0099:GCEWTP>2.3.CO;2.
Soulsby, R. (1997), Dynamics of marine sands: A manual for practical applications, 1st ed., Telford, London.
SC
7
RI PT
5
Stammermann, R., and M. Piasecki (2012), Influence of Sediment Availability, Vegetation, and Sea Level Rise on the Development of Tidal Marshes in the Delaware Bay: A Review, Journal of
13
Coastal Research, 285, 1536–1549, doi:10.2112/JCOASTRES-D-11-00143.1.
14
M AN U
12
Stanev, E. V., R. Al-Nadhairi, J. Staneva, J. Schulz-Stellenfleth, and A. Valle-Levinson (2014), Tidal
15
wave transformations in the German Bight, Ocean Dynamics, 64(7), 951–968, doi:10.1007/s10236-
16
014-0733-6.
Stanev, E. V., M. Dobrynin, A. Pleskachevsky, S. Grayek, and H. Günther (2009), Bed shear stress
TE D
17 18
in the southern North Sea as an important driver for suspended sediment dynamics, Ocean
19
Dynamics, 59(2), 183–194, doi:10.1007/s10236-008-0171-4. Stips, A., K. Bolding, T. Pohlmann, and H. Burchard (2004), Simulating the temporal and spatial
21
dynamics of the North Sea using the new model GETM (general estuarine transport model), Ocean
22
Dynamics, 54(2), 266–283, doi:10.1007/s10236-003-0077-0. Umlauf, L., and H. Burchard (2005), Second-order turbulence closure models for geophysical
AC C
23
EP
20
24
boundary layers. A review of recent work, Continental Shelf Research, 25(7-8), 795–827,
25
doi:10.1016/j.csr.2004.08.004.
26
van der Molen, J., K. Bolding, N. Greenwood, and D. K. Mills (2009), A 1-D vertical multiple grain
27
size model of suspended particulate matter in combined currents and waves in shelf seas, J. Geophys.
28
Res., 114(F1), doi:10.1029/2008JF001150.
35
ACCEPTED MANUSCRIPT 1 2
van Duren, L. A., P. M. J. Herman, et al. (2006), Effects of mussel filtering activity on boundary layer structure, Journal of Sea Research, 55(1), 3-14.
3
van Ledden, M. (2003), Sand-mud segregation in estuaries and tidal basins, xvi, 218,
4
Communications on hydraulic and geotechnical engineering, report no. 03-2, Dept. of Civil
5
Engineering and Geosciences, Delft University of Technology, Delft. van Ledden, M., W. van Kesteren, and J. Winterwerp (2004), A conceptual framework for the
RI PT
6 7
erosion behaviour of sand–mud mixtures, Continental Shelf Research, 24(1), 1–11,
8
doi:10.1016/j.csr.2003.09.002.
9
van Leeuwen, S. M., J. van der Molen, P. Ruardij, L. Fernand, and T. Jickells (2013), Modelling the contribution of deep chlorophyll maxima to annual primary production in the North Sea,
11
Biogeochemistry, 113(1-3), 137–152, doi:10.1007/s10533-012-9704-5.
SC
10
13 14
van Moorsel, G. (2011), Species and habitats of the international Dogger Bank, ecosub, Doorn. 74 p.
M AN U
12
van Prooijen, Bram C., F. Montserrat, and P. M. Herman (2011), A process-based model for
15
erosion of Macoma balthica-affected mud beds, Continental Shelf Research, 31(6), 527–538,
16
doi:10.1016/j.csr.2010.12.008.
18 19
Wentworth, C.K., 1922, A scale of grade and class terms for clastic sediments: Journal of Geology, v. 30, p. 377-392.
TE D
17
Wheatcroft, R. A., and C. A. Butman (1997), Spatial and temporal variability in aggregated grainsize distributions, with implications for sediment dynamics, Continental Shelf Research, 17(4), 367–
21
390, doi:10.1016/S0278-4343(96)00035-0.
22
EP
20
Widdows, J., M. Brinsley, and M. Elliott (1998), Use of in situ flume to quantify particle flux (biodeposition rates and sediment erosion) for an intertidal mudflat in relation to changes in current
24
velocity and benthic macrofauna, Geological Society, London, Special Publications, 139(1), 85–97,
25
doi:10.1144/GSL.SP.1998.139.01.07.
26
AC C
23
Widdows, J., M. Brinsley, P. N. Salkeld, and C. H. Lucas (2000a), Influence of biota on spatial and
27
temporal variation in sediment erodability and material flux on a tidal flat (Westerschelde, The
28
Netherlands), Marine Ecology Progress Series, 194, 23–37.
36
ACCEPTED MANUSCRIPT 1
Widdows, J., S. Brown, M. Brinsley, P. Salkeld, and M. Elliott (2000b), Temporal changes in
2
intertidal sediment erodability: influence of biological and climatic factors, Continental Shelf
3
Research, 20(10-11), 1275–1289, doi:10.1016/S0278-4343(00)00023-6. Winterwerp, J. C. (2007), On the sedimentation rate of cohesive sediment, in Estuarine and
5
Coastal Fine Sediments Dynamics - Intercoh 2003, Proceedings in Marine Science, pp. 209–226,
6
Elsevier.
RI PT
4
7
Winterwerp, J. C., and W. G. M. van Kesteren (2004), Introduction to the physics of cohesive
8
sediment in the marine environment, 1st ed., 1 online resource (1 volume (various, Developments in
9
sedimentology, vol. 56, Elsevier, Amsterdam.
Wood, R., and J. Widdows (2002), A model of sediment transport over an intertidal transect,
SC
10
comparing the influences of biological and physical factors, Limnol. Oceangr., 47(3), 848–855,
12
doi:10.4319/lo.2002.47.3.0848.
M AN U
11
13
AC C
EP
TE D
14
37
ACCEPTED MANUSCRIPT
1
1
2
1.1
3
Appendix – Hydrodynamic model validation Introduction A brief hydrodynamic model validation for water levels and waves is presented in the following
for the reference year 2010. The validation – as far as possible on available data – for suspended
5
sediment concentration can be found in the main article. A detailed validation for the simulation of
6
temperature and salinity characteristics (e.g. thermohaline stratification, residual circulation) can be
7
found in Kerimoglu et al. (2017).
8
1.2
SC
RI PT
4
Observational data
The model validation was carried out for a number of stations along the Dutch and German
10
coasts. The analysis was restricted to positions representing the dynamics of the German Bight, thus
11
locations within the estuaries have not been analyzed. An overview of observational data used for
12
the model validation is given in Table 4. The location and measured properties of the individual
13
stations are shown in Figure 8.
TE D
waves (WAV), suspended sediment concentration (SSC) and total suspended matter (TSM).
Station Name
Abbreviation
Terschelling-Noordzee Huibertgat Borkum Fischerbalje Norderney-Riffgat Wangerooge-Nord Bake Z Cuxhaven Helgoland Wyk List
TSN HUI BOF NOY WAN BKZ CUX HEL WYK LIH
AC C
15
Table 4: Overview of observational data used for the model validation. Measured parameters are water levels (WL),
16
Position Latitude (deg.) Longitude (deg.) 53.450 5.333 53.567 6.400 53.558 6.748 53.697 7.158 53.806 7.929 54.014 8.315 53.868 8.718 54.179 7.890 54.694 8.577 55.017 8.441
EP
14
M AN U
9
Fireship "Elbe" List wave buoy FINO-I plattform
ELB LWB FN1
54.000 55.054 54.015
8.117 8.390 6.588
Terschelling T10 Terschelling T50 Rottumerplaat R3 Rottumerplaat R50 Sylt S9 Sylt S62
T10 T50 R3 R50 S9 S62
53.454 53.764 53.557 53.947 54.909 54.867
5.077 4.801 6.547 6.314 8.150 7.271
Measured Parameter WL WAV SSC TSM x x x x x x x x x x x x x
Data Source Rijkswaterstaat Rijkswaterstaat WSV WSV WSV WSV WSV WSV LKN WSV BSH COSYNA BSH
x x x x
x x x x x x
Rijkswaterstaat, Brockmann Consult / ESA Rijkswaterstaat, Brockmann Consult / ESA Rijkswaterstaat, Brockmann Consult / ESA Rijkswaterstaat, Brockmann Consult / ESA Brockmann Consult / ESA Brockmann Consult / ESA
38
ACCEPTED MANUSCRIPT 1
Observational data for water levels are based on tide gauge stations which were provided for the Netherlands by Rijkswaterstaat and for Germany by the Waterways and Shipping Administration
3
(WSV) and the Landesbetrieb für Küstenschutz, Nationalpark und Meeresschutz Schleswig-Holstein
4
(LKN). Wave measurements are based on wave buoy data provided by the German Federal Maritime
5
and Hydrographic Agency (BSH) and the Helmholtz-Zentrum Geesthacht (HZG) within the COSYNA
6
framework. Suspended sediment concentrations were obtained from monitoring stations provided
7
by Rijkswaterstaat and extracted from satellite measurements of total suspended matter carried out
8
by the European Space Agency (ESA) and processed by Brockmann Consult.
TE D
M AN U
SC
RI PT
2
EP
9 10
waves (WAV) and suspended sediment concentration (SSC / TSM).
AC C
11
Figure 8: Location of observation stations considered for the model validation for parameters water level (WL),
12
The propagation of the tidal wave from west to east along the coast can be fairly well covered
13
with the available data. Wave data is less frequently monitored but for the time period considered
14
here positions representing the inner German Bight (Fino-I), the outer estuaries (Elbe) and a near
15
coastal position (List) were available.
16
1.3
17
1.3.1 Water levels
Model validation
39
ACCEPTED MANUSCRIPT 1
As validation period for water levels a period of about two spring-neap cycles from July 19, 2010 to August 20, 2010 was chosen. Inspection of individual time series of water levels shows general
3
good agreement of observations and simulation (see Figure 9 for station Norderney Riffgat as an
4
example) in terms of tidal range and phase of the tidal wave.
M AN U
SC
RI PT
2
7 8 9
Figure 9: Simulated (red) and observed (blue) water levels for station Norderney Riffgat for validation period July 19, 2010 to August 20, 2010
As can be seen from the quantitative comparison of simulations and observations of water levels given in Table 5 the model is generally capable of reproducing the tidal range well. The difference
EP
6
TE D
5
between simulated and observed values is less than 10 % at most stations, larger differences occur at
11
station Wyk, a location strongly influenced by tidal flat dynamics and Cuxhaven. The correlation
12
between model and measurements is with a mean value for the correlation coefficient of 0.92 high.
13
There is a small bias in the model in terms of mean water level which is on average 12 cm too high.
14
There is also a phase difference for the propagation of the tidal wave by on average 25 minutes.
15
Generally the model lags observations but this trend is not present at all stations. Correcting the
16
simulated water levels at individual stations for this phase shift we obtain an RMSE of less than 30 cm
17
on average and for the standard RMSE of less than 40 cm.
AC C
10
40
ACCEPTED MANUSCRIPT 1
Table 5: Water level statistics comparing simulated (sim) and observed (obs) values of tidal range (TR), phase
2
difference for the propagation of the tidal wave, root mean square error (RMSE), phase corrected root mean square
3
error (RMSE*), model bias and correlation coefficient (Pearson R) for individual stations and mean over all stations
4 5
1.4
TR sim (m) 1.75 1.86 1.97 2.00 2.46 2.54 2.96 1.97 2.00 1.41 2.09
∆ TR obs-sim (m / %) 0.02 1% -0.10 -5% -0.12 -6% -0.16 -7% -0.04 -1% -0.01 0% 0.31 12% -0.11 -5% -0.43 -18% -0.16 -10% -0.08 -4%
Phase diff. (min) -26 -36 -23 -21 -27 -32 -30 -26 -64 35 -25
RMSE (m) 0.28 0.33 0.32 0.31 0.34 0.37 0.46 0.29 0.61 0.29 0.38
RMSE* (m) 0.23 0.23 0.28 0.27 0.25 0.24 0.36 0.23 0.40 0.23 0.27
Bias (m) 0.12 0.09 0.09 0.14 0.12 0.09 0.12 0.11 0.19 0.12 0.12
Pearson R 0.95 0.92 0.94 0.95 0.95 0.94 0.93 0.95 0.82 0.93 0.92
RI PT
Terschelling-Noordzee Huibertgat Borkum Fischerbalje Norderney-Riffgat Wangerooge-Nord Bake Z Cuxhaven Helgoland Wyk List Mean
TR obs (m) 1.74 1.97 2.09 2.16 2.50 2.55 2.65 2.08 2.43 1.56 2.17
SC
Station Name
Waves
As validation period for the wave model a time span from July 25, 2010 to August 20, 2010 was
7
chosen based on data availability and the presence of waves. The analysis is restricted to significant
8
wave heights as most relevant wave-related parameter for sediment transport modelling, the
9
evaluation of peak periods is not shown here. As can be seen from the comparison of observed and
10
simulated significant wave heights for station Fino-I in Figure 10 our wave model is able to roughly
11
capture the wave climate but cannot reproduce individual wave events reliably.
AC C
EP
TE D
M AN U
6
41
SC
RI PT
ACCEPTED MANUSCRIPT
M AN U
1
Figure 10: Simulated (red) and observed (blue) significant wave height (SWH) at station Fino-I
3
A statistical analysis for observed and modelled significant wave heights is given in Error!
4
Reference source not found.. The analysis is taking into account the frequency distribution of
5
significant wave height in 1 m classes, mean and maximum values as well as the 90th-percentile.
6
Table 6: Wave statistics comparing observed and simulated significant wave heights frequency distribution, mean, th
maximum and 90 -percentile analyzed for July 25, 2010 to August 20, 2010
Significant wave height distribution (%) Mean 0 m - 1 m 1 m - 2 m 2 m - 3 m 3 m - 4 m (m) 68.4 23.2 8.4 0.0 1.02 56.9 26.1 14.3 2.7 1.07 71.8 24.4 3.7 0.1 0.85 60.9 27.1 12.0 0.0 0.92 86.4 13.6 0.0 0.0 0.61 66.2 33.8 0.0 0.0 0.77
EP
7
TE D
2
9 10
AC C
8
FINO-I observed simulated Elbe observed simulated List observed simulated
Max (m) 2.73 3.48 3.16 2.81 1.67 1.82
P90 (m) 1.93 2.62 1.62 2.07 1.11 1.57
The mean wave height is simulated well within 10 % of the observed values for offshore locations, although for the nearshore location (station List) the difference is larger. Considering the
42
ACCEPTED MANUSCRIPT 1
90th-percentile and maximum wave height the model tends to overestimate wave heights. The same
2
tendency can be found in the frequency distribution.
3
1.5
4
Discussion Evaluating the model performance given above we find water levels sufficiently well reproduced
by the model for the scope of this study. Non-systematic differences in phase lag can probably
6
attributed to the not fully resolved coastline but do not indicate a general misrepresentation of
7
hydrodynamics. Most importantly the tidal energy is represented well; certain systematic errors do
8
exist (mean water level, phase lag) but are of minor importance for our sensitivity studies of
9
sediment transport.
SC
10
RI PT
5
The parametric wave model in the current set-up is only capable of providing a rough estimate of the actual wave climate. While the mean wave height is reproduced rather well the model tends to
12
overestimate wave heights. However, as the focus of this study is on estimating the effect of changes
13
in sediment resuspension due to the presence of macrofauna and the wave model provides the same
14
forcing for all sensitivity runs there will be no significant impact of the wave model imperfections on
15
our results.
TE D
16
M AN U
11
Taking into account the model´s resolution of about 0.5-1 nautical mile results are in parts surprisingly good. However, we can also clearly identify the need for model improvements for future
18
studies which are primarily related to the wave model and resolution and boundary conditions
19
(mean water level) of the hydrodynamic model.
AC C
20
EP
17
Overall the model performance can be considered sufficient for the scope of this study
21
particularly as we focus on a comparison between model experiments with and without macrofauna
22
effects based on the same hydrodynamic model.
43
S0272-7714(17)30098-7
DOI:
10.1016/j.ecss.2017.11.002
Reference:
YECSS 5659
To appear in:
Estuarine, Coastal and Shelf Science
Received Date: 27 January 2017 Revised Date:
23 October 2017
Accepted Date: 1 November 2017
Please cite this article as: Nasermoaddeli, M.H., Lemmen, C., Kösters, F., Stigge, G., Kerimoglu, O., Burchard, H., Klingbeil, K., Hofmeister, R., Kreus, M., Wirtz, K.W., A model study on the large-scale effect of macrofauna on the suspended sediment concentration in a shallow shelf sea, Estuarine, Coastal and Shelf Science (2017), doi: 10.1016/j.ecss.2017.11.002. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
A model study on the large-scale effect of macrofauna on the
2
suspended sediment concentration in a shallow shelf sea
3
Nasermoaddeli, M. H.a, Lemmen, C.b, Kösters, F.a, Stigge, G.c, Kerimoglu, O.b, Burchard, H.d, Klingbeil,
4
K.d,e, Hofmeister, R.b, Kreus, M. a, Wirtz, K. W.b
5
RI PT
1
a
Federal Waterways Engineering and Research Institute (BAW), Hamburg, Germany
7
b
Helmholtz-Zentrum Geesthacht Zentrum für Material- und Küstenforschung, Institute of Coastal
8
Research, Geesthacht, Germany
9
c
Institute for Applied Ecology (IfAÖ GmbH), Rostock, Germany
10
d
Leibniz Institute for Baltic Sea Research Warnemünde, Dept. for Physical Oceanography and
11
Instrumentation, Rostock, Germany
12
e
M AN U
Dept. of Mathematics, University of Hamburg, Germany
SC
6
13 Abstract
15
The activity of macrofauna on the sea floor is since long known to mediate deposition and erosion of
16
sediment, but so far most studies addressed this effect at a local scale. In the present paper, the
17
contribution of the observed macrofauna distribution (exemplified by a bivalve, the bean-like
18
Fabulina fabula, formerly known as Tellina fabula) on large-scale sediment transport in the southern
19
North Sea is investigated by means of a model study. Macrofauna effects are considered with respect
20
to the critical bed shear stress and erodibility, which are two important factors that control the
21
resuspension rate. Simulation results for a typical winter month revealed for the first time that the
22
suspended sediment concentration (SSC) is increased not only locally but beyond the inhabited
23
zones. This alteration is not confined to near-bed zones but can be observed throughout the entire
24
water column, especially during storm events. These effects are most prominent in the fine silt
25
fraction, coarser and finer fractions are less affected. For a selected storm event in February 2010,
26
we explain the counter-intuitive decrease in near-bed SSC in some areas with a high macrofauna
27
abundance compared to a simulation excluding such macrofauna: A high macrofauna-induced
28
entrainment rate leads to rapid exhaustion of available sediments at the bed in the model and
29
consequently limits the near-bed SSC.
AC C
EP
TE D
14
1
ACCEPTED MANUSCRIPT 1
Key words: macrofauna, suspended sediment, North Sea, Fabulina fabula, coupled modeling
2
1
3
Introduction Suspended sediment dynamics in coastal and shelf seas critically affects major biological
processes and has large economic implications (Puls et al., 1997, Winterwerp and van Kesteren 2004,
5
Gayer et al., 2006, van der Molen et al., 2009). The transport of suspended sediment contributes
6
significantly to material fluxes between the seabed and the water column and between rivers,
7
shallow water systems, and the open sea. It depends on physical factors such as tides and their
8
asymmetry (Burchard et al., 2008; Gräwe et al., 2016), wind and wave forcing (Lettmann et al., 2009),
9
and the morphological characteristics of the seafloor (van Ledden et al., 2004). Sediment transport is
SC
RI PT
4
also affected by biological factors such as microphytobenthos and macrofauna (Jumars and Nowell,
11
1984, Krumbein, 1994, Widdows et al., 1998, Orvain et al., 2006, Briggs et al., 2015, Harris et al.,
12
2016). Benthic organisms exert a considerable influence on sediment dynamics (a three- to sixfold
13
increase in the re-suspension rate as reported by Davis (1993)) since their activity is linked to a large
14
number of biological, geological, and physical factors. For example, macrobenthic animals graze on
15
stabilizing microphytobenthos (Orvain et al., 2014), alter the roughness of the seafloor (Peine et al.,
16
2005), or directly release particles into the water column (Graf and Rosenberg, 1997). In this study, we focus on two effects of macrobenthic communities on suspended sediments:
TE D
17
M AN U
10
changes in (1) the critical bed shear stress for erosion and (2) the erodibility of the bed sediment.
19
These two parameters, together with actual bottom shear stress, largely determine the erosion and
20
resuspension rates (Pleskachevsky et al., 2005 Stanev et al., 2009) and have been identified as
21
sensitive parameters in previous modeling studies (Knaapen et al., 2003, Lumborg et al., 2006, Le Hir
22
et al., 2007, van Prooijen et al., 2011, Orvain et al., 2012).
AC C
EP
18
23
Sediment erodibility of the few upper centimeters of the bed layer is modified (Andersen and
24
Pejrup, 2011), for example, by moving, sheltering and feeding macrofauna. These highly seasonal
25
processes lead to sediment displacement and are subsumed under the term bioturbation (Le Hir et
26
al., 2007). They also alter sediment aggregation (Nowell and Jumars, 1984), porosity and
27
permeability (Graf and Rosenberg, 1997), and the erodibility of the bed layer through
28
biostratification. For example, the vertical transport by upward conveyors (e.g. tube building
29
lugworm Arenicola marina) increases the fine fraction at the water-sediment interface. Conversely, 2
ACCEPTED MANUSCRIPT 1
downward conveyors (e.g. sipunculid worms) decrease the fine fraction (Wheatcroft and Butman,
2
1997).
3
Burrowing and grazing activities of the deposit feeder Fabulina fabula (F. fabula) - formerly known as Tellina fabula - for example, increase the erodibility of surface sediments (Austen et al.,
5
1999, Borsje et al., 2009). Sediment erodibility can be further altered by the aggregation of bulk
6
sediment into fecal pellets or pseudo-feces by deposit feeders. This bio-aggregation increases
7
erodibility since the aggregated material is less cohesive than before consumption (Austen et al.,
8
1999). Fecal pellets are also more prone to erosion due to the reduction in specific surface area, i.e.
9
the ratio of surface area to mass (Andersen and Pejrup, 2011).
SC
RI PT
4
The critical bed shear stress of the uppermost bed layer is also reduced by the generation of a
11
fluff layer, which may be formed by snail tracks, fecal pellets and disintegrated sediment particles
12
due to deposit feeding activities (Orvain et al., 2003). Particles from this layer are brought into
13
suspension far below the critical bed shear stress of the underlying sediment ("fluff layer erosion")
14
(Shimeta et al., 2002).
M AN U
10
Research on suspended sediment dynamics has started to include macrobenthic effects,
16
specifically on erodibility, in both observational and modeling studies. Most of these past studies,
17
however, focused on local effects of benthic fauna (benthos) on sediment transport or were
18
restricted to smaller, e.g. intertidal, scales (Paarlberg et al., 2005; Borsje et al., 2007; Le Hir et al.,
19
2007; Sanford, 2008). To date, there have been very few studies of the large-scale effects of benthic
20
fauna on suspended sediment dynamics (e. g. in Seifert et al. (2009) for the Baltic Sea).
EP
21
TE D
15
In this paper, we quantify local effects of macrobenthos on sediment transport and study how these impact the large-scale distribution of suspended sediments by means of a model study. In
23
particular, the present study aims to yield model based sensitivities in view of the following research
24
questions: (1) To what extent and magnitude do macrobenthic effects on critical bed shear stress and
25
erodibility influence the large-scale near-bed suspended sediment distribution, i.e. sediment
26
concentration patterns, in a shallow shelf-sea (southern North Sea)? (2) Are the effects similar for
27
different sediment classes? (3) Do effects differ for sediment transport near the bed and in the whole
28
water column? (4) Are the impacts of macrofauna limited to events or characteristic for longer time
29
periods?
AC C
22
3
ACCEPTED MANUSCRIPT 1
We address these questions based on a coupled model system taking into account 3D hydrodynamics (tides, baroclinic flow), a simplified wave model and multi-fractioned sediments thus
3
resolving the most relevant processes for coastal and shelf sediment transport. In addition this
4
coupled system includes effects of macrofauna by using the observed spatial distribution of the
5
bivalve F. fabula, which is a dominant species and eponymous characteristic of a macrozoobenthos
6
community of the southern North Sea. We then compare the simulated surface suspended sediment
7
concentration (SSC) with satellite observations and in-situ measurements and analyze the simulation
8
results for a reference scenario including mentioned processes and another scenario that disregards
9
the macrobenthic effects.
11
2.1
12
SC
2
Materials and methods
M AN U
10
RI PT
2
Study Area
The study area comprises the southern North Sea, a shallow shelf system with water depths of up to 50 m, intertidal flats and a number of estuaries at the coastline (Figure 1). Currents are
14
predominantly influenced by the semi-diurnal tides that propagate through the German Bight in an
15
anti-clockwise direction, with amplifications in the estuaries and complex bathymetric interactions
16
(e.g. Becker et al., 1992; Stanev et al., 2014). Substantial quarter-diurnal tidal components exhibiting
17
a significant seasonal signal (Gräwe et al., 2016) are generated (Stanev et al., 2014). The tidal range
18
varies from 2 m near the Dutch coast to about 4 m in the German estuaries. Tidal mixing inhibits
19
thermal and haline stratification in the shallower parts of the German Bight (Schrum, 1997). Coastal
20
upwelling and downwelling due to wind forcing affect the pathways of the discharge from the Elbe,
21
Weser and Ems rivers (Krause et al., 1986). Stratification only occurs seasonally in the deeper parts of
22
the southern North Sea and occasionally near the estuaries (van Leeuwen et al., 2013). At the fronts
23
between the generally vertically well mixed German Bight waters and the stratified North Sea waters,
24
a so-called line of no return (Postma, 1984) defines an area within which sediments are trapped in
25
the German Bight. Between this line of no return and the coast, a thermohaline overturning
26
circulation transports sediments towards the coast such that sediments accumulate in the Wadden
27
Sea (Burchard et al., 2008) and contribute to the formation of estuarine turbidity maxima in the
28
estuaries (e. g. Kappenberg and Grabemann, 2001).
AC C
EP
TE D
13
4
SC
RI PT
ACCEPTED MANUSCRIPT
2 3
M AN U
1
Figure 1: Bathymetry of the southern North Sea as represented on the numerical model grid. The study area of the German Bight is depicted by the black rectangle.
4
Macrobenthic infauna of the German EEZ has classically been categorized as belonging to one of five main benthic infauna communities based on preferred habitat conditions (depth, sediment
6
type), namely, Amphiura filifornis, Tellina (Fabulina) fabula, Nucula nitidosa, Goniadella-Spisula, Spio
7
filicornis (Salzwedel et al., 1985). The categorization has been roughly confirmed by the ICES North
8
Sea Benthos Survey (Duineveld et al., 1991) and the North Sea Benthos Project 2000 (Kröncke et al.,
9
2011) using a grid of sampling points, and by more refined cluster analysis (Neumann et al., 2012).
EP
TE D
5
10
2.2
11
2.2.1 Sedimentological data
AC C
12
Observational data
Sedimentological input data is based on measurements of the Federal Maritime and
13
Hydrographic Agency (BSH). These and other data sources have been compiled and are available as a
14
habitat atlas published by the ‘North Sea Observation and Assessment of Habitats (NOAH)’ project
15
(http://www.noah-project.de).
16 17
Near-surface in-situ SSC time series in the area of West Frisia provided by the Dutch authority Rijkswaterstaat was accessed through OpenEarth (Rijkswaterstaat, 2017). The two cross-shore 5
ACCEPTED MANUSCRIPT 1
transects, Terschelling and Rottumerplate (Figure 3), were selected because of their close proximity
2
to the East Anglian plume and data availability for the period of study.
3
Suspended sediment concentrations close to the water surface can be inferred from Envisat satellite images provided by the European Space Agency (ESA) at a 300 m resolution. The satellite
5
images were pre- and post-processed by ESA and Brockmann Consult using MERIS regional case 2
6
water algorithms (C2R) explained in Doerffer et al. (2006). It should be noted that the calculated
7
concentrations represent the total suspended matter in the near-surface layers, which also includes
8
organic matter not considered in the simulated SSC. The penetration depth of the light emitted from
9
the satellite-borne MERIS sensor depends on the turbidity of the water, for example at a depth of 1 m in the case of high turbidity in our area of interest.
11
2.2.2 Macrofauna data
We accessed the macrozoobenthos database of the Institute for Applied Ecology (IfAÖ GmbH),
M AN U
12
SC
10
RI PT
4
which included 17,710 hauls at 2,347 infauna sampling stations and 2,798 hauls at 1,049 epifauna
14
sampling stations in the German Bight by the time of the data request. Based on an analysis of
15
species abundances, biomasses and presences derived from this database and by means of a
16
literature review, we selected species that are particularly important for sediment transport
17
processes. Due to its bioturbation potential the bivalve F. fabula is one of the most distinctive species
18
in the study area (Borsje et al., 2009), which is also reflected by the naming of the F. fabula
19
community occurring in the German Bight (Rachor and Nehmer, 2003). The distribution of F. fabula
20
in the southern North Sea is based on the data by Creutzberg (1986) as well as the occurrence
21
probability by Reiss et al. (2011), converted to abundance by field data compiled by van Moorsel
22
(2011) and supplemented for the German Bight by data of Dannheim (2014) as shown in Figure 2.
23
The comparison of sedimentological data, in this case the median grain size d50 in mm, with the
24
abundance of F. fabula illustrates that this species not only appears in large numbers in sandy
25
sediments between 0.125 mm and 0.250 mm but also in regions with coarser or finer median grain
26
sizes.
AC C
EP
TE D
13
6
1 2
-
Figure 2: Estimated mean abundance of the bivalve F. fabula in the southern North Sea (color scale, individuals m ²); contours denote median grain size (mm).
4
2.3
Numerical models
TE D
3
5
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
To investigate large-scale macrofauna effects on the suspended sediment transport in the southern North Sea, a generic benthos module (Nasermoaddeli et al., 2014) was coupled to
7
hydrodynamics, sediment transport, and erosion—sedimentation models via the Modular System for
8
Shelves and Coasts (MOSSCO, Lemmen et al., 2017) coupling framework. The MOSSCO framework is
9
built on and extends the Earth System Modeling Framework (ESMF, Hill et al., 2004) and enables
AC C
10
EP
6
flexible and relatively seamless coupling of a data and model components.
11
In our application, MOSSCO includes five model components (hydrodynamics, suspended
12
sediment, erosion—sedimentation, waves and benthos effect) and generic input components that
13
provide data on ocean boundary particle concentration, river loads and macrofauna distribution.
14
2.3.1 Hydrodynamic model
7
ACCEPTED MANUSCRIPT 1
Hydrodynamics are simulated by the General Estuarine Transport Model (GETM; Burchard and Bolding, 2002). GETM is an open source coastal ocean model which has been shown to be of high
3
skill in various studies for the North Sea and Wadden Sea (Stips et al., 2004; Burchard et al., 2008;
4
Lettmann et al., 2009; van Leeuwen et al., 2013; Gräwe et al., 2016). It solves the incompressible
5
Navier-Stokes-Equations and additional prognostic equations for temperature and salinity (Klingbeil
6
and Burchard, 2013). The free surface is calculated efficiently in an explicit mode-splitting algorithm
7
and supports drying-and-flooding of intertidal flats. Quantities are transported with a high-order
8
advection scheme (e. g. Klingbeil et al., 2014). Horizontal sub-grid scale dynamics are parameterized
9
by a Smagorinsky closure (Smagorinsky, 1963). Turbulent vertical viscosities and diffusivities are
RI PT
2
provided by the General Ocean Turbulence Model (GOTM) which offers state-of-the-art turbulence
11
closure from Umlauf and Burchard (2005). For the present study, a dynamic k-ε model with the 2nd-
12
order closure of Canuto et al. (2001) “Model A” was applied. Bottom stresses are calculated on the
13
basis of the law of the wall with a spatially uniform roughness length of z0=1 mm, which was chosen
14
as a typical roughness value for the southern North Sea (e.g. Gräwe et al., 2016). Note that
15
roughness is a calibration factor for the model. The roughness length z0 chosen here relates to a
16
reasonable ripple height of about 30 mm. Spatial variations of the grain related roughness can be
17
neglected as typical sediment grain size diameters for the German Bight are in the order of 0.3 – 0.5
18
mm (medium sand), thus accounting for only 1% of total roughness. Variations in form roughness
19
have been also omitted in order to facilitate the interpretation of the results.
TE D
M AN U
SC
10
In order to include the effect of wind waves, local wave conditions (significant wave height, peak
21
period, wave direction) were parameterized in terms of wind speed, wind direction and water depth
22
according to Breugem and Holthuijsen (2007). These current-only bottom stresses are modified
23
according to Soulsby (1997) to consider the combined wave—current stresses.
24
2.3.2 Sediment transport model
AC C
25
EP
20
Suspended sediment is represented here by three size classes. Sediment mass is transported
26
along the current vector computed by the hydrodynamic model, while diffusion due to turbulence
27
and downward movement with a constant, size-dependent sinking velocity are applied additionally in
28
the vertical. For the transport of the SSC, the same high-order advection scheme is used as for the
29
transport of salinity and temperature in the hydrodynamic model. At open sea and river boundaries,
30
a constant sediment concentration is prescribed as proposed by Gayer et al. (2006). At the water-
8
ACCEPTED MANUSCRIPT 1
sediment interface, the net mass flux for each suspended sediment size class is prescribed as the flux
2
boundary condition for the sediment transport model.
3
The modularity within MOSSCO allows the coupling of a separate model for the calculation of the net sediment flux at the water—sediment interface. For this purpose, the erosion—sedimentation
5
routines of the Deltares Delft3d model were encapsulated and coupled via an ESMF interface to
6
MOSSCO (Nasermoaddeli et al., 2014). The model uses the well-known Partheniades-Krone equation
7
(Partheniades, 1965) extended for a sand-mud mixture to calculate the net sediment flux of cohesive
8
sediment at the water-sediment interface (see Table 1). A similar approach was proposed by van
9
Ledden (2004) or Le Hir et al. (2011) and was applied in Paarlberg et al. (2005).
= ,
12
,
,
∙
∙
∙
= =
,
,
,
,
∙ 1−
,
−1 ,
>
,
<
0,
,
,
≤
Eq. 1 Eq. 2 Eq. 3
,
,
Eq. 4
0,
≥
,
EP
,
11
SC
= ∙
M AN U
Table 1: Erosion and deposition fluxes as implemented in the numerical model according to Partheniades (1965)
TE D
10
RI PT
4
The main parameters taken into account are the erosion flux E (kg m-2 s-1) and the deposition flux D (kg m-2 s-1). In the equations, l is the index of sediment class, g the biological destabilization factor
14
for erodibility (-), M the erodibility factor (kg m-2 s-1), S the Heavy-Side function, ws the settling
15
velocity for each sediment class l, cb the near bed suspended sediment concentration, τ the bed
16
shear stress, τcr,er the critical bed shear stress for erosion, f the biological destabilization factor for
17
critical bed shear stress for erosion, and τcr,depr the critical bed shear stress for deposition. The factors
18
f and g are introduced into the original formula here to account for the macrofauna effects
19
presented in the following section.
AC C
13
9
ACCEPTED MANUSCRIPT The net vertical sediment flux for each class results from the difference between erosion and
2
deposition, both of which are applied to the active single layer of the bed. In this approach, the bed is
3
conceptualized as a well-mixed, non-stratified layer with a given depth. An arbitrary number of
4
sediment size classes can be considered within this highly idealized sediment inventory. The mass
5
fraction of each sediment size class within the well-mixed layer is the ratio of its mass to the total
6
mass of all fractions within the layer. The mass of each sediment size class for a unit area is initialized
7
in the model by a given fraction, thickness, porosity and sediment dry density which is spatially
8
constant. The deposited mass of each sediment size class is added to the rest of the mass of the
9
same class within the layer and mixed instantaneously with the bed material throughout the whole
RI PT
1
layer thickness. The mass fraction of the size class is altered accordingly for the whole depth. The
11
erosion of bed material for each sediment size class is calculated in a similar manner.
12
2.3.3 Benthos model
M AN U
13
SC
10
The inclusion of biotic effects on sediment transport modeling is complicated partly because of the lack of data and knowledge so that it is currently not possible to quantify the effects of
15
macrofauna on sediment transport, especially at the community level. Further complicacy stems
16
from the seasonal variation of biological components, spatial patchiness, and non-linear interactions
17
within the benthic community (Le Hir et al., 2007). In addition, effects on sediment transport
18
properties are highly specific and are thus dependent on the function of the considered benthic
19
species, while collective effects of different species on sediments are difficult to predict due to
20
reciprocal (antagonistic) influences in benthic communities (Kristensen et al., 2013). Currently available models of sediment transport only depict in a very simplified manner the high
EP
21
TE D
14
complexity of effects that macrozoobenthos in their entirety exert on the water-sediment boundary
23
layer. Another issue is the lack of species- or community-specific parameterizations of macrofauna
24
effects on sediment transport parameters. Most sediment transport models that resolve
25
macrozoobenthos have so far only been parameterized for few selected macrofauna species and
26
with few affected sediment transport parameters (e.g. Wood and Widdows, 2002; Knaapen et al.,
27
2003; Paarlberg et al., 2005) or they quantify such effects using process-based differential equations
28
(e.g. François et al., 1997; Orvain et al., 2003; Orvain, 2005; Montserrat Trotsenburg, 2011; Orvain et
29
al., 2012).
AC C
22
10
ACCEPTED MANUSCRIPT Due to the lack of suitable parameterizations for F. fabula, we applied parameterizations which
2
were originally derived for the effect of Macoma balthica on critical bed shear stress and erodibility
3
(Knaapen et al., 2003; Paarlberg et al., 2005). This approach was justified on the basis of a functional
4
classification of a total of 153 macrozoobenthos species from the German Bight according to a trait
5
matrix. Each species was numerically characterized according to seven traits (motility, nutritional
6
habits, habitat structure, sediment transport, size, shape and position in the sediment) as proposed
7
by Darr et al. (2014). Based on the assigned traits, a similarity analysis using a fuzzy coding approach
8
(Chevenet et al., 1994) allowed an aggregated description of macrofauna effects on sediment
9
transport at the community level by clustering species with similar autecology. F. fabula was found to
10
be the species most similar to Macoma balthica, following bivalves Abra alba and Astarte montagui.
11
Our analysis of macrobenthic communities based on the trait matrix by Darr et al. (2014) permitted a
12
generalization of a species specific parametrization. Underlying (laboratory) studies are usually
13
performed using single model species and now can be extrapolated to other dominant species in the
14
region of interest. Given the high number of macrobenthic species in coastal and shelf habitats, we
15
advocate the further development of trait-based approaches, also to permit a reduction in model
16
complexity and parameterization effort.
M AN U
SC
RI PT
1
These two formulations are used as linear modification factors of the corresponding abiotic
18
parameters. It should be noted that both relations rely only on limited data collected by Widdows et
19
al. (2000a, b) in an intertidal basin. The destabilization factors for critical bed shear stress and
20
erodibility are given in Table 2.
Table 2: Factors accounting for destabilization of critical bed shear stress (fd) and erodibility (gd) # $M&
22
&=
+&
− 0.085 ∙ ln$ & + 1
/+ ∙ 0 $/+ + 0 ∙ /12 & 3
AC C
#$
= 0.0016 ∙ ln$
EP
21
TE D
17
Eq. 5 Eq. 6
23
fd is the destabilizing factor for critical bed shear stress, M is the dimensionless abundance
24
(actual abundance of individuals (ind.) divided by the reference abundance of 1 ind. m-2), and gd the
25
destabilizing factor for erodibility. The derivation of gd is based on the assumption that the
26
macrofauna effect reaches a maximum with an increasing abundance M, after which it remains 11
ACCEPTED MANUSCRIPT 1
constant (sigmoid function). The parameters applied here are the maximum biological erosion
2
coefficient γ = 6 x 10-7 ms-1, the erosion coefficient without macrofauna influence I = 4.68 × 10-8 ms-1
3
and the regression coefficients b1 = 0.995 and b2 = 5.08 10-8 ms-1.
4
The aforementioned parameterizations were implemented in MOSSCO’s benthos module, which is a generic object-oriented library. It provides an adaptive structure for linear superposition of
6
macrofauna effects on sediment transport parameters such as critical shear stress, erodibility,
7
roughness, settling velocity, biodeposition and bio-resuspension. Currently, only the first two
8
parameterizations are included in the benthos model. In the following, this parameterization is
9
referred to as the macrofauna effect but the reader should bear in mind that this is based on a rather
RI PT
5
simplistic approach which can only give an indication of the complex interaction between
11
macrofauna, or biology in general, and sediment dynamics.
12
2.4
M AN U
13
Model setup
SC
10
The model was set-up in a way to represent the dynamics of fine sediments in the German Bight.
14
We focus on the topmost layer of easily erodible material, i.e. a layer of stationary suspension (e.g.
15
Hayter and Mehta, 1986). Thus large-scale sand transport and associated morphological changes are
16
not considered here.
The model resolves three silt classes, which differ in settling velocity and transport parameters,
TE D
17 18
similar to Gayer et al. (2006) and in accordance with the measurements of Puls et al. (1995) in the
19
German Bight. Other sediment transport parameters for each silt class are presented in Table 3. The initial fraction of each silt class at the bed and in the water column was assumed to be
21
spatially constant in the computational domain. The silt classes were determined by multiplying silt
22
fractions measured in the shallow water (0-20 m) zone by the ratio of the area of the shallow zone to
23
the total area of the southern North Sea. The silt fractions at open sea boundaries and rivers were
24
selected according to Gayer et al. (2006).
AC C
EP
20
25
The amount of mobile sediments in nature is not only spatially variable but also depends on the
26
vertical structure. In spite of intensive efforts to map surface sediments and their properties for the
27
German Bight (e.g. Arfai et al., 2014) information for the top few centimeters is not available yet.
28
Therefore the initial thickness of the active layer had to be selected and was set at 10 mm. This is in
29
line with comparable studies such as Lumborg et al. (2006) who used 2-5 mm obtained from field 12
ACCEPTED MANUSCRIPT 1
observations for the upper bed layer in Kongsmark, a tidal mudflat in the Danish Wadden Sea.
2
Limiting the erodible layer to a fixed value of 10 mm reflects, in a simplified manner, the limitation of
3
erosion in nature due, for example, to the consolidation or presence of coarse sand in the soil matrix
4
which may prevent the removal of fines by armoring. Critical bed shear stress for erosion (τc,er ) can generally be regarded as a calibration parameter; it
6
was adjusted in order to obtain a comparable structure of simulated SSC with available satellite data
7
and in-situ measurements. We found the calibration values of Gayer et al. (2006) to be a good
8
starting point but adjusted the critical bed shear stress for fine silt to 0.6 N m-2.
The existence of a critical bed shear stress for deposition (τc,dep ) is under debate. Winterwerp
SC
9
RI PT
5
(2007) showed the irrelevance of τc,dep by successfully reproducing experimental results without
11
including this parameter in his model. In contrast, Maa et al. (2008) argue, on the basis of their
12
laboratory experiments, that deposition ceases beyond a specific bed shear stress. We conducted a
13
sensitivity analysis by testing a range of values of τc,dep for each silt class varying between 0.096 N m-2
14
to 2 N m-2, as well as ignoring this parameter. The selected range reflects literature values (i.e. Gayer
15
et al., 2006, and Lumborg et al., 2006). The best fit between simulated and measured SSC was
16
reached by the values of τc,dep given in Table 3. The erodibility, which was constant for the three silt
17
fractions, was also calibrated.
TE D
Table 3: Sediment parameters of the numerical model (SNS setup)
Settling velocity
Initial
Initial
Fraction*
concentration
(m s )
(%)
(g m )
(m)
0.001
8.24
2.5
0.0001
5.46
0.00002
1.21
-1
Coarse silt
Initial
EP
Silt class
AC C
18
M AN U
10
τc,dep -
Erodibility -
-2 -1
(N m ²)
(N m ²)
(kg m s )
0.01
0.78
-
1.0×10-5
2.5
0.01
0.6
1.5
1.0×10-5
0.5
0.01
0.1
1.5
1.0×10-5
-3
thickness
τc,er
(0.031 -0.062 mm) Fine silt
(0.008-0.016 mm) Very fine silt (0.004 – 0.008
13
ACCEPTED MANUSCRIPT mm)
1 2
*fraction in respect of the total mass including all other sediment size classes (sediment classes according to Wentworth, 1922) The model domain covers large parts of the southern North Sea (SNS). We use a curvilinear
4
horizontal grid with 140x100 grid points in the horizontal to align the main axes with the coastline
5
and to increase the resolution in the German Bight up to 1.5 km at the south-east corner and 4.5 km
6
at the north-west corner (see Figure 1 for the model grid). In the vertical, 20 sigma layers are used,
7
resulting in a vertical resolution of 0.5 m at a water depth of 10 m and 2.0 m at a water depth of
8
40 m. This strategy has been shown to work successfully in hindcast simulations with GETM (e. g.
9
Hofmeister et al., 2013).
The numerical time step, as well as a coupling time step between the components, was set at 2
M AN U
10
SC
RI PT
3
minutes. The setup is forced supplied with output data from a regional atmospheric model at the sea
12
surface and the sea surface height at the open boundaries is taken from a barotropic hindcast
13
simulation. A more detailed description of the setup and its validation is given in Kerimoglu et al.
14
(2017). The effect of bottom friction is implemented as temporally and horizontally constant. The
15
bathymetry of the model domain and model grid is shown in Figure 1. Ten major rivers along the
16
coasts on the southern North Sea were considered with respect to their freshwater fluxes and the
17
suspended sediment loads (Radach and Pätsch (2007), Gayer et al. (2006)). We prescribe the
18
suspended sediment concentration at the northern and western open boundaries of the model
19
domain using estimates of total SSC by Heath et al. (2002).
EP
TE D
11
20
The simulation period from January to December 2010 covered both winter as well as summer months (for qualitative assessments) but also matches the availability of validation data for
22
hydrodynamic and partly for suspended sediment. For the fully coupled simulation of sediment
23
transport including macrobenthic effects, the observed distribution of the bivalve F. fabula was
24
implemented as external forcing and implemented in a modular manner. The configuration used for
25
this study can be reproduced by downloading the open source MOSSCO setups repository from
26
https://sf.net/p/mossco/setups , selecting the „sns“ example setup with the „spm“ configuration.
27
3
AC C
21
Results 14
ACCEPTED MANUSCRIPT 1
3.1
Model validation Observed water-levels are reproduced sufficiently well, e.g. the difference in tidal range, as a
3
measure of energy input by the tidal wave, between observed and modelled values is in the order of
4
5 % in the focus domain. Further details are given in the Appendix. The hydrodynamic model also
5
provides a realistic description of horizontal mixing as inferred from domain-scale salinity gradients
6
and vertical stratification from May to September within the deeper (> ~30 m) regions of the German
7
Bight (Kerimoglu et al., 2017).
8 9
RI PT
2
The simulation of waves is less satisfying as observed values are overestimated by the parametric wave model. However, for the scope of this study our rather simple approach provides a sufficient estimate for temporary wave induced excess bed shear stress in addition to the periodic tide induced
11
bed shear stress. Further details are given also in the Appendix.
For the SSC validation, two satellite images recorded on March 1, and April 14, 2010 were
M AN U
12
SC
10
13
selected (Figure 3 a, c). The former is representative of a storm event and the latter of a typical neap
14
tide. Both images display the typically observed coastal gradient (strongly decreasing offshore
15
concentration) as well as the East Anglian plume.
16
Their comparison with simulated near-surface SSC shows a good agreement in terms of reproducing the coastal gradients in the eastern part of the German Bight; but SSC is overestimated
18
at the Elbe and Weser estuaries and underestimated at the Ems estuary, as estuarine dynamics are -
19
due to the relatively coarse model resolution of 0.5 to 1 nautical mile - only resolved as climatological
20
boundaries. It was possible to reproduce the East Anglian plume on March 1 fairly well (Figure 3 a, b).
21
It is, however, missing in the simulation results on April 14 (Figure 3 c, d). It should be noted that the
22
satellite images show the total suspended matter while the simulations only show the suspended
23
sediment concentration.
EP
AC C
24
TE D
17
15
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
1
Figure 3: Snapshots of total suspended matter (TSM) calculated from satellite images on March 1, 2010 (a) and April
TE D
2 3
14, 2010 (c) vs simulated suspended sediment concentration (SSC) near-surface at the same days (b and d)
The sediment transport model was further validated using the time series from satellite images
5
as well as in-situ near-surface SSC at several locations nearshore and offshore of West Frisia. Near-
6
surface simulated SSC with and without macrofauna compare relatively well to satellite derived data
7
(Figure 4). At least the order of magnitude of observations and the simulated near-surface SSC agrees
8
in near shore and offshore regions of West and North Frisia. It was possible to reproduce the
9
temporal variation in SSC well in some periods and locations (for example at Terschelling T50 and
AC C
EP
4
10
Rottumerplate R70, Figure 4 b, c), but SSC was overestimated at 10 km off Terschelling (T10, Figure
11
4 a). The simulated SSC scenario without macrofauna performed better at T10, which could be due to
12
the overestimation of F. fabula abundance by the interpolation at this location.
16
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
1
8 9
simulations (green lines) of concentration of suspended material for coastal transects represented by stations T10 and T50 (10 and 50 km off the West Frisian island of Terschelling), R3 and R70 (3 and 70 km off the island of Rottumerplate outside the Dollart estuary, and S9 and S62 (9 and 62 km off the North Frisian island of Sylt); The simulation was sampled in a 2500 m radius around the station and the satellite (with higher resolution than the simulation) in a 500 m radius.
3.2
TE D
7
Figure 4: Comparison between in-situ measurements (blue diamond), remote sensing observations (red crosses) and
Simulations with macrofauna effects
In the following, the results are presented for a storm event as well as the temporal mean effect of macrofauna on SSC. Simulation results for SSC including macrofauna effects (i.e. using the
EP
2 3 4 5 6
observed spatial distribution of an average abundance of F. fabula) reveal a number of differences
11
between the scenarios with and without macrofauna.
12
3.2.1 Storm event
13
AC C
10
Simulated near-bed SSC with macrofauna differs from the SSC in the scenario without
14
macrofauna, depending on the sediment class, macrofauna abundance, and location. The near-bed
15
SSC of the coarse silt class during a storm event (Figure 5 a) is distributed over a large area in the
16
shallow zone and partly offshore. This patchiness seems to be the imprint of F. fabula inhabitation
17
zones when compared to the contour lines of F. fabula abundance in Figure 5 b. Additionally,
18
macrofauna in the model has led not only to an increase in the near-bed SSC in some inhabited areas 17
ACCEPTED MANUSCRIPT but also to a decrease in other areas. Zones with reduced near-bed SSC correspond mostly to
2
inhabited zones with a higher abundance of F. fabula (shown by thicker contour lines). While few
3
individuals may increase the erosion rate many individuals may induce biodeposition, as found by
4
Friedrichs et al. (2009). Increases in near-bed SSC due to macrofauna effects were more frequent
5
than reductions but the absolute magnitude of 95 percentiles ( 1.65 g m-3) was less than the absolute
6
magnitude of (p5 =-2.56 g m-3) as shown in Figure 5 b, reflecting the macrofauna-induced increase
7
and decrease in the near-bed SSC.
8 9
RI PT
1
A much lower degree of patchiness is observed in the near-bed SSC distribution of the fine silt class (Figure 5 c). Figure 5 d shows a significant macrofauna-related increase in SSC, primarily off the coast of Sylt and also in some parts of the shallow waters along East and West Frisia. These zones
11
correspond to the inhabited regions with F. fabula (shown by contour lines). A closer comparison of
12
macrofauna-enhanced SSC zones with F. fabula abundance contours in Figure 5 d shows that near-
13
bed SSC has been modified in the model results beyond the zones inhabited by macrofauna, altering
14
sediment erosion and deposition properties. This is particularly obvious in the shallow zones such as
15
north of Sylt or offshore of the Ems estuary, which suggests that the contribution of macrofauna to
16
sediment transport may affect large-scale patterns of near-bed SSC of fine silt.
M AN U
17
SC
10
The near-bed SSC of fine silt is also reduced by macrofauna in a number of regions (Figure 5 d). The pattern of near-bed SSC of fine and coarse silt is similar, with the former exhibiting greater
19
magnitudes of change and changes over larger areas. The near-bed SSC of fine silt has been
20
significantly reduced offshore of West Frisia and south-east of Helgoland. Similar to coarse silt
21
macrofauna-induced increase of SSC can be observed to the west and north of Sylt as well as
22
offshore of East Frisia. Seaward of West Frisia, however, macrofauna activities have resulted in a
23
decrease in the near-bed SSC of fine silt and coarse silt. Finally, a significant increase in differences
24
due to macrofauna effects (p95 1.65-7.86 g m-3) can be inferred from Figure 5 b and d.
EP
AC C
25
TE D
18
No patchiness can be detected from near-bed SSC distribution for very fine silt in Figure 5 e. The
26
modified near-bed SSC due to macrofauna for very fine silt, depicted in Figure 5 f, shows a very slight
27
difference (almost negligible) and mostly a reduced concentration in coastal shallow areas,
28
irrespective of the F. fabula distribution. It is inferred that the macrofauna effect does not play any
29
significant role for near-bed SSC for very fine silt during a storm event, in contrast to the other two
30
silt classes. 18
ACCEPTED MANUSCRIPT 1
The concentration of near-bed total silt (sum of all three silt classes) shown in Figure 5 g represents similar gradients and magnitude to those of the fine silt fraction. Patchiness of the near-
3
bed SSC offshore along the coasts and the SSC plume west and south west of Helgoland can still be
4
observed. The contribution of macrofauna to the near-bed SSC (Figure 5 h) is quite similar for near-
5
bed total silt and fine silt class. This implies that the predominant macrofauna effects on the near-
6
bed SSC may be represented by fine silt fractions in the case of a storm event. However, it should be
7
noted that the magnitude of enhancement of near-bed SSC due to macrofauna decreases when
8
integrating the concentration over all silt fractions since in some regions the concentration increased
9
for fine silt, but decreased for coarse silt, such as to the west of Sylt.
AC C
EP
TE D
M AN U
SC
RI PT
2
19
1 2 3 4 5
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Figure 5: Simulated near-bed suspended sediment concentration (SSC) during a storm event in February 2010. The left-hand panels show the near-bed SSC for coarse (a), fine (b), very fine (e) and total (g) SSC; the right-hand panels (b, d, f, h) show the calculated modification due to benthos effects. Contours denote the presence of F. fabula at 10 (thin line), -2
100, and 500 (thick line) individuals m . Summary statistics for n grid cells shown as mean µ and 5-95 percentiles p5,95
20
ACCEPTED MANUSCRIPT 1 2
3.2.2 Storm event, depth-averaged SSC Depth-averaged SSC is especially important for both ecosystem modeling and sediment management and was therefore subject to further investigations. As results for all three silt classes
4
were quite similar to those for near-bed SSC, we present only the total depth-averaged SSC below.
5
Model results of depth averaged SSC shown in Figure 6 are temporal means over a spring-neap-cycle
6
after one month of model spin-up time.
RI PT
3
7
The total depth-averaged SSC in Figure 6 exhibits similar patterns to the total near-bed SSC,
8
except for a weaker patchiness, which is a consequence of the vertical and horizontal mixing of SSC.
9
The main difference seems to be the plume around Helgoland, which lacks in total depth-averaged SSC and thus appears to be specific to the near-bed layer. The plume may hence be due to local
11
resuspension of silt near the bed, usually deposited south-east of Helgoland (Puls et al., 1999). The
12
horizontal distribution of the contribution of macrofauna to the depth-averaged SSC shown in Figure
13
6 b is similar to that of near-bed SSC depicted in Figure 5 h. Thus biological activities of macrofauna
14
as parameterized in the model impact not only the concentration at the sea floor but throughout the
15
water column. Both model results (near bed and depth averaged SSC) show a similar pattern of SSC
16
reduction where macrofauna is present (bluish colors in Figures 5 and 6) and increase adjacent to
17
these regions (reddish colors in Figures 5 and 6). Thus the modelled macrofauna effect extends
18
beyond the inhabitation zone.
AC C
EP
TE D
M AN U
SC
10
21
ACCEPTED MANUSCRIPT
M AN U
SC
RI PT
1
10
spring-neap-cycle (including a storm event) in February 2010 (Jan 31-Feb 13, shown as inset, upper panels a, b) and at the storm event in February 2010 (Feb 3, 2:00 am, location in tidal cycle marked with a red dot, lower panels c, d). The lefthand panels (a, c) show the concentration; the right-hand panels (b, d) show the included contribution of biotic benthic -2
modification. Contours denote the presence of F. fabula at 10 (thin line), 100, and 500 (thick line) individuals m .
EP
9
Figure 6: Simulated depth-averaged total suspended sediment concentration (SSC) temporally averaged over a
Summary statistics for n grid cells shown as mean µ and 5-95 percentiles p5,95.
3.2.3 Temporal mean macrofauna effects
AC C
3 4 5 6 7 8
TE D
2
For a spring-neap tidal cycle during a typical winter month (February 2010) including a storm
11
event, the patchiness of total mean depth-averaged SSC disappears (Figure 6 a), in contrast to the
12
storm event (Figure 6 c). This may indicate strong horizontal mixing of macrofauna-induced changes
13
in SSC during a spring-neap tidal cycle. Furthermore, it can be inferred from Figure 6 b that mean
14
depth-averaged SSC is generally enhanced by macrofauna in most areas of the southern North Sea,
15
also in contrast to the storm event. Specifically, contrary to the storm event, the SSC is increased
16
offshore of West Frisia due to a macrofauna contribution. The only area with slightly decreased SSC is 22
ACCEPTED MANUSCRIPT 1
south-east of Helgoland, i.e. the mouth of the Elbe estuary. The SSC has been increased due to
2
macrofauna mostly along the coastal belt and the effect is highest around the East Frisian Islands and
3
Sylt. The magnitude of increased SSC due to macrofauna effects is still significant (up to 5.24 g m-3),
4
but the maximum reduction of SSC reaches only -1 g m-3. These model results suggest that macrofauna affects the spatio-temporal distribution of SSC
6
beyond the inhabited areas, thus inducing a large-scale impact on SSC in the southern North Sea
7
regardless of physical condition.
8
4
RI PT
5
SC
Discussion
The modelled suspended sediment concentration increases due to macrofauna effects (Davis,
10
1993, Le Hir et al., 2007) beyond the inhabitation zones in the case of the fine silt class but not for
11
coarse silt (Figure 5). This can be explained by the lower settling velocity of this sediment class
12
(0.0001 m s-1 for dispersed grains, note that flocculation is not taken into account in the model)
13
compared to that of coarse silt, allowing transportation beyond their origin of erosion. For example,
14
for a residual tidal velocity (root mean square) of 0.2 m s-1 and a water depth of 10 m in shallow
15
areas, the fine silt can be transported 20 km. Moreover, the macrofauna contribution to fine silt is
16
almost five times greater than for coarse silt, which provides more suspended sediment to be
17
transported much farther, triggering a large-scale impact on SSC. The very fine silt at the bed was
18
rapidly exhausted due to its relatively low initial mass (1/5 of fine or coarse silt) and its very low
19
critical bed shear stress (0.1 N m-2), making this class of sediment in our model runs less sensitive to
20
the biologically mediated sediment properties at the bed (erodibility, critical bed shear stress).
21
Therefore, the difference of very fine silt in near-bed SSC between macrofaunal mediated sediment
22
transport and without macrofauna effect is small. Moreover, the very low settling velocity of these
23
particles (0.0002 m s-1) permitted a large spatial distribution of the slightly modified near-bed SSC by
24
horizontal transport processes beyond the inhabited zones and rendered effects virtually invisible in
25
the model runs. When interpreting the results, it has to be kept in mind that sediment availability,
26
i.e. the thickness of the initial sediment layer, is a calibration parameter in the model. Here we have
27
focused on the topmost layer of sediment only, which accounts for deposition and resuspension of
28
fine sediments but not for the more complex interaction between consolidated bed and active
29
sediment transport layer. In future studies the extension of the simple one-layer bed model towards
30
a multi-layer bed model representing stratigraphy is needed.
AC C
EP
TE D
M AN U
9
23
ACCEPTED MANUSCRIPT 1
The implication of the choice of model parameters becomes especially evident when considering sediment availability. This is essentially set by the initial sediment thickness, here chosen to be 10
3
mm. The unexpected decrease in modelled near-bed SSC in some areas, especially those with high
4
macrofauna abundance (>100 ind. m-2), can now be attributed either to a not sufficient initial
5
sediment thickness in the model or taken as process present in nature as well due to high
6
macrofauna mediated entrainment rates and limited sediment availability (e.g. Stammermann and
7
Piasecki, 2012). In model areas covered or dominated by F. fabula, erodibility increases and the
8
critical bed shear stress decreases. These two processes lead to higher and more frequent
9
resuspension, primarily over these zones. Consequently, the sediment mass is eroded more
RI PT
2
frequently and finally exhausted more quickly than is the case when macrofauna-related effects are
11
not considered. Figure 7 demonstrates this situation (referred to as the reference when the
12
macrofauna effect is included) for the area west of Sylt before and after the storm event.
M AN U
13
SC
10
It was found that, at the same location west of Sylt (Figure 5 d), macrofauna resulted in enhancement of the near-bed SSC of fine silt compared with a reduction for near-bed SSC of coarse
15
silt. The reason can be demonstrated by sediment availability in both cases (with and without
16
macrofauna effects) presented in Figure 7 a. These two figures indicate the importance of sediment
17
availability in the model as a limiting factor. Moreover, it can be observed in Figure 7 a that the
18
sediment mass decreases more rapidly when macrofauna effects on sediment erodibility and critical
19
bed shear stress are included (compare the slope of the two curves of the lower diagram just before
20
the storm event shown with a vertical green line). Sediment mass of coarse silt had been already
21
exhausted prior to the storm event on February 3 so that insufficient sediment was available for
22
erosion during the storm (Figure 7 b). Compared to the case without macrofauna effects, the
23
sediment mass is at least two orders of magnitude greater and the erosion process is not limited by
24
sediment availability. Consequently, the concentration of near-bed SSC during the storm event is
25
higher for the case without any macrofauna effect (light shaded area in the upper Figure 7 b).
AC C
EP
TE D
14
24
RI PT
ACCEPTED MANUSCRIPT
1 2 3 4
the contribution of biotic modification (blue and pink shading, upper panel) in the first two weeks of February 2010.
5
especially after the storm event (green line).
SC
Available mass is shown in the lower panel indicating frequent depletion of sediment available for resuspension,
M AN U
6
Figure 7: Near-bed simulated suspended sediment concentration (SSC, for fine and coarse silt fractions) including
Local in-situ measurements would be needed to further study the interaction of sediment
7
availability and macrofauna effects on sediment transport. The complexity of the interaction of
8
macrofauna and sediment transport was already highlighted by the laboratory flume experiments of
9
Friedrichs et al. (2009). They were able to quantify how the interaction of biogenic structures with the near-bed flow regime affects the sediment flux. At low population densities (roughness densities
11
below 2 %), there was an increase in the erosion fluxes and deposition of suspended material. At
12
densities above 4 %, reduced particle exchange was found and therefore erosion almost stopped
13
inside the test arrays. For roughness densities of about 5 %, it can therefore be assumed that a
14
skimming flow exists where a new benthic boundary layer eventually forms above the biogenic
15
structures. Furthermore filtration activities, especially by bivalves, could affect SSC in nature, even
16
though the extent of this effect is highly dependent on the abundance and the ambient flow
17
conditions (van Duren et al., 2006; Le Hir et al. 2007).
EP
AC C
18
TE D
10
Our results have important implications for ecosystem modeling. Light penetration through
19
the water column is affected by increased turbidity induced by macrofauna at the seabed, which
20
impacts primary production and population dynamics of phytoplankton. This finding indicates the
21
indirect coupling between macrofauna and phytoplankton through SSC. It should, however, be
22
investigated whether such a linkage is limited to storm events or may have long-term effects.
25
ACCEPTED MANUSCRIPT The model experiments which have been presented here partly depend on our choices of model
2
parameters. In the absence of sufficient data for the determination of sediment parameters (critical
3
bed shear stress, erodibility and sediment fractions as well as thickness in a few centimeters of the
4
upper bed layers), these were calibrated by comparing the simulated SSC with measured values or
5
values adopted from literature. The same shortage holds for F. fabula abundance data. This was
6
restricted to the locations of observation and was therefore extrapolated and interpolated for other
7
regions. Seasonal variations in macrofauna abundance were not considered. Furthermore, in the
8
present study, the macrofauna effect of a single dominant species was considered, ignoring the
9
effect of several other species and their non-linear interaction within the community. Most
RI PT
1
importantly the effect of suspension feeders which will introduce a net sink of SSC has not been
11
included so far. The parameterizations adopted for the inclusion of macrofauna effects were based
12
on very limited data for Macoma balthica from literature, which was adopted for F. fabula. These
13
effects can be highly non-linear. As presented in Eq. 6 erodibility increases exponentially between an
14
abundance of 100 to 800 ind. m-2,following a sigmoid function, leading to greatly increased erosion
15
rates in inhabited areas and finally to exhaustion of the available bed material. Another possible
16
limitation of our model is that macrofauna effects were assumed to be constant over time and other
17
macrofauna effects such as those on roughness or biodeposition were not considered. In the case of
18
F. fabula, however, the temporal variability of abundance is known to be limited (Creutzberg, 1986).
19
We did not consider the effect of biological reworking on the migration of bed forms (Borsje et al.,
20
2009) since we focused on suspended sediment transport.
TE D
M AN U
SC
10
Finally, we propose more laboratory experiments addressing the issue of how macrobenthic
22
organisms (also in the form of communities) alter sediment transport properties. This would greatly
23
reduce the uncertainty in critical model formulations. In addition, future work would greatly benefit
24
by the availability of more SSC measurements in the German Bight and ideally data based large-scale
25
SSC estimates in addition to small-scale studies measuring simultaneously hydrodynamic, SSC as well
26
as bed sediment and biological parameters. This may result in new and more accurate and integral
27
(because biological) parameterizations of sediment transport models.
28
5
AC C
EP
21
Conclusions
29
The observed influence of macrofauna on sediments at the seabed and the seabed structure
30
itself has been investigated in terms of large-scale impacts on suspended sediment concentration in 26
ACCEPTED MANUSCRIPT a model study of a shallow shelf sea. The southern North Sea has been taken as an example in order
2
to study the impact of macrofauna in a numerical model by integrating a generic benthos module in a
3
coupled 3D model system. Modification of critical bed shear stress and erodibility constituted the
4
major proxies for macrofauna impacts on sediment. A feature of the present study is the focus on
5
large-scale effects (>100 km), combined with the application of the measured spatial distribution of
6
the bivalve F. fabula as a chosen characteristic species.
RI PT
1
7
The simulation results indicate that macrofauna is not only able to significantly modify SSC locally
8
but also beyond the inhabited zones. As expected, the enhanced SSC due to macrofauna is most
9
pronounced for high energy conditions such as storm events but also persists over a spring-neap tidal cycle. Changes in SSC were not confined to near-bed zones but extended throughout the water
11
column due to vertical mixing.
Furthermore, it was shown that the magnitude and horizontal extent of macrofauna modified
M AN U
12
SC
10
SSC are constrained by sediment availability (in the model) and depend on the sediment properties
14
such as settling velocity and critical bed shear stress for deposition and erosion. Fine silt, being the
15
dominant fraction for suspended sediment in the water column, was found to be very sensitive to
16
the occurrence of macrofauna in the model. While the present study contributes to the improvement
17
of ecosystem and sediment transport modeling by offering new insights into large-scale effects of
18
macrofauna on SSC in the southern North Sea, it should be noted that physical and biological
19
processes in the model are less complex than in nature and rely partly on simplified
20
parameterizations, which opens up various new research issues, both in empirical science and
21
modeling.
22
6
EP
Acknowledgement
AC C
23
TE D
13
This paper is the result of collaboration between the MOSSCO project partners funded by the
24
German Federal Ministry of Education and Research (BMBF) as part of the Coastal Research Agenda
25
for the North Sea and Baltic Sea (grant numbers 03F0668A, 03F0667A, 03F0667B ). Further financial
26
support of K. Klingbeil has been provided by the Collaborative Research Center TRR181 on Energy
27
Transfers in Atmosphere and Ocean, funded by the German Research Foundation. We acknowledge
28
the contribution of J. Dannheim and M. Zeiler of the German Federal Maritime and Hydrographic
29
Agency (BSH) for providing F. fabula abundance data and Wenyan Zhang from the Helmholtz27
ACCEPTED MANUSCRIPT Zentrum Geesthacht (PACES Programme of the Helmholtz-Gemeinschaft) for completing the data set
2
for the southern North Sea. We thank A. Hammrich for contributions to the model validation.
3
Processed satellite data were provided by Brockmann Consult. Moreover, we thank Stefan Foster
4
and two anonymous reviewers as well as Christian Winter as guest editor for their helpful
5
recommendations and suggestions which improved our manuscript significantly.
6
7
7 8
Literature
RI PT
1
Andersen, T., and M. Pejrup (2011), Biological Influences on Sediment Behavior and Transport, in Treatise on Estuarine and Coastal Science, pp. 289–309, Elsevier.
Arfai, J., Jähne, F., Lutz, R., Franke, D., Gaedicke, C. & Kley, J., (2014). Late Palaeozoic to Early
10
Cenozoic geological evolution of the northwestern German North Sea (Entenschnabel): New results
11
and insights. Netherlands Journal of Geosciences -Geologie en Mijnbouw, 93(4): 147-174.
M AN U
12
SC
9
Austen, I., T. J. Andersen, and K. Edelvang (1999), The Influence of Benthic Diatoms and
13
Invertebrates on the Erodibility of an Intertidal Mudflat, the Danish Wadden Sea, Estuarine, Coastal
14
and Shelf Science, 49(1), 99–111, doi:10.1006/ecss.1998.0491.
16 17
Becker, G. A., S. Dick and J. W. Dippner (1992), Hydrography of the German Bight, Marine Ecology Progress Series, 91, 9-18.
TE D
15
Borsje, B. W., S. Hulscher, M. B. de Vries, and G. J. de Boer (2007), Modeling large scale cohesive
18
sediment transport by including biological activity, River, Coastal and Estuarine Morphodynamics:
19
RCEM 2007, 255–262.
Borsje, B. W., Hulscher,Suzanne Jacqueline Marie Hélène, P. M. J. Herman, and M. B. Vries
EP
20
(2009), On the parameterization of biological influences on offshore sand wave dynamics, Ocean
22
Dynamics, 59(5), 659–670, doi:10.1007/s10236-009-0199-0.
AC C
21
23
Bremner, J., S. Rogers, and C. Frid (2006), Methods for describing ecological functioning of
24
marine benthic assemblages using biological traits analysis (BTA), Ecological Indicators, 6(3), 609–
25
622, doi:10.1016/j.ecolind.2005.08.026.
26
Breugem, W. A., and L. H. Holthuijsen (2007), Generalized Shallow Water Wave Growth from
27
Lake George, J. Waterway, Port, Coastal, Ocean Eng., 133(3), 173–182, doi:10.1061/(ASCE)0733-
28
950X(2007)133:3(173). 28
ACCEPTED MANUSCRIPT 1
Briggs, K. B., G. Cartwright, C. T. Friedrichs, and S. Shivarudruppa (2015), Biogenic effects on
2
cohesive sediment erodibility resulting from recurring seasonal hypoxia on the Louisiana shelf,
3
Continental Shelf Research, 93, 17–26, doi:10.1016/j.csr.2014.11.005.
4 5
Burchard, H., and K. Bolding (2002), GETM, A General Estuarine Transport Model: Scientific Documentation, Technical Report EUR 20253 EN. European Commission. Burchard, H., G. Flöser, J. V. Staneva, T. H. Badewien, and R. Riethmüller (2008), Impact of
7
Density Gradients on Net Sediment Transport into the Wadden Sea, J. Phys. Oceanogr., 38(3), 566–
8
587, doi:10.1175/2007JPO3796.1.
9
RI PT
6
Canuto, V. M., A. Howard, Y. Cheng, and M. S. Dubovikov (2001), Ocean Turbulence. Part I: OnePoint Closure Model—Momentum and Heat Vertical Diffusivities, J. Phys. Oceanogr., 31(6), 1413–
11
1426, doi:10.1175/1520-0485(2001)031<1413:OTPIOP>2.0.CO;2.
SC
10
Creutzberg, F. (1986), Distribution patterns of two bivalve species (Nucula turgida, Tellina fabula)
13
along a frontal system in the Southern North Sea, Netherlands Journal of Sea Research, 20(2-3), 305–
14
311, doi:10.1016/0077-7579(86)90052-9.
15
M AN U
12
Dannheim, J. (2014), Bewertungsansätze für Raumordnung und Genehmigungsverfahren im Hinblick auf das benthische System und Habitatstrukturen, Statusbericht und Web-Dienst im
17
GeoSeaPortal des BSH: Forschungsvorhaben im Auftrag des BSH.
18
TE D
16
Darr, A., M. Gogina, and M. L. Zettler (2014), Functional changes in benthic communities along a
19
salinity gradient– a western Baltic case study, Journal of Sea Research, 85, 315–324,
20
doi:10.1016/j.seares.2013.06.003.
Davis, W. R. (1993), The role of bioturbation in sediment resuspension and its interaction with
EP
21
physical shearing, Journal of Experimental Marine Biology and Ecology, 171(2), 187–200,
23
doi:10.1016/0022-0981(93)90003-7.
24 25 26
AC C
22
Doerffer, R., H. Schiller, and M. Peters (2006), Meris regional case 2 water algorithms (c2r), www.brockmann-consult.de/beam/software/plugins/merisc2r-1.1. Duineveld, G., A. Künitzer, U. Niermann, P. de Wilde, and J. S. Gray (1991), The macrobenthos of
27
the north sea, Netherlands Journal of Sea Research, 28(1-2), 53–65, doi:10.1016/0077-
28
7579(91)90004-K.
29
ACCEPTED MANUSCRIPT 1
François, F., J.-C. Poggiale, J.-P. Durbec, and G. Stora (1997), A New Approach for the Modelling
2
of Sediment Reworking Induced by a Macrobenthic Community, Acta Biotheoretica, 45(3/4), 295–
3
319, doi:10.1023/A:1000636109604.
5 6
Friedrichs, M., T. Leipe, F. Peine and G. Graf (2009). "Impact of macrozoobenthic structures on near-bed sediment fluxes." Journal of Marine Systems 75(3-4): 336-347. Gayer, G., S. Dick, A. Pleskachevsky, and W. Rosenthal (2006), Numerical modeling of suspended
RI PT
4
7
matter transport in the North Sea, Ocean Dynamics, 56(1), 62–77, doi:10.1007/s10236-006-0070-5.
8
Graf, G., and R. Rosenberg (1997), Bioresuspension and biodeposition: a review, Journal of
10
Marine Systems, 11(3-4), 269–278, doi:10.1016/S0924-7963(96)00126-1.
SC
9
Gräwe, U., G. Flöser, T. Gerkema, M. Duran-Matute, T. H. Badewien, E. Schulz, and H. Burchard (2016), A numerical model for the entire Wadden Sea: Skill assessment and analysis of
12
hydrodynamics, J. Geophys. Res. Oceans, 121(7), 5231–5251, doi:10.1002/2016JC011655.
13
M AN U
11
Harris, R. J., C. A. Pilditch, B. L. Greenfield, V. Moon, and I. Kröncke (2016), The Influence of
14
Benthic Macrofauna on the Erodibility of Intertidal Sediments with Varying mud Content in Three
15
New Zealand Estuaries, Estuaries and Coasts, 39(3), 815–828, doi:10.1007/s12237-015-0036-2.
17 18
Hayter, E. J. and A. J. Mehta (1986), Modelling cohesive sediment transport in estuarial waters. In: Applied Mathematical Modelling, 10 (4), 294–303, doi: 10.1016/0307-904X(86)90061-2.
TE D
16
Heath, M. R., A. C. Edwards, J. Patsch, and Turrell Scottish Executive Central Research Unit (2002), Modelling the behaviour of nutrients in the coastal waters of Scotland, nowel,
20
http://strathprints.strath.ac.uk/18568/.
22
Hill, C., C. DeLuca, Balaji, M. Suarez, and A. Da Silva (2004), The architecture of the earth system modeling framework, Comput. Sci. Eng., 6(1), 18–28, doi:10.1109/MCISE.2004.1255817.
AC C
21
EP
19
23
Hofmeister, R., K. Bolding, R. D. Hetland, G. Schernewski, H. Siegel, and H. Burchard (2013), The
24
dynamics of cooling water discharge in a shallow, non-tidal embayment, Continental Shelf Research,
25
71, 68–77, doi:10.1016/j.csr.2013.10.006.
26
Jumars, P. A., and A. Nowell (1984), Effects of benthos on sediment transport: difficulties with
27
functional grouping, Continental Shelf Research, 3(2), 115–130, doi:10.1016/0278-4343(84)90002-5.
28
Kappenberg, J., and I. Grabemann (2001), Variability of the Mixing Zones and Estuarine Turbidity
29
Maxima in the Elbe and Weser Estuaries, Estuaries, 24(5), 699, doi:10.2307/1352878. 30
ACCEPTED MANUSCRIPT 1
Kerimoglu, O., R. Hofmeister, J. Maerz, and K. Wirtz (2017), A novel acclimative biogeochemical
2
model and its implementation to the southern North Sea, Biogeosciences Discuss., doi:10.5194/bg-
3
2017-104.
4
Klingbeil, K., and H. Burchard (2013), Implementation of a direct nonhydrostatic pressure gradient discretisation into a layered ocean model, Ocean Modelling, 65, 64–77,
6
doi:10.1016/j.ocemod.2013.02.002.
RI PT
5
7
Klingbeil, K., M. Mohammadi-Aragh, U. Gräwe, and H. Burchard (2014), Quantification of
8
spurious dissipation and mixing – Discrete variance decay in a Finite-Volume framework, Ocean
9
Modelling, 81, 49–64, doi:10.1016/j.ocemod.2014.06.001.
Knaapen, M., H. Holzhauer, S. Hulscher, M. J. Baptist, M. d. Vries, and M. van Ledden (2003), On
11
the modelling of biological effects on morphology in estuaries and seas, edited by A. Sánchez-Arcilla
12
and A. Bateman, pp. 773–783, IHAR.
M AN U
13
SC
10
Krause, G., G. Budeus, D. Gerdes, K. Schaumann, K. Hesse (1986), Frontal systems in the German
14
Bight and their physical and biological effects, in: Nihoul, J.C.J. (Ed.) Marine interfaces
15
ecohydrodynamics: proceedings of the 17th International Liège Colloquium on Ocean
16
Hydrodynamics. Elsevier Oceanography Series, 42, 119-140
Kristensen, E., J. M. Neto, M. Lundkvist, L. Frederiksen, M. Â. Pardal, T. Valdemarsen, and M. R.
TE D
17 18
Flindt (2013), Influence of benthic macroinvertebrates on the erodability of estuarine cohesive
19
sediments: Density- and biomass-specific responses, Estuarine, Coastal and Shelf Science,
20
doi:10.1016/j.ecss.2013.09.020.
Kröncke, I., H. Reiss, J. D. Eggleton, J. Aldridge, M. J. Bergman, S. Cochrane, J. A. Craeymeersch, S.
EP
21
Degraer, N. Desroy, J.-M. Dewarumez, G. C. Duineveld, K. Essink, H. Hillewaert, M. S. Lavaleye, A.
23
Moll, S. Nehring, R. Newell, E. Oug, T. Pohlmann, E. Rachor, M. Robertson, H. Rumohr, M.
24
Schratzberger, R. Smith, E. V. Berghe, J. van Dalfsen, G. van Hoey, M. Vincx, W. Willems, and H. L.
25
Rees (2011), Changes in North Sea macrofauna communities and species distribution between 1986
26
and 2000, Estuarine, Coastal and Shelf Science, 94(1), 1–15, doi:10.1016/j.ecss.2011.04.008.
27
Krumbein, W. E. (1994), Biostabilization of sediments, 526 pp., Carl-von-Ossietzky-Univ.
28
AC C
22
Oldenburg, Oldenburg.
31
ACCEPTED MANUSCRIPT 1
Le Hir, P., Y. Monbet, and F. Orvain (2007), Sediment erodability in sediment transport modelling:
2
Can we account for biota effects?, Continental Shelf Research, 27(8), 1116–1142,
3
doi:10.1016/j.csr.2005.11.016.
5 6
Le Hir, P., Cayocca, F., & Waeles, B. (2011). Dynamics of sand and mud mixtures: a multiprocessbased modelling strategy. Continental Shelf Research, 31(10), S135-S149. Lemmen, C., Hofmeister, R., Klingbeil, K., Nasermoaddeli, M.H., Kerimoglu, O., Burchard, H.,
RI PT
4
7
Kösters, F., Wirtz, K.W. (2017), Modular System for Shelves and Coasts (MOSSCO v1.0) - a flexible and
8
multi-component framework for coupled coastal ocean ecosystem modelling, Geosci. Model Dev.
9
submitted., http://arxiv.org/abs/1706.04224
Lettmann, K. A., J.-O. Wolff, and T. H. Badewien (2009), Modeling the impact of wind and waves
11
on suspended particulate matter fluxes in the East Frisian Wadden Sea (southern North Sea), Ocean
12
Dynamics, 59(2), 239–262, doi:10.1007/s10236-009-0194-5.
M AN U
13
SC
10
Lumborg, U., T. J. Andersen, and M. Pejrup (2006), The effect of Hydrobia ulvae and
14
microphytobenthos on cohesive sediment dynamics on an intertidal mudflat described by means of
15
numerical modelling, Estuarine, Coastal and Shelf Science, 68(1–2), 208–220,
16
doi:10.1016/j.ecss.2005.11.039.
Maa, J. P.-Y., J.-I. Kwon, K.-N. Hwang, and H.-K. Ha (2008), Critical Bed-Shear Stress for Cohesive
TE D
17 18
Sediment Deposition under Steady Flows, J. Hydraul. Eng., 134(12), 1767–1771,
19
doi:10.1061/(ASCE)0733-9429(2008)134:12(1767).
20
Montserrat Trotsenburg, F. (2011), Estuarine ecosystem engineering: biogeomorphology in the estuarine intertidal, Dissertation, Civil Engineering and Geosciences, Hydraulic Engineering,
22
Technische Universiteit Delft, Delft.
Nasermoaddeli, M. H., C. Lemmen, R. Hofmeister, F. Koesters, and K. Klingbeil (2014), The
AC C
23
EP
21
24
benthic geoecology model within the modular system for shelves and coasts (MOSSCO), in HIC2014-
25
11. International Conference on Hydroinformatics: Informatics and the Environment: Data and Model
26
Integration in a Heterogeneous Hydro World, edited by M. Piaseki, City University of New York, New
27
York, USA.
28
Neumann, H., H. Reiss, S. Ehrich, A. Sell, K. Panten, M. Kloppmann, I. Wilhelms, and I. Kröncke
29
(2012), Benthos and demersal fish habitats in the German Exclusive Economic Zone (EEZ) of the
30
North Sea, Helgol Mar Res, doi:10.1007/s10152-012-0334-z. 32
ACCEPTED MANUSCRIPT 1 2
Nowell, A. R. M., and P. A. Jumars (1984), Flow environments of aquatic benthos, Annual Review of Ecology and Systematics, 15, 303–328.
3
Orvain, F. (2005), A model of sediment transport under the influence of surface bioturbation:
4
generalisation to the facultative suspension-feeder Scrobicularia plana, Mar. Ecol. Prog. Ser., 286,
5
43–56, doi:10.3354/meps286043. Orvain, F., K. Guizien, S. Lefebvre, M. Bréret, and C. Dupuy (2014), Relevance of macrozoobenthic
RI PT
6 7
grazers to understand the dynamic behaviour of sediment erodibility and microphytobenthos
8
resuspension in sunny summer conditions, Journal of Sea Research,
9
doi:10.1016/j.seares.2014.03.004.
Orvain, F., P. Le Hir, and P.-G. Sauriau (2003), A model of fluff layer erosion and subsequent bed
SC
10
erosion in the presence of the bioturbator, Hydrobia ulvae, J Mar Res, 61(6), 821–849,
12
doi:10.1357/002224003322981165.
13
M AN U
11
Orvain, F., P. Le Hir, P.-G. Sauriau, and S. Lefebvre (2012), Modelling the effects of macrofauna on sediment transport and bed elevation: Application over a cross-shore mudflat profile and model
15
validation, Estuarine, Coastal and Shelf Science, 108, 64–75, doi:10.1016/j.ecss.2011.12.036.
16
Orvain, F., P.-G. Sauriau, C. Bacher, and M. Prineau (2006), The influence of sediment
17
cohesiveness on bioturbation effects due to Hydrobia ulvae on the initial erosion of intertidal
18
sediments: A study combining flume and model approaches, Journal of Sea Research, 55(1), 54–73,
19
doi:10.1016/j.seares.2005.10.002.
TE D
14
Paarlberg, A. J., M. Knaapen, M. B. de Vries, S. Hulscher, and Z. B. Wang (2005), Biological
21
influences on morphology and bed composition of an intertidal flat, Estuarine, Coastal and Shelf
22
Science, 64(4), 577–590, doi:10.1016/j.ecss.2005.04.008.
24 25 26 27
Partheniades, E. (1965), Erosion and Deposition of Cohesive Soils, Journal of the Hydraulics
AC C
23
EP
20
Division, 91(1), 105–139.
Peine, F., B. Bobertz, and G. Graf (2005), Influence of the blue mussel Mytilus edulis (Linnaeus) on the bottom roughness length (z0) in the south-western Baltic Sea, Baltica, 18(1), 13–22. Pleskachevsky, A., G. Gayer, J. Horstmann, and W. Rosenthal (2005), Synergy of satellite remote
28
sensing and numerical modeling for monitoring of suspended particulate matter, Ocean Dynamics,
29
55(1), 2–9, doi:10.1007/s10236-004-0101-z. 33
ACCEPTED MANUSCRIPT 1 2 3
Postma, H. (1984), Introduction to the symposium on organic matter in the Wadden Sea, Neth. Inst. Sea Res. Pub. Ser., 10, 15–22. Puls, W., J. Beusekom, U. Brockmann, R. Doerffer, U. Hentschke, P. König, D. Murphy, B. Mayer, A. Müller, T. Pohlmann, A. Reimer, R. Schmidt-Nia, and J. Sündermann (1999), SPM concentrations in
5
the German Bight: Comparison between a model simulation and measurements, Deutsche
6
Hydrographische Zeitschrift, 51(2-3), 221–244, doi:10.1007/BF02764175.
7
RI PT
4
Puls, W., H. Kühl, A. Frohse, and P. König (1995), Measurements of the suspended matter settling
8
velocity in the German Bight (North Sea), Deutsche Hydrographische Zeitschrift, 47(4), 259–276,
9
doi:10.1007/BF02737787.
Puls, W., T. Pohlmann, and J. Sündermann (1997), Suspended particulate matter in the Southern
SC
10
North Sea: Application of a numerical model to extend NERC North Sea project data interpretation,
12
Deutsche Hydrographische Zeitschrift, 49(2-3), 307–327, doi:10.1007/BF02764041.
13 14 15
M AN U
11
Rachor, E., and P. Nehmer (2003), Erfassung und Bewertung ökologisch wertvoller Lebensräume in der Nordsee, 175 pp., Bundesamt für Naturschutz, Bremerhaven, hdl:10013/epic.36744.d001. Radach, G., and J. Pätsch (2007), Variability of continental riverine freshwater and nutrient inputs into the North Sea for the years 1977–2000 and its consequences for the assessment of
17
eutrophication, Estuaries and Coasts, 30(1), 66–81, doi:10.1007/BF02782968.
18 19
TE D
16
Reiss, H., S. Cunze, K. König, H. Neumann, and I. Kröncke (2011), Species distribution modelling of marine benthos: a North Sea case study, Mar. Ecol. Prog. Ser., 442, 71–86, doi:10.3354/meps09391. Rijkswaterstaat (2017), Waterbase, http://live.waterbase.nl/ [accessed 19 Jan 2017].
21
Salzwedel, H., E. Rachor, and D. Gerdes (1985), Benthic macrofauna communities in the German
23
Bight, Veröff. Inst. Meeresforsch. Bremerh., 20, 199–267.
AC C
22
EP
20
Sanford, L. P. (2008), Modeling a dynamically varying mixed sediment bed with erosion,
24
deposition, bioturbation, consolidation, and armoring, Computers & Geosciences, 34(10), 1263–1283,
25
doi:10.1016/j.cageo.2008.02.011.
26
Schrum, C. (1997), Thermohaline stratification and instabilities at tidal mixing fronts: Results of
27
an eddy resolving model for the German Bight, Continental Shelf Research, 17(6), 689–716,
28
doi:10.1016/S0278-4343(96)00051-9.
34
ACCEPTED MANUSCRIPT 1
Seifert, T., W. Fennel, and C. Kuhrts (2009), High resolution model studies of transport of
2
sedimentary material in the south-western Baltic, Journal of Marine Systems, 75(3-4), 382–396,
3
doi:10.1016/j.jmarsys.2007.01.017.
4
Shimeta, J., C. L. Amos, S. E. Beaulieu, and O. M. Ashiru (2002), Sequential resuspension of protists by accelerating tidal flow: implications for community structure in the benthic boundary
6
layer, Limnol. Oceangr., 47, 1152–1164.
8 9 10 11
Smagorinsky, J. (1963), General circulation experiments with the primitive equations, Mon. Wea. Rev., 91(3), 99–164, doi:10.1175/1520-0493(1963)091<0099:GCEWTP>2.3.CO;2.
Soulsby, R. (1997), Dynamics of marine sands: A manual for practical applications, 1st ed., Telford, London.
SC
7
RI PT
5
Stammermann, R., and M. Piasecki (2012), Influence of Sediment Availability, Vegetation, and Sea Level Rise on the Development of Tidal Marshes in the Delaware Bay: A Review, Journal of
13
Coastal Research, 285, 1536–1549, doi:10.2112/JCOASTRES-D-11-00143.1.
14
M AN U
12
Stanev, E. V., R. Al-Nadhairi, J. Staneva, J. Schulz-Stellenfleth, and A. Valle-Levinson (2014), Tidal
15
wave transformations in the German Bight, Ocean Dynamics, 64(7), 951–968, doi:10.1007/s10236-
16
014-0733-6.
Stanev, E. V., M. Dobrynin, A. Pleskachevsky, S. Grayek, and H. Günther (2009), Bed shear stress
TE D
17 18
in the southern North Sea as an important driver for suspended sediment dynamics, Ocean
19
Dynamics, 59(2), 183–194, doi:10.1007/s10236-008-0171-4. Stips, A., K. Bolding, T. Pohlmann, and H. Burchard (2004), Simulating the temporal and spatial
21
dynamics of the North Sea using the new model GETM (general estuarine transport model), Ocean
22
Dynamics, 54(2), 266–283, doi:10.1007/s10236-003-0077-0. Umlauf, L., and H. Burchard (2005), Second-order turbulence closure models for geophysical
AC C
23
EP
20
24
boundary layers. A review of recent work, Continental Shelf Research, 25(7-8), 795–827,
25
doi:10.1016/j.csr.2004.08.004.
26
van der Molen, J., K. Bolding, N. Greenwood, and D. K. Mills (2009), A 1-D vertical multiple grain
27
size model of suspended particulate matter in combined currents and waves in shelf seas, J. Geophys.
28
Res., 114(F1), doi:10.1029/2008JF001150.
35
ACCEPTED MANUSCRIPT 1 2
van Duren, L. A., P. M. J. Herman, et al. (2006), Effects of mussel filtering activity on boundary layer structure, Journal of Sea Research, 55(1), 3-14.
3
van Ledden, M. (2003), Sand-mud segregation in estuaries and tidal basins, xvi, 218,
4
Communications on hydraulic and geotechnical engineering, report no. 03-2, Dept. of Civil
5
Engineering and Geosciences, Delft University of Technology, Delft. van Ledden, M., W. van Kesteren, and J. Winterwerp (2004), A conceptual framework for the
RI PT
6 7
erosion behaviour of sand–mud mixtures, Continental Shelf Research, 24(1), 1–11,
8
doi:10.1016/j.csr.2003.09.002.
9
van Leeuwen, S. M., J. van der Molen, P. Ruardij, L. Fernand, and T. Jickells (2013), Modelling the contribution of deep chlorophyll maxima to annual primary production in the North Sea,
11
Biogeochemistry, 113(1-3), 137–152, doi:10.1007/s10533-012-9704-5.
SC
10
13 14
van Moorsel, G. (2011), Species and habitats of the international Dogger Bank, ecosub, Doorn. 74 p.
M AN U
12
van Prooijen, Bram C., F. Montserrat, and P. M. Herman (2011), A process-based model for
15
erosion of Macoma balthica-affected mud beds, Continental Shelf Research, 31(6), 527–538,
16
doi:10.1016/j.csr.2010.12.008.
18 19
Wentworth, C.K., 1922, A scale of grade and class terms for clastic sediments: Journal of Geology, v. 30, p. 377-392.
TE D
17
Wheatcroft, R. A., and C. A. Butman (1997), Spatial and temporal variability in aggregated grainsize distributions, with implications for sediment dynamics, Continental Shelf Research, 17(4), 367–
21
390, doi:10.1016/S0278-4343(96)00035-0.
22
EP
20
Widdows, J., M. Brinsley, and M. Elliott (1998), Use of in situ flume to quantify particle flux (biodeposition rates and sediment erosion) for an intertidal mudflat in relation to changes in current
24
velocity and benthic macrofauna, Geological Society, London, Special Publications, 139(1), 85–97,
25
doi:10.1144/GSL.SP.1998.139.01.07.
26
AC C
23
Widdows, J., M. Brinsley, P. N. Salkeld, and C. H. Lucas (2000a), Influence of biota on spatial and
27
temporal variation in sediment erodability and material flux on a tidal flat (Westerschelde, The
28
Netherlands), Marine Ecology Progress Series, 194, 23–37.
36
ACCEPTED MANUSCRIPT 1
Widdows, J., S. Brown, M. Brinsley, P. Salkeld, and M. Elliott (2000b), Temporal changes in
2
intertidal sediment erodability: influence of biological and climatic factors, Continental Shelf
3
Research, 20(10-11), 1275–1289, doi:10.1016/S0278-4343(00)00023-6. Winterwerp, J. C. (2007), On the sedimentation rate of cohesive sediment, in Estuarine and
5
Coastal Fine Sediments Dynamics - Intercoh 2003, Proceedings in Marine Science, pp. 209–226,
6
Elsevier.
RI PT
4
7
Winterwerp, J. C., and W. G. M. van Kesteren (2004), Introduction to the physics of cohesive
8
sediment in the marine environment, 1st ed., 1 online resource (1 volume (various, Developments in
9
sedimentology, vol. 56, Elsevier, Amsterdam.
Wood, R., and J. Widdows (2002), A model of sediment transport over an intertidal transect,
SC
10
comparing the influences of biological and physical factors, Limnol. Oceangr., 47(3), 848–855,
12
doi:10.4319/lo.2002.47.3.0848.
M AN U
11
13
AC C
EP
TE D
14
37
ACCEPTED MANUSCRIPT
1
1
2
1.1
3
Appendix – Hydrodynamic model validation Introduction A brief hydrodynamic model validation for water levels and waves is presented in the following
for the reference year 2010. The validation – as far as possible on available data – for suspended
5
sediment concentration can be found in the main article. A detailed validation for the simulation of
6
temperature and salinity characteristics (e.g. thermohaline stratification, residual circulation) can be
7
found in Kerimoglu et al. (2017).
8
1.2
SC
RI PT
4
Observational data
The model validation was carried out for a number of stations along the Dutch and German
10
coasts. The analysis was restricted to positions representing the dynamics of the German Bight, thus
11
locations within the estuaries have not been analyzed. An overview of observational data used for
12
the model validation is given in Table 4. The location and measured properties of the individual
13
stations are shown in Figure 8.
TE D
waves (WAV), suspended sediment concentration (SSC) and total suspended matter (TSM).
Station Name
Abbreviation
Terschelling-Noordzee Huibertgat Borkum Fischerbalje Norderney-Riffgat Wangerooge-Nord Bake Z Cuxhaven Helgoland Wyk List
TSN HUI BOF NOY WAN BKZ CUX HEL WYK LIH
AC C
15
Table 4: Overview of observational data used for the model validation. Measured parameters are water levels (WL),
16
Position Latitude (deg.) Longitude (deg.) 53.450 5.333 53.567 6.400 53.558 6.748 53.697 7.158 53.806 7.929 54.014 8.315 53.868 8.718 54.179 7.890 54.694 8.577 55.017 8.441
EP
14
M AN U
9
Fireship "Elbe" List wave buoy FINO-I plattform
ELB LWB FN1
54.000 55.054 54.015
8.117 8.390 6.588
Terschelling T10 Terschelling T50 Rottumerplaat R3 Rottumerplaat R50 Sylt S9 Sylt S62
T10 T50 R3 R50 S9 S62
53.454 53.764 53.557 53.947 54.909 54.867
5.077 4.801 6.547 6.314 8.150 7.271
Measured Parameter WL WAV SSC TSM x x x x x x x x x x x x x
Data Source Rijkswaterstaat Rijkswaterstaat WSV WSV WSV WSV WSV WSV LKN WSV BSH COSYNA BSH
x x x x
x x x x x x
Rijkswaterstaat, Brockmann Consult / ESA Rijkswaterstaat, Brockmann Consult / ESA Rijkswaterstaat, Brockmann Consult / ESA Rijkswaterstaat, Brockmann Consult / ESA Brockmann Consult / ESA Brockmann Consult / ESA
38
ACCEPTED MANUSCRIPT 1
Observational data for water levels are based on tide gauge stations which were provided for the Netherlands by Rijkswaterstaat and for Germany by the Waterways and Shipping Administration
3
(WSV) and the Landesbetrieb für Küstenschutz, Nationalpark und Meeresschutz Schleswig-Holstein
4
(LKN). Wave measurements are based on wave buoy data provided by the German Federal Maritime
5
and Hydrographic Agency (BSH) and the Helmholtz-Zentrum Geesthacht (HZG) within the COSYNA
6
framework. Suspended sediment concentrations were obtained from monitoring stations provided
7
by Rijkswaterstaat and extracted from satellite measurements of total suspended matter carried out
8
by the European Space Agency (ESA) and processed by Brockmann Consult.
TE D
M AN U
SC
RI PT
2
EP
9 10
waves (WAV) and suspended sediment concentration (SSC / TSM).
AC C
11
Figure 8: Location of observation stations considered for the model validation for parameters water level (WL),
12
The propagation of the tidal wave from west to east along the coast can be fairly well covered
13
with the available data. Wave data is less frequently monitored but for the time period considered
14
here positions representing the inner German Bight (Fino-I), the outer estuaries (Elbe) and a near
15
coastal position (List) were available.
16
1.3
17
1.3.1 Water levels
Model validation
39
ACCEPTED MANUSCRIPT 1
As validation period for water levels a period of about two spring-neap cycles from July 19, 2010 to August 20, 2010 was chosen. Inspection of individual time series of water levels shows general
3
good agreement of observations and simulation (see Figure 9 for station Norderney Riffgat as an
4
example) in terms of tidal range and phase of the tidal wave.
M AN U
SC
RI PT
2
7 8 9
Figure 9: Simulated (red) and observed (blue) water levels for station Norderney Riffgat for validation period July 19, 2010 to August 20, 2010
As can be seen from the quantitative comparison of simulations and observations of water levels given in Table 5 the model is generally capable of reproducing the tidal range well. The difference
EP
6
TE D
5
between simulated and observed values is less than 10 % at most stations, larger differences occur at
11
station Wyk, a location strongly influenced by tidal flat dynamics and Cuxhaven. The correlation
12
between model and measurements is with a mean value for the correlation coefficient of 0.92 high.
13
There is a small bias in the model in terms of mean water level which is on average 12 cm too high.
14
There is also a phase difference for the propagation of the tidal wave by on average 25 minutes.
15
Generally the model lags observations but this trend is not present at all stations. Correcting the
16
simulated water levels at individual stations for this phase shift we obtain an RMSE of less than 30 cm
17
on average and for the standard RMSE of less than 40 cm.
AC C
10
40
ACCEPTED MANUSCRIPT 1
Table 5: Water level statistics comparing simulated (sim) and observed (obs) values of tidal range (TR), phase
2
difference for the propagation of the tidal wave, root mean square error (RMSE), phase corrected root mean square
3
error (RMSE*), model bias and correlation coefficient (Pearson R) for individual stations and mean over all stations
4 5
1.4
TR sim (m) 1.75 1.86 1.97 2.00 2.46 2.54 2.96 1.97 2.00 1.41 2.09
∆ TR obs-sim (m / %) 0.02 1% -0.10 -5% -0.12 -6% -0.16 -7% -0.04 -1% -0.01 0% 0.31 12% -0.11 -5% -0.43 -18% -0.16 -10% -0.08 -4%
Phase diff. (min) -26 -36 -23 -21 -27 -32 -30 -26 -64 35 -25
RMSE (m) 0.28 0.33 0.32 0.31 0.34 0.37 0.46 0.29 0.61 0.29 0.38
RMSE* (m) 0.23 0.23 0.28 0.27 0.25 0.24 0.36 0.23 0.40 0.23 0.27
Bias (m) 0.12 0.09 0.09 0.14 0.12 0.09 0.12 0.11 0.19 0.12 0.12
Pearson R 0.95 0.92 0.94 0.95 0.95 0.94 0.93 0.95 0.82 0.93 0.92
RI PT
Terschelling-Noordzee Huibertgat Borkum Fischerbalje Norderney-Riffgat Wangerooge-Nord Bake Z Cuxhaven Helgoland Wyk List Mean
TR obs (m) 1.74 1.97 2.09 2.16 2.50 2.55 2.65 2.08 2.43 1.56 2.17
SC
Station Name
Waves
As validation period for the wave model a time span from July 25, 2010 to August 20, 2010 was
7
chosen based on data availability and the presence of waves. The analysis is restricted to significant
8
wave heights as most relevant wave-related parameter for sediment transport modelling, the
9
evaluation of peak periods is not shown here. As can be seen from the comparison of observed and
10
simulated significant wave heights for station Fino-I in Figure 10 our wave model is able to roughly
11
capture the wave climate but cannot reproduce individual wave events reliably.
AC C
EP
TE D
M AN U
6
41
SC
RI PT
ACCEPTED MANUSCRIPT
M AN U
1
Figure 10: Simulated (red) and observed (blue) significant wave height (SWH) at station Fino-I
3
A statistical analysis for observed and modelled significant wave heights is given in Error!
4
Reference source not found.. The analysis is taking into account the frequency distribution of
5
significant wave height in 1 m classes, mean and maximum values as well as the 90th-percentile.
6
Table 6: Wave statistics comparing observed and simulated significant wave heights frequency distribution, mean, th
maximum and 90 -percentile analyzed for July 25, 2010 to August 20, 2010
Significant wave height distribution (%) Mean 0 m - 1 m 1 m - 2 m 2 m - 3 m 3 m - 4 m (m) 68.4 23.2 8.4 0.0 1.02 56.9 26.1 14.3 2.7 1.07 71.8 24.4 3.7 0.1 0.85 60.9 27.1 12.0 0.0 0.92 86.4 13.6 0.0 0.0 0.61 66.2 33.8 0.0 0.0 0.77
EP
7
TE D
2
9 10
AC C
8
FINO-I observed simulated Elbe observed simulated List observed simulated
Max (m) 2.73 3.48 3.16 2.81 1.67 1.82
P90 (m) 1.93 2.62 1.62 2.07 1.11 1.57
The mean wave height is simulated well within 10 % of the observed values for offshore locations, although for the nearshore location (station List) the difference is larger. Considering the
42
ACCEPTED MANUSCRIPT 1
90th-percentile and maximum wave height the model tends to overestimate wave heights. The same
2
tendency can be found in the frequency distribution.
3
1.5
4
Discussion Evaluating the model performance given above we find water levels sufficiently well reproduced
by the model for the scope of this study. Non-systematic differences in phase lag can probably
6
attributed to the not fully resolved coastline but do not indicate a general misrepresentation of
7
hydrodynamics. Most importantly the tidal energy is represented well; certain systematic errors do
8
exist (mean water level, phase lag) but are of minor importance for our sensitivity studies of
9
sediment transport.
SC
10
RI PT
5
The parametric wave model in the current set-up is only capable of providing a rough estimate of the actual wave climate. While the mean wave height is reproduced rather well the model tends to
12
overestimate wave heights. However, as the focus of this study is on estimating the effect of changes
13
in sediment resuspension due to the presence of macrofauna and the wave model provides the same
14
forcing for all sensitivity runs there will be no significant impact of the wave model imperfections on
15
our results.
TE D
16
M AN U
11
Taking into account the model´s resolution of about 0.5-1 nautical mile results are in parts surprisingly good. However, we can also clearly identify the need for model improvements for future
18
studies which are primarily related to the wave model and resolution and boundary conditions
19
(mean water level) of the hydrodynamic model.
AC C
20
EP
17
Overall the model performance can be considered sufficient for the scope of this study
21
particularly as we focus on a comparison between model experiments with and without macrofauna
22
effects based on the same hydrodynamic model.
43