Download Nutrient dynamics and phytoplankton development along an

Journal of Marine Systems 84 (2011) 49–66
Contents lists available at ScienceDirect
Journal of Marine Systems
j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j m a r s y s
Nutrient dynamics and phytoplankton development along an estuary–coastal zone
continuum: A model study
Sandra Arndt a,⁎, Geneviève Lacroix b, Nathalie Gypens c, Pierre Regnier a,d, Christiane Lancelot c
a
Department of Earth-Sciences, Utrecht University, Utrecht, The Netherlands
Management Unit of the North Sea Mathematical Models , Royal Belgian Institute for Natural Sciences, Brussels, Belgium
Ecologie des systèmes aquatiques, Université Libre de Bruxelles, Bruxelles, Belgium
d
Department of Earth and Environmental Sciences, Université Libre de Bruxelles, Bruxelles, Belgium
b
c
a r t i c l e
i n f o
Article history:
Received 3 November 2009
Received in revised form 18 August 2010
Accepted 23 August 2010
Available online 6 September 2010
Keywords:
Land–ocean continuum
Scheldt
Reactive-transport model
Nutrient cycling
Phytoplankton bloom
Estuarine filter
a b s t r a c t
This study presents a first attempt to quantify the biogeochemical transformations and fluxes of carbon and
nutrients along the entire mixing zone of the shallow, tidally-dominated estuary–coastal zone continuum of
the Scheldt (Belgium/The Netherlands). A fully transient, two-dimensional, nested-grid hydrodynamic
model of the continuum is coupled to the biogeochemical MIRO model for the coastal zone and the
CONTRASTE model for the estuary. Transient model simulations are performed with a high spatial (80–
750 m) and temporal (30 min) resolution over a period of one year (January–December 1995). The high
temporal resolution allows including the short-term variability triggered by the tides, the freshwater
discharge and the wind stress. System scale simulations provide time series of nutrient transformations and
fluxes along the entire estuary–coastal zone continuum, as well as highly resolved nutrient inventories for
the estuarine and the coastal zone sub-domains. Simulation results reveal that the balance between highly
variable estuarine nutrient inputs and physical constrains set by the unsteady residual transport field exert
an important control on the magnitude and succession of phytoplankton blooms and the ecosystem
structure in the coastal zone. In addition, they suggest that the poorly surveyed estuarine–coastal zone
interface plays a central role in the continuum. In this dynamic area, marked spatial concentration gradients
develop and episodically lead to a reversal of material fluxes from the coast into the estuary. During distinct
episodes of the productive period, euryhaline coastal diatoms intrude far upstream into the saline estuary.
This intrusion reduces the estuarine nutrient concentrations and export fluxes, thereby reinforcing the
nutrient limitation in the coastal area. As a consequence, the estuarine filter does not operate independently
from the processes in the coastal zone. The dynamic interplay between the two ecosystems and the intense
process rates operating at their transition, therefore, strongly supports our continuum approach.
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
Coastal environments play an important role in the global
biogeochemical cycles of carbon and macronutrients (DIN, SiðOH Þ4 ,
and PO4 ). Despite their relatively small volume, they contribute
disproportionately to global primary production (Killops and Killops,
1993). In coastal transition systems, tightly coupled geological,
hydrodynamic, geochemical and biological processes are modulated
by a wide array of forcing mechanisms, which include, light,
temperature, wind stress, waves, tides, freshwater discharge, as well
as continental and oceanic nutrient inputs. These forcings operate on
very different temporal and spatial scales and, together, set strong
constraints on the biogeochemical functioning of the coastal ocean.
⁎ Corresponding author. Department of Earth Sciences, Geochemistry, Faculty of
Geosciences, Utrecht University, P.O. Box 80.021, 3508 TA Utrecht, The Netherlands.
Tel.: +33 56133 2644.
E-mail address: [email protected] (S. Arndt).
0924-7963/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.jmarsys.2010.08.005
They determine the spatial and temporal distributions of dissolved
and particulate species and markedly influence the magnitude of
biogeochemical reaction rates. Achieving a detailed and mechanistic
understanding of the processes that control the biogeochemical
dynamics along the river–ocean continuum is thus particularly
challenging, yet, essential for predicting the response of the coastal
ecosystem to the continuously evolving anthropogenic pressure.
Over the past decades, human activity has significantly accelerated
the supply of land-derived nutrients to the coastal ocean and, as a
result, nearshore areas bordering highly populated regions have
experienced increased eutrophication (Neal et al., 2000; Cloern, 2001;
Turner et al., 2003). In north-western Europe, massive developments
of Phaeocystsis colonies have regularly been observed along the
coastline of the Eastern Channel and the Southern Bight of the North
Sea (Lancelot et al., 1987; Riegman et al., 1992; Mills et al., 1994;
Rousseau et al., 2002). A characteristic and recurrent diatom–
Phaeocystis–diatom succession develops in response to the seasonal
availability of nutrients (Rousseau et al., 2002). In this region, nutrient
50
S. Arndt et al. / Journal of Marine Systems 84 (2011) 49–66
levels are controlled by the mixing of Atlantic waters with nutrientrich freshwater from the Scheldt, the Rhine/Meuse and the Seine
estuary (Pingree and Maddock, 1977; Salomon and Breton, 1993;
Lacroix et al., 2004, 2007b). The present study focuses on the Belgian
coastal zone where phytoplankton and nutrient concentrations have
been intensively monitored at one sampling station (St. 330;
N51o26.05; E2o48.50) between 1988 and 2000, at time intervals
ranging from twice a week to twice a month (Breton et al., 2006).The
long-term time series thus provide key information on the variable
nutrient supply from the Scheldt estuary in this highly dynamic area.
Nonetheless, high-resolution data are only available at a single
sampling station and can thus not easily be extrapolated at the scale
of the entire coastal ecosystem. To date, observations on nutrient and
phytoplankton distributions in the Belgian coastal zone are available
but they are limited in time span and spatial coverage and merely
resolve the monthly timescale (Borges and Frankignoulle, 1999, 2002;
Muylaert et al., 2006; Van der Zee and Chou, 2005).
Reactive-transport models (RTMs) are particularly suitable to
develop a quantitative understanding of biogeochemical transformations and fluxes in coastal environments. In particular, they complement
field measurements, because they provide a mechanistic description of
process interactions over a spectrum of scales that cannot be resolved by
field observations. More than two decades ago, Nihoul and Hecq (1984)
identified the dominant patterns in the ecosystem structure of the
Belgian coastal zone using hydrodynamic simulations to support the
interpretation of ten year-averaged field observations of nutrients,
phytoplankton, zooplankton and suspended sediments. Since then,
valuable insights on the biogeochemical dynamics of the Belgian coastal
zone have been gained through further modeling efforts. In particular,
the MIRO model (Lancelot et al., 2005) was designed to describe diatom
and Phaeocystis colony blooms and the related carbon and nutrient (N, P,
and Si) cycles in the eastern channel and southern North Sea with a
focus on the Belgian coastal zone. It was first implemented in a multibox framework to study the link between anthropogenic nutrient loads
by the Seine and Scheldt rivers and the magnitude and extent of diatom/
Phaeocystis blooms (Lancelot et al., 2005), as well as its evolution over
the last decade (Gypens et al., 2007) and the last 50 years (Lancelot et al.,
2007). The carbon dynamics and CO2 air/water exchange which result
from this coupling has also been addressed (Gypens et al., 2004, 2009).
Recently, MIRO has been coupled to the COHERENS-3D hydrodynamic
model (Lacroix et al., 2007b). The coupled model has been used to
simulate and assess the relative impact of river (Seine, Somme, Scheldt,
and Rhine/Meuse) and Atlantic waters on the contemporary (1991–
2003) spatio-temporal distribution of diatom/Phaeocystis blooms and
the related carbon and nutrient cycling in the Channel and Southern
Bight of the North Sea (Lacroix et al., 2007a). At the larger scale of the
North Sea, three-dimensional box-averaged biogeochemical models
forced by spatially-resolved transport models (Baretta et al., 1995;
Patsch and Radach, 1997), two-dimensional simplified biogeochemical
models (Hydes et al., 1996) and three-dimensional coupled transport
and biogeochemical models (Skogen et al., 1995; Delhez, 1998; Moll,
1998) have also been used to simulate primary production (Moll, 1998;
Skogen and Soiland, 1998; Luyten et al., 1999; Skogen and Moll, 2005),
toxic algae blooms (Cugier et al., 2005), nutrient cycling (Allen et al.,
2001; Vichi et al., 2004) and long-term eutrophication (Patsch and
Radach, 1997) in the entire North Sea. In parallel to the development of
North Sea models, a suite of RTMs that couple the physics and
biogeochemistry of estuaries have been developed over the last two
decades. The physical extension of these estuarine models was generally
restricted to the area between the estuarine mouth and the upper limit
of the salt intrusion (Soetaert and Herman, 1995a, b; Regnier et al., 1997;
Regnier and Steefel, 1999; Vanderborght et al., 2002; Tappin et al., 2003;
Arndt et al., 2009). For instance, Regnier and Steefel (1999) used the 1D
model CONTRASTE to calculate the nitrogen export fluxes to the coastal
zone in an estuarine system subject to fully-transient conditions.
However, the control of these fluxes on coastal biogeochemistry was not
explicitly simulated. Although, integrated representations of the
morphologically complex river–ocean continuum are increasingly
used (Proctor et al., 2000; Tappin et al., 2003; Cugier et al., 2005;
Lancelot et al., 2007; Vanderborght et al., 2007) they generally
compromise on the spatial, temporal and dynamical resolution. In
particular, no study has addressed so far the seasonal evolution in
ecosystem structure and functioning of the Scheldt estuary and adjacent
coastal zone using a continuum approach, that is, none has considered
the influence of the estuarine domain on the coastal environment and
vice-versa. Nutrient supplies are either specified by measured riverine
fluxes discharging directly into the coastal zone or, at best, transiting
through a poorly resolved estuarine system. Therefore, these
approaches do not provide a realistic description of the retention and
transformation of land-derived nutrient inputs within the estuary,
which may influence both the magnitude and timing of export fluxes to
the coastal zone. In addition, the feedbacks of coastal processes on the
estuarine environment are currently not accounted for. Estuarine
models generally assume that the biogeochemistry of the estuary is
exclusively influenced by the riverine inputs from the watershed.
Therefore, combined observational and modeling efforts do not provide
a comprehensive, synoptic view of the biogeochemical dynamics along
the entire mixing-zone between river and marine waters and the
quantitative significance of the estuary for the functioning of the
adjacent coastal zone remains poorly evaluated. As a result, the
mechanisms that lead to the emergence and short-term temporal
variability in ecosystem structure along such an estuary–coastal zone
continuum are still largely unknown.
In this study, a fully transient two-dimensional, nested-grid
hydrodynamic model of the Scheldt (B/NL) and adjacent nearshore
coastal zone continuum is coupled to the biogeochemical models MIRO
and CONTRASTE for the coastal zone and the estuary, respectively.
Transient model simulations are performed with a high spatial (80–
750 m) and temporal (30 min) resolution over a period of one year
(January–December 1995). The system scale simulations provide time
series of nutrient transformations and fluxes along the entire estuary–
coastal zone continuum, as well as weekly-resolved nutrient inventories
for the estuarine and the coastal zone sub-domains. These results are
then used to investigate the respective importance of the physical and
biogeochemical controls on the spatio-temporal structure of an entire
estuary–coastal zone transition continuum.
2. The model
2.1. Hydrodynamics
The estuarine and coastal hydrodynamics is described using the
nested hydrodynamic model MIKE 21 NHD© (DHI). The model
simulates unsteady two-dimensional flows in a vertically homogeneous water column. It solves simultaneously the set of depthaveraged Saint-Venant equations for barotropic flow:
∂ξ
∂p
∂q
+
+
=0
∂t
∂x
∂y
!
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
p2 + q2
∂ pq
∂ξ
+ gh
+ gp
∂y h !
∂x
h2 Ch2
∂2 p
∂2 q
h ∂
−Ωx q−E
+
p =0
−f ⋅w⋅wx +
2
ρw ∂x a
∂x
∂y2
∂p
∂ p2
+
∂t
∂x h
+
ð2Þ
!
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
∂ pq
∂ξ
p2 + q2
+ gh
+ gq
∂x h
∂y
h2 Ch2
!
∂2 p
∂2 q
h ∂
+ Ωy p−E
+
−f ⋅w⋅wy +
p =0
ρw ∂y a
∂x2
∂y2
2
∂q
∂ q
+
∂t
∂y h
ð1Þ
+
ð3Þ
S. Arndt et al. / Journal of Marine Systems 84 (2011) 49–66
where ξ is the water elevation (m); p = vx ⋅ h, q = vy ⋅ h are the flux
densities (m2 /s); h is the water depth (m); g is the acceleration due to
gravity (m/s2); Ch is the Chézy coefficient (m1/2 /s); Ωx, Ωy are the
Coriolis coefficients (1/s); E is the eddy viscosity (m2 /s); f is the wind
friction factor; w is the wind speed; wx,wy are the wind components in
x- and y-directions (m s − 1); ρw is the water density (kg m − 3) and pa
the atmospheric pressure (kg m − 1 s − 2). Eqs. (1)–(3) are spatially
discretised on dynamically nested, rectangular, staggered grids and
integrated using a ‘fractioned-step’ technique combined with an
Alternating Direction Implicit (ADI) algorithm (Leendertse, 1967).
The ADI scheme leads to a tri-diagonal system which is solved by
applying a double-sweep algorithm. Three dynamically nested grids
which cover the brackish and saline zones of the Scheldt estuary, as
well as the well-mixed part of the Belgian coastal zone form the model
support. The grid size is 83.33 × 83.33 m for the channel up to the
Belgian/Dutch border (grid 3, Fig. 1), 250 × 250 m for the lower
estuary (grid 2, Fig. 1) and 750 × 750 m for the coastal zone (grid 1,
Fig. 1). The total amount of grid points is larger than 60,000 for a total
surface area of 8700 km2. The temporal resolution of the hydrodynamic simulations is limited by the Courant criterium and set to 30 s.
Simulations are performed from June 1994 to December 1995 and
results are processed for the period between January 1995 to
December 1995, after the model has reached a hydrodynamic regime.
The eddy viscosity coefficient, E, in Eqs. (2)–(3) is calculated
according to the Smagorinsky formula, which relates E to the local
flow velocities using a Smagorinsky constant of 0.5 (Smagorinsky,
1963). Bed friction exerted on the moving water is characterized by
means of a roughness formulation with a Chézy coefficient, Ch. This
coefficient is the key parameter for the hydrodynamic model
calibration, performed by comparing observed and simulated tidal
amplitudes and phases. The best calibration is achieved with
Ch = 90 m1/2s − 1 for the coastal zone and Ch = 75 m1/2s − 1 for the
Scheldt estuary. The wind friction factor f is set to 0.006.
2.2. Transport
Coupled reaction and mass transport is described by the twodimensional continuity equation for scalar quantities:
∂hCi
∂
∂
∂
ðvx hCi Þ =
+
h⋅Dsx ⋅ Ci + h ∑ Rn;i
∂t
∂x
∂x
∂x
n
ð4Þ
∂hCi
∂ ∂
∂
+
vy hCi =
h⋅Dsy ⋅ Ci + h ∑ Rn;i
∂t
∂y
∂y
∂y
n
ð5Þ
51
where C is the concentration of species i; vx, vy are the horizontal
velocity vector components; Ds, x, Ds, y are the horizontal dispersion
coefficients and Ri = ∑ n Rn, i denotes the sum of all transformation
processes, n, affecting species i. Eqs. (4) and (5) are integrated
numerically using an operator splitting approach. The advection–
dispersion terms are solved with the third order finite difference
scheme QUICKEST (Ekebjaerg and Justesen, 1991), which avoids
instabilities associated with central differencing and numerical
dissipation. The source/sink term in the continuity equation is
integrated using the Euler method (e.g. Press et al., 1992) with a
time step of 1800 s, and is discussed in further detail in the following
section. The dispersion coefficients account mainly for the nonresolved sub-grid scale processes and are calibrated using longitudinal salinity profiles as an indicator of the mixing of water masses. A
good agreement with observations is achieved with dispersion
coefficients of 40 m2 /s, 100 m2 /s and 150 m2 /s for the upper estuary,
the lower estuary and the coastal zone, respectively.
2.3. Biogeochemistry
The biogeochemical reaction-network is coupled to the hydrodynamic and transport processes using the ECOLab©environmental
modeling tool (DHI). The reaction network is based on a combination
of the biogeochemical MIRO model (Lancelot et al., 2005), which is
implemented for the coastal zone (grid 1, Fig. 1) and the biogeochemical reaction network of the CONTRASTE model (Regnier and
Steefel, 1999; Vanderborght et al., 2007), which simulates the
biogeochemical transformations along the estuarine gradient (grids
2 and 3, Fig. 1).
2.3.1. The MIRO model
MIRO is a mechanistic biogeochemical model describing carbon,
nitrogen, phosphorous and silica cycling through aggregated components of the planktonic and benthic realms of Phaeocystis dominated
ecosystem (Lancelot et al., 2005). Its structure includes thirty-eight
state variables assembled in four modules describing the dynamics of
phytoplankton (diatoms, nanoflagellates and Phaeocystis colonies) in a
nutrient (DIN; DIP; SiðOH Þ4 )- and light-limited environment, zooplankton (copepods and microzooplankton), dissolved and particulate
organic matter (each with two classes of biodegradability) degradation and nutrients (NO3 , NH4 , PO4 and SiðOH Þ4 ) regeneration by
bacteria in the water column and the sediment. Equations and
parameters were formulated based on current knowledge on the
kinetics and the factors controlling the main auto- and heterotrophic
processes involved in the functioning of the coastal marine ecosystem.
Fig. 1. Bathymetric map of the Scheldt (Belgium/The Netherlands) estuary (grid 1 and grid 2) and the adjacent Dutch/Belgian coastal zone (grid 3). The limits of the three nested
grids are indicated by black boxes.
52
S. Arndt et al. / Journal of Marine Systems 84 (2011) 49–66
These are fully documented in Lancelot et al. (2005) and www.int-res.
com/journals/suppl/appendix_lancelot.pdf. Compared to the original
MIRO model, the two-dimensional implementation described in this
work does not account for diagenetic recycling processes, since at the
intra-annual timescale the benthic regeneration of nutrients is
generally negligible compared to the pelagic recycling processes and
the estuarine nutrient fluxes in the nearshore coastal zone (Lancelot
et al., 2005). In addition, the description of photosynthesis was
modified to account for the coupling between primary production,
light availability and SPM (see Section 2.3.3). Furthermore, the MIRO
model was extended by an explicit description of oxygen to allow for a
realistic description of O2 dynamics in the estuary. Oxygen is affected
by the following source/sink terms:
∑RO2 = Roxtr;O2 + RNPP;O2 −2⋅Rnitr;O2 −Rox;O2
ð6Þ
In Eq. (6), Roxtr, O2 denotes the rate of O2-transfer through the air/
water interface, RNPP, O2, is the oxygen production by net primary
production, while Rnitr, O2 and Rox, O2 are the rates of pelagic oxygen
consumption through nitrification and aerobic respiration,
respectively.
2.3.2. The CONTRASTE model
The description of biogeochemical processes in the estuary with
the exception of primary production is based on the latest version of
the 1D-CONTRASTE model (Regnier et al., 1997; Regnier and Steefel,
1999; Vanderborght et al., 2002, 2007; Arndt et al., 2007, 2009) and
includes organic matter, nutrients (ammonium, nitrate, phosphate
and silica) and oxygen as state variables. Processes included are
heterotrophic respiration, nitrification, denitrification and gas transfer
at the water/air interface. Further details about their respective
formulation can be found in Regnier and Steefel (1999) and
Vanderborght et al. (2002, 2007)). The CONTRASTE model is extended
by a simple description of phosphate dynamics, where phosphate is
taken up by phytoplankton and released during aerobic respiration
and denitrification. The description of estuarine phytoplankton
dynamics is adapted from the MIRO approach. Based on field
observations (Koeman et al., 1992; Kromkamp and Peepe, 1995;
Muylaert and Sabbe, 1999; Muylaert et al., 2009), only diatoms are
considered as an explicit state variable. Phaeocystis and nanoflagellates are assumed to lyse and remineralize within the estuarine
domain. Studies of phytoplankton speciation along the Scheldt
estuarine gradient show a complex distribution from typical freshwater/estuarine species (e.g. Cyclotella spp.) in the freshwater tidal
reaches to marine species (e.g. Chaetoceros spp., Rhizosolenia spp.,
Guinardia delicatula) in the vicinity of the estuarine mouth (Rijstenbil
et al., 1993; Kromkamp and Peepe, 1995; Muylaert and Sabbe, 1999;
Muylaert et al., 2000, 2009). Distinct riverine and coastal communities
are imported into the estuary through river discharge and tidal
mixing, respectively (Muylaert et al., 2009). To account for this
process, a salinity threshold that allows distinguishing between
marine and estuarine diatom communities was applied. It has been
shown that coastal diatoms are generally characterized by a growth
optimum at salinities between 25 and 33 (Brandt, 1984). However,
some marine species tolerate lower salinities, but their growth rate
generally decreases with increasing salinity (Brandt, 1984; Roubeix
et al., 2008). Muylaert et al. (2009) showed that marine species
mainly contribute to the species diversity in the polyhaline areas of
the Scheldt estuary, while a complex mixture of estuarine, freshwater
and coastal diatom species dominates at lower salinities. In addition,
observations reveal that the spring phytoplankton development
never intrudes further than 50 km (salinity ca. 2–15) into the estuary
and spring maximum chlorophyll a concentrations are generally
measured around km 40 (salinity ca. 15–25) (Van Damme et al.,
2005). Based on these observations, the threshold value was chosen to
allow for the intrusion of coastal diatom species down to salinities of
15. Parameter values for estuarine diatom growth (Table 1) where
chosen according to Billen and Garnier (1997) and adjusted in order
to better describe existing datasets on estuarine spring diatom
biomasses (Van Damme et al., 2005).
2.3.3. Coupling SPM dynamics, light availability and primary production
The description of phytoplankton growth in the estuary–coastal
zone continuum was modified to incorporate an explicit coupling
with the suspended matter (SPM) dynamics. The depth-integrated
photosynthetic activity φi (mg C m − 3 d − 1) of phytoplankton species
i is calculated according to:
i
φi = kmax ⋅iF∫
zmax
0
1− exp
!
αi
exp
ð
K
z
Þ
dz
I
0⋅
d⋅
kimax
ð7Þ
where kimax (d − 1) is the maximum rate of photosynthesis; iF (mg Cm − 3)
is the fraction of functional and structural metabolites in the biomass of
species i; I0 is the incident photosynthetically active radiation; Kd (m − 1) is
Table 1
Parameter values used in the SPM transport model and in the biogeochemical model. Only the parameter values which differ from the MIRO implementation given in (Lancelot et al.,
2005) are reported here.
Symbol
Description
Estuary
Value coastal zone
Unit
Reference
ϕ
ρ
M
τcrd
ρSPM
η
g
ρw
kNDA
kSiDA
kPDA
k0
kSPM
DA
Topt
Porosity
Density
Erosion constant
Critical shear stress
Sediment density
Water viscosity
Gravitational acceleration
Water density
Half-saturation constant for DIN uptake
Half-saturation constant for SiðOH Þ4 uptake
Half-saturation constant for PO4 uptake
Apparent background attenuation coefficient
Specific attenuation of SPM
Optimal growth temperature
Days 1 to 160
Days 160 to 365
DA temperature interval
Days 1 to 160
Days 160 to 365
0.7
2650
5.5 ⋅ 10 − 4
0.2
1760
0.0013
9.82
1000
5
7
0.5
1.4
0.0592
0.7
2650
5.5 ⋅ 10 − 4
0.3–0.5
1760
0.0013
9.82
1000
0.8
0.4
0.3
0.188
0.024
–
kg ⋅ m − 3
m⋅d
N⋅m−2
kg ⋅ m − 3
N⋅s⋅m−2
m⋅s−2
kg ⋅ m − 3
mmol m − 3
mmol m − 3
mmol m − 3
m−1
mgl − 1 m − 1
Estimated
–
1, calibrated
1, calibrated
Estimated
–
–
–
2,3
2,3
2,3
4,5
4,5
25
25
5.5
15
. °C
. °C
Estimated from 4,3
Estimated from 4, 3
5
5
5
12
. °C
. °C
Estimated from 4, calib.
Estimated from 4,3
dTDA
1: (Arndt et al., 2007); 2: (Billen and Garnier, 1997); 3: (Lancelot, 1995); 4: (Desmit et al., 2005); 5: (Rousseau, 2000).
−1
S. Arndt et al. / Journal of Marine Systems 84 (2011) 49–66
the light extinction coefficient and αi (μE m2 s − 1) is the photosynthetic
efficiency. Eq. (7) is integrated numerically using a cost-efficient analytical
approximation (Vanderborght et al., 2007). The extinction coefficient Kd in
Eq. (7) is an explicit function of the SPM concentration:
Kd = k0 + kSPM ⋅SPM
ð8Þ
where k0 is the apparent background attenuation and kSPM denotes the
SPM specific attenuation constants, respectively. In the coupled model,
SPM concentrations are calculated by a fully-formulated SPM transport
model (Arndt et al., 2007). The erosion and deposition fluxes are
expressed according to the formulations of Ariathurai (1974), Partheniades (1962) and Einstein and Krone (1962):
∂hSPM
∂
∂ ∂
∂
ðvx hSPM Þ +
+
vy hSPM =
h⋅Dsx ⋅ SPM
∂t
∂x
∂y
∂x
∂x
∂
∂
h⋅Dsy ⋅ SPM + Rero −Rdep
+
∂y
∂y
ð9Þ
where
Rero = pero ⋅M⋅SPMb
ð10Þ
Rdep = pdep ⋅ws ⋅SPM
ð11Þ
In Eqs. (10) and (11), pero and pdep stand for the probabilities of
erosion and deposition, respectively, M denotes the erosion constant
(m/s), ws represents the settling velocity of particles (m/s) and SPMb
refers to the concentration of total solids in the sediment. The settling
velocity ws depends directly on the diameter of suspended particles.
The dynamical adaptation of the SPM diameter to processes such as
deposition and erosion is described using the effective variable
approach proposed by Wirtz and Eckhardt (1996) and Wirtz (1997).
The wide range of different particle diameters is represented by a
continuous gaussian grain size distribution, which is completely
characterized by the dynamically adapting average grain diameter
and its standard deviation (Arndt et al., 2007). Parameter values of the
SPM transport model are given in Table 1.
53
addition, the three-dimensional model provides the depth-averaged
concentrations of biogeochemical state variables, temperature and
salinity along the open sea boundaries. At the upstream boundary, the
water flux through the estuarine cross-section, calculated by the onedimensional estuarine model CONTRASTE (Regnier et al., 1998), is
specified. Weekly measurements from the Hemiksen station (Regnier
and Steefel, 1999) provide the boundary concentrations and determine the typical resolution of freshwater scalar inputs to the
estuarine–coastal zone model. At the boundaries SPM concentrations
are set to a constant background value of 50 mg l − 1 and 100 mg l − 1
at the open sea and estuarine limits, respectively. A two dimensional
wind-field is specified by interpolated maps, which are extracted from
the coarser three-dimensional model (Lacroix et al., 2007b, a). The
incident light intensity I0 (Royal Meteorological Institute of Belgium)
is given as time-dependent external forcing, which resolves the daily
variations in light regime.
2.5. Validation data
Time series of surface elevation (data source: Hydro Meteo
Centrum Zeeland (HMCZ), www.hmcz.nl) at six stations along the
estuary–coastal zone continuum and observed phase lags of the tidal
wave in the Scheldt estuary (Horrevoets et al., 2004) are used to
validate the hydrodynamics in the Scheldt estuary and the adjacent
coastal zone. Transport is validated on the basis of salinity time series
from three monitoring stations along the estuarine–coastal zone
gradient (data source: Hydro Meteo Centrum Zeeland (HMCZ), www.
hmcz.nl). Simulated SPM concentrations are compared to field
observations at six different coastal and estuarine stations (data
source: Ministerie van Verkeer en Waterstaat, www.waterbase.nl). In
addition, time series from five estuarine and coastal zone monitoring
stations (data source: Ministerie van Verkeer en Waterstaat, www.
waterbase.nl) and reference station 330 (N51o26.05 ; E2o48.50, Breton
et al. (2006)) are used for the validation of the seasonal dynamics of
nutrients (NH4 , NO3 , SiðOH Þ4 , PO4 ), phytoplankton mphChl a and O2
concentrations.
3. Results and discussion
2.4. Boundary conditions and forcings
3.1. Hydrodynamics and transport
Elevations and current directions are extracted from a coarser
three-dimensional model (3D-MIRO&CO) for the year 1995 (Lacroix
et al., 2007b, a) and imposed at the two open North Sea boundaries. In
The hydrodynamics in the Scheldt estuary and the adjacent coastal
zone is determined by strong tidal forcing. Over a spring/neap cycle, the
tidal wave reaches an amplitude of 4.3/2.8 m in the coastal zone close to
Fig. 2. Comparison of observed (gray, data from Hydro Meteo Centrum Zeeland (HMCZ)) and simulated (black) surface elevations at different locations (see Fig. 1) along the estuary–
coastal zone continuum in March 1995. Surface elevations are given in the Belgian ordnance datum (TAW), which refers to mean low water level about 2.3 m below mean sea level.
54
S. Arndt et al. / Journal of Marine Systems 84 (2011) 49–66
Zeebrugge and 4.6/3.0 m near Oostende. In the estuary, the tidal range
gradually increases due to convergence to a maximum amplitude of 5.8/
4.6 m, until friction-induced dampening exceeds its amplification
upstream of Antwerp. Fig. 2 compares measured and simulated surface
elevations at different stations along the estuarine–coastal zone
continuum. The time series reflect the dominant tidal forcing, on
which local wind speed variations are superimposed. The average
deviation between observations and simulations at all stations is ca.
6 cm and the simulated elevations are thus in good agreement with
observations. The tidal wave propagates relatively fast (ca. 15 m s − 1)
through the coastal zone and the lower estuary, where waters are still
relatively deep (average depth 10 m). Close to the Dutch/Belgian border,
the wave is partially reflected due to the sudden reduction in estuarine
width. The superposition of the reflected and the incoming wave results
in almost infinite propagation speeds. Further upstream, the propagation speed decreases (10 m s − 1) as a consequence of the decreasing
average water depths (h = 4.5 m). The good agreement between
simulated and observed propagation of the tidal wave in the estuary
indicates that the model captures the main hydrodynamic features in
this system Fig. 3.
The large tidal currents lead to a relatively well-mixed water
column in the estuary and the adjacent coastal zone throughout the
year. However, the freshwater discharge of the Scheldt produces a
pronounced horizontal salinity gradient along the estuarine curvilinear axis and in a significant portion of the coastal zone. The salinity
distribution depends on the freshwater discharge of the Scheldt, the
tidal excursion and the prevailing wind speed and direction. Salinity
increases gradually from the upstream boundary to the mouth. The
less saline estuarine water then flows to the south–west along the
coastline, forming the Scheldt plume (Fig. 4-a). In the estuary, the
daily mean salinity variations reflect the dominant influence of
freshwater discharge (Fig. 4-d). At the beginning of February,
extremely high river discharges (N600 m3 s − 1) push less saline
water far downstream and lead to a strong decrease in salinity at the
estuarine stations (Fig. 4-c and -d). At the Baalhoek station, located
roughly 50 km upstream from the estuarine mouth, salinity normally
varies between 15 and 20. Yet, the winter freshwater pulse results in
unusually low salinities, which drop down to 3 (Fig. 4-d). The offshore
station shows much less variability (Fig. 4-b). Here, the salinity is
mainly controlled by local wind conditions, which determine the
extension of the Scheldt plume and the influx of less saline water
through the northern boundary. Fig. 4-a illustrates this dominant
influence of the local wind field (direction and speed) on the
Fig. 3. Comparison of observed (dots, (Horrevoets et al., 2004)) and simulated (line)
phase lags of the tidal wave along the estuarine gradient.
spreading of the estuarine plume. At the beginning of February, very
strong south-westerly winds limit the extension of the plume, despite
the extremely high freshwater discharge (N600 m3 s−1) and are
responsible for the very sharp longitudinal salinity gradient in the
estuary (Fig. 4-a). On the other hand, well established northerly winds
early in April push the Scheldt plume further south and thus lead to
lower salinities in the coastal zone (Fig. 4-a, also shown by Lacroix
et al. (2004)). In general, long-term residual transport processes
determine the spatial structure of the estuarine plume. The Eulerian
residual transport velocities (Fig. 5-a) reveal a general northwards
circulation, resulting mainly from the large-scale sea surface slope
across the shelf, and to a lesser extent, meteorological forcing (Delhez
and Carabin, 2001). However, the Eulerian residual transport field
does not explain the clockwise rotation of the Scheldt plume.
Interactions between tidal currents and residual eddies, arising from
tide-topography interactions, result in a strong stokes drift and, thus,
in a pronounced difference between the Eulerian and the Lagrangian
residual transports. Therefore, it has long been known that the
Lagrangian residual transport provides a better representation of the
coastal mean circulation in environments characterized by a complex,
irregular bathymetry. However, the computation of the Lagrangian
residual transport field is difficult due to its unsteady character. It is
usually determined by releasing particles in the time-varying velocity
field. Yet, the net-displacement of released particles strongly depends
on the hydrodynamic history and thus on their release time. Fig. 5
shows the general structure of the Lagrangian residual transport field,
according to a first-order approximation of the surface Lagrangian
residual transport velocity calculated by Delhez and Carabin (2001)
and the results of tracer experiments (Van den Eynde, 2004). In
addition, it illustrates the variability of the transport field due to the
effect of different wind directions by analyzing the tracks of seven
particles, released at different locations in the estuary and the coastal
zone at the end of January and in April 1995. The simulated tracks
agree well with the first-order approximation of the surface
Lagrangian residual transport velocity calculated by Delhez and
Carabin (2001) and the results of tracer experiments (Van den
Eynde, 2004). In the estuary (tracks 1 and 2; Fig. 5-c), most of the
residual transport is limited to the main tidal channels. In addition,
the oscillatory tidal forcing results in small residual velocities and,
therefore, long residence times. Close to the mouth (track 3; Fig. 5-c),
the residual circulation follows the deep navigational channel along
the coast until the bathymetry induces an offshore drift close to
Zeebrugge (Fig. 5-c). Further offshore (tracks 4 and 5; Fig. 5-c), the
difference between Lagrangian and Eulerian transport is reduced by
greater water depths and a more regular bathymetry. The Lagrangian
circulation thus resembles the northeastern Eulerian circulation. A
residual gyre forms offshore of the mouth, and increases the residence
time in this area. In the southern part of the model area (track 6;
Fig. 5-c), the residual transport is directed along the coastline in a
narrow band close to the coast. Further offshore (track 7; Fig. 5-c), the
complex bathymetry directs the particle northwards. Local wind
conditions modify this general pattern and, therefore, result in
differences between the simulated Langragian particle tracks
recorded at the end of January and April 1995 (Fig. 5-c). At the end
of January, predominantly southwesterly winds lead to a stronger
alongshore, residual transport in the southern part of the model area
(track 6; Fig. 5-c), but limit the offshore transport at the mouth (track
3; Fig. 5-c). In April (Fig. 5-c), when northwesterly winds prevail, the
particle released at the estuarine mouth is transported further
offshore into the residual gyre (track 3; Fig. 5-c), while the residual
transport in the southern area of the coastal zone is weaker (track 6;
Fig. 5-c). Offshore, particles are less influenced by the wind conditions
and are generally transported in northward direction (Fig. 5-c).
However, well established southerly winds reinforce the residual
transport, while weaker or northerly winds trigger a more complex
transport pattern (tracks 4, 5, and 7; Fig. 5-c).
S. Arndt et al. / Journal of Marine Systems 84 (2011) 49–66
55
Fig. 4. a—Instantaneous horizontal salinity distribution on February 1st and April 9th 1995. The respective wind directions and speed are indicated in the right lower corner of
the respective panels. Time series of observed (gray, Hydro Meteo Centrum Zeeland (HMCZ), www.hmcz.nl) and simulated (black) daily mean salinities at b—Vlakte van der
Raan, c—Hoofdplaat, d—Baalhoek.
3.2. Suspended matter
3.3. Spatio-temporal trends in nutrient, O2 and Chl a concentration
Fig. 6 illustrates the spatio-temporal dynamics of SPM in the
estuary and the adjacent coastal zone. The SPM distribution in the
estuarine coastal-zone continuum is controlled by the interplay
between hydrodynamic and meteorological conditions. In the estuary,
strong tidal currents and shallow water depths result in relatively
high turbidities (30–200 mg l − 1, Fig. 6-e and -f). A turbidity
maximum forms in the vicinity of Antwerp due to a maximum in
total energy dissipation (Chen, 2003; Chen et al., 2005; Arndt et al.,
2007). In the adjacent coastal zone, turbidity is generally lower and
falls within the range of observed SPM concentrations (Fig. 6-a to -d;
Fettweis and Van den Eynde (2003), Fettweis et al. (2006), Ruddick
et al. (2003). SPM concentrations show a characteristic distribution
(Fig. 6-g) with higher concentrations (50–200 mg l − 1) in the shallow
nearshore areas and decreasing concentrations (10–30 mg l − 1) in
offshore direction (Fettweis and Van den Eynde, 2003; Ruddick et al.,
2003). The decrease in residual transport between Oostende and
Zeebrugge, as well as shallow water depths in this area and the
difference between spring and neap tidal currents, lead to the
formation of a coastal turbidity maximum between Oostende and
the mouth of the Scheldt estuary (Nihoul, 1975; Gullentops et al.,
1976; Fettweis and Van den Eynde, 2003; Fettweis et al., 2006; Van
den Eynde, 2004).
SPM concentrations show a significant temporal variability (Fig. 6-a
to -e). Previously, a spectral analysis of estuarine SPM time series
revealed that fluctuations are mainly aggregated at tidal frequencies
(Arndt et al., 2007). In the coastal zone, the temporal variability of SPM
concentrations is not only controlled by the amplitude of the tide, but
also by wind conditions. In general, maximum daily mean concentrations are significantly higher during spring tides (Fig. 6-a to -e,
Fettweis et al. (2006)). However, the seasonal trend in the SPM record
is mainly induced by long-term variations in wind speed. In autumn,
winter and spring, higher wind speeds lead to increased maximum
SPM concentrations, while summer concentrations are generally
lower (Fig. 6; Fettweis et al., (2006)).
Fig. 7 compares the observed and simulated spatio-temporal
variability in nutrient (NH4 , NO3 , SiðOH Þ4 , PO4 ), phytoplankton Chl a
and O2 concentrations at selected locations along the estuarine–
coastal zone continuum over the year 1995. In general, nutrient and
O2 concentrations show a clear seasonal trend characterized by high
values in winter, a progressive decrease during the productive spring
and summer period, and a replenishment of nutrient levels in late
summer and autumn (de Galan et al., 2004; Van der Zee and Chou,
2005; Lancelot et al., 2005). In spring, nutrient depletion is significant
for NH4 , SiðOH Þ4 and PO4 , whileNO3 shows a similar but damped
evolution (Fig. 7-a to -d). Chl a time series reflect the development of
characteristic phytoplankton spring and summer blooms along the
estuary–coastal zone continuum that coincide with the periods of low
nutrient levels (Fig. 7-a to -d). Yet, the magnitude and the timing of
Chl a maxima differ along the estuary–coastal zone continuum. The
upper estuarine reaches (Fig. 7-a) are characterized by a relatively
broad early summer Chl a maximum whereas a sharp spring
maximum, followed by a smaller summer phytoplankton development are observed and simulated in the lower estuary (Fig. 7-b to -d)
and the adjacent coastal zone (Fig. 7-e and -f). These general seasonal
trends in nutrient and Chl a concentrations are superimposed by the
characteristic oscillatory tidal variability in the estuary, while
stochastic fluctuations induced by the wind forcing become more
important close the estuarine mouth and in the coastal zone. A
pronounced decrease in nutrient concentrations can also be observed
along the estuary–coastal zone continuum due to dilution and
consumption along the salinity gradient. At the mouth, nutrientenriched estuarine waters generally turn southwards and form the
Scheldt plume, following the Lagrangian residual transport along the
coastline (Fig. 5). It subsequently mixes with nutrient-poor Atlantic
waters offshore, resulting in a pronounced onshore/offshore and
north–east/south–west gradient (Van der Zee and Chou, 2005; de
Galan et al., 2004). However, during nutrient depletion in the vicinity
of the estuarine mouth, a reversed gradient in SiðOH Þ4 and NH4
56
S. Arndt et al. / Journal of Marine Systems 84 (2011) 49–66
a50
km
40
30
20
10
0
20
40
60
km
b-
c-
Fig. 5. a—Eulerian residual velocity field in the coastal zone. b—Schematic presentation of the general Lagrangian residual transport field according to (Delhez and Carabin, 2001) and
c—simulated Langragian particle tracks recorded over a period of one week at the end of January and April 1995.
concentrations can be observed, with slightly higher concentrations
offshore. O2 concentrations are generally low at the upstream limit of
the model domain (Fig. 7-a), where nitrification and respiration rates
exceed the uptake capacity of oxygen from the atmosphere. In
downstream direction, an intensification of mixing processes progressively improves water column oxygenation and ultimately results
in oxygen concentrations close to saturation (Fig. 7-b and -f) or even
slightly oversaturated in the coastal zone and estuarine mouth in early
spring (Fig. 7-d to -f). Later in the year, concentrations drop to lower
values due to enhanced respiration rates induced by the increased
availability of fresh organic matter from the declining phytoplankton
blooms (Fig. 7-c and -f).
The model performance is quantitatively evaluated on the basis of
different statistical measures (e.g. Allen et al., 2007). Correlation
coefficients, model efficiencies, percentage model biases and cost
functions are calculated as defined in Allen et al. (2007) on the basis of
tidally-averaged simulation results and the corresponding field data
for the entire estuary–coastal zone transition (Fig. 8). Correlation
coefficients and model efficiencies are high for macro-nutrients and
somewhat lower for phytoplankton and oxygen (Fig. 8-a and -b). The
percentage model bias shows that the model has the tendency to
slightly overestimate ammonium and underestimate phytoplankton
(Fig. 8-c). The cost function quantifies the difference between model
results and measurement data and reveals a very good fit for all
analyzed variables. The summary of basic model data fit metrics (e.g.
Allen et al., 2007) for the spatio-temporal dynamics of macronutrients
(NH4 , NO3 , SiðOHÞ4 ,and PO4 ), phytoplankton and oxygen (Fig. 8)
reveals that the model has skill for all variables analyzed. It indicates
that model performance slightly deteriorates from nitrate, silicate to
ammonium, phosphate and finally to phytoplankton and oxygen.
However, all evaluated variables reach good to excellent results and
the model thus performs well in capturing the spatio-temporal
dynamics along the estuary–coastal zone.
3.4. Seasonal evolution of estuarine and coastal primary producers
Fig. 9 illustrates the evolution of the simulated, volume-integrated
phytoplankton biomass in the coastal zone (model grid 1) and the
estuary (model grids 2 and 3). In both the estuary and the coastal
zone, the evolution of the phytoplankton biomass is characterized by
a spring and a summer bloom (Fig. 9-a and -b). Simulation results
reveal that in 1995, the annual Phaeocystis and diatom biomass
S. Arndt et al. / Journal of Marine Systems 84 (2011) 49–66
57
Fig. 6. a to f—Observed (data source: Ministerie van Verkeer en Waterstaat, www.waterbase.nl) and simulated daily maximum and minimum SPM concentrations at different
stations in the Scheldt estuary and the adjacent coastal zone. g—Two-dimensional distribution of mean, depth-integrated SPM concentrations. The locations of the measurement
sites shown in a to f are indicated by their respective letters.
represent 58% and 38% of the total annual, coastal phytoplankton
biomass, while nanophytoflagellates only contribute 4%. The estuarine
phytoplankton community is almost entirely dominated by diatoms.
In the coastal zone, the onset of the spring phytoplankton bloom
occurs in March with the development of a diatom bloom (Fig. 9-b).
Shortly after, both a small nanophytoflagellate and a pronounced
Phaeocystis colony bloom start to develop (Fig . 9-b). Coastal diatom
biomass peaks in mid-April before it rapidly decreases to low values
(Fig. 9-b). The simulated Phaeocystis colony biomass reaches several
maxima in late march and at the end of April and slowly declines to
lower values in May. A summer/fall diatom bloom, follows the decline
of the Phaeocystis colonies (Fig. 9-b). The simulated succession of
coastal phytoplankton agrees well with the previously observed
evolution of phytoplankton at a single monitoring station (Rousseau
et al., 2000, 2002; Sabbe et al., 2006) and in the entire coastal zone
(Lancelot et al., 2005; Muylaert et al., 2006). These field observations
showed that the coastal spring phytoplankton community is clearly
dominated by a bloom of Phaeocystis colonies (Lancelot et al., 1987,
1998; Cadee and Hegeman, 1991; Lancelot, 1995; Rousseau et al.,
2000). Diatoms are usually present throughout the productive period
and dominate the summer phytoplankton bloom (Lancelot et al.,
1998; Rousseau et al., 2000).
In the estuary, maximum spring diatom biomass coincides with the
spring bloom in the coastal zone (Fig. 9-a). Observed and simulated Chl
a time series reveal that diatoms mainly develop in the lower estuary
(Vlissingen-Hanswert, see Fig. 1), while, further upstream, environmental conditions do not meet the requirements for a spring diatom
bloom (Fig. 7). Multiple year monitoring results show a similar,
repetitive pattern with spring Chl a maxima of variable magnitude
(20–100 μg Chl a l − 1) in the lower Scheldt estuary (bkm 60) (Van
Damme et al., 2005). The simulated and observed spatio-temporal
pattern thus indicates that the increase in estuarine spring biomass is
initiated by the import of euryhaline coastal diatoms, which encounter
suitable growth conditions in the relatively nutrient-rich saline
estuary. Observations indicate that typical coastal phytoplankton
species contribute up to 20% to the total diversity of phytoplankton
58
S. Arndt et al. / Journal of Marine Systems 84 (2011) 49–66
a-
b-
c-
d-
e-
f-
Fig. 7. Annual evolution of measured (points) and simulated (gray line) nutrient, phytoplankton and oxygen concentrations at different stations in the model area (see Fig. 1).
Source: Ministerie van Verkeer en Waterstaat, www.waterbase.nl except for station 330 Rousseau (2000), Lancelot et al. (2005), Breton et al. (2006).
a
c
b
d
Fig. 8. Model performance summary statistics for the entire estuary–coastal zone transition (Fig. 7) data set, a—correlation coefficient, b—model efficiency, c—model bias, and d—cost
function.
S. Arndt et al. / Journal of Marine Systems 84 (2011) 49–66
a-
3.5
DA
2.5
PH
1.5
0
0
100
150
200
250
300
350
bDA
60
PH
NF
40
0
0
50
100
150
200
250
300
350
Fig. 9. Seasonal evolution of simulated volume-integrated a—estuarine phytoplankton
b—coastal zone phytoplankton biomass (DA = diatoms, PH = Phaeocystis, NF =
Nanoflagellates).
species in the polyhaline zone (salinity 30–18) of the Scheldt estuary
Muylaert et al. (2009). Similarly, Phaeocystis colonies are imported
into the estuary by tidal mixing at the time of their spring maxima. The
simulated import leads to a small accumulation of Phaeocystis biomass
in the vicinity of the estuarine mouth (Fig. 9-a). Therefore, they
periodically contribute to the estuarine phytoplankton biomass in the
vicinity of the estuarine mouth (Kromkamp and Peepe, 1995). In
summer, diatom biomass reveals a second increase (Fig. 9-a). This
increase coincides with the peak stage of the periodically observed
freshwater diatom bloom in the upper reaches of the Scheldt estuary
(Muylaert et al., 2000; Van Damme et al., 2005; Muylaert et al., 2005).
High summer Chl a concentrations at the upstream boundary, as well
as a rapid downstream concentration decrease (Fig. 7) indicate that
the simulated estuarine summer maximum is mainly triggered by the
import of freshwater diatoms (Arndt et al., 2007). In what follows,
these general trends in primary production dynamics are discussed in
the light of the constraints set by the seasonal evolution of nutrient
availability along the estuary–coastal zone continuum.
3.5. Nutrient cycling along the estuary–coastal zone continuum
Fig. 10A–C (a–g) summarizes the DIN, SiðOH Þ4 and PO4 dynamics
along the estuary–coastal zone continuum, respectively. The estuarine
domain comprises the area between the estuarine mouth in the
vicinity of Vlissingen and the upstream boundary (grids 2 and 3,
Fig. 1), while the coastal domain covers the entire coastal area
between the estuarine mouth and the open sea boundaries (grid 1,
Fig. 1). Fig. 10A–C, illustrate the seasonal evolution of the low-pass
filtered (cut-off frequency fc = 16d) nutrient influx through the
upstream boundary at Hemiksen (a), the low-pass filtered flux across
the estuarine mouth in the vicinity of Vlissingen (d) as well as at the
low-pass filtered flux through the open sea boundaries (g). In
addition, the spatially integrated, low-pass filtered net nutrient
transformation and the respective contribution of the various
individual processes (c and f) are shown together with the
instantaneous and low-pass filtered mass variations in each subsystem (b and e). The following paragraph discusses the evolution of
nutrients fluxes and transformations from the estuary (Fig. 10A–C (a–
d)) to the coastal zone (Fig. 10A–C (d–g)).
In the Scheldt estuary, the freshwater reaches are characterized by
short residence times and dominated by advective transport. At the
59
upstream boundary, nutrient fluxes thus correlate well with the
riverine inflow and well-defined winter maxima coincide with peak
river discharges (Fig. 10A–C-a). However, this correlation decreases in
summer, when enhanced process rates consume a larger fraction of
the nutrient inventory in the freshwater reaches, upstream of the
model domain (Arndt et al., 2009). Sharp summer minima in SiðOH Þ4
and PO4 inflow can be observed (Fig. 10B-a and C-a) and result from
the nutrient consumption upstream of the model domain during the
freshwater diatom bloom (Arndt et al., 2009). DIN consumption also
occurs and is largely dominated by denitrification, which maintains
the DIN inflow at low values between late spring and early autumn
(Fig. 10A-a). In the estuary, DIN transformation is controlled by
aerobic respiration and denitrification which are of similar absolute
magnitude in terms of carbon but result in a net consumption of DIN.
This net consumption is amplified by the DIN uptake during the
spring, and to a lesser extent, during the summer diatom bloom
(Figs. 10A-c and 9-a). SiðOH Þ4 consumption correlates with the
primary production signal and thus reveals a pronounced peak in
spring followed by smaller peaks in summer (Fig. 10B-c). Silica
dissolution partly compensates the uptake during the summer bloom
and results in slightly reduced net-consumption rates during this
period. PO4 transformation is dominated by net production, due to
organic matter degradation in the estuarine domain (Fig. 10C-c). Yet,
a net consumption is simulated during the spring diatom bloom in the
saline estuary when diatom PO4 uptake largely exceeds PO4
regeneration. The nutrient riverine inflow is thus significantly
modified by biogeochemical processes during the estuarine transit
and low-pass filtered nutrient fluxes across the estuarine mouth differ
markedly from the fluxes at the upstream boundary (Fig. 10A–C-d).
This variability is however not only the result of the evolution in the
magnitude of biogeochemical processes, but also results from flux
imbalances, which arise from the time-lagged response of the nutrient
scalar fields to hydrological perturbations (Regnier and Steefel, 1999;
Arndt et al., 2009). The magnitude of this imbalance strongly depends
on the specific shape of the respective nutrient scalar fields. The
increase in residual transport driven by the pronounced winter
discharges results in a storage of SiðOHÞ4 (Fig. 10B-b). During winter
and spring, SiðOH Þ4 fluxes (Fig. 10B-d) therefore show a damped and
smeared response to input variations (Fig. 10B-a). The time-lagged
release of the stored SiðOH Þ4 initially supports the spring diatom
bloom in the coastal zone. However, the enhanced nutrient
consumption, due to the coastal diatom bloom intruding into the
estuary in late spring leads to a pronounced SiðOHÞ4 and PO4 mass
decrease in the estuarine domain (Fig. 10B-b and C-b). During the
terminal stage of the estuarine diatom bloom, the lower estuary
becomes completely depleted in SiðOHÞ4 and PO4 (Fig. 7) and a small
influx of these elements occurs across the estuarine mouth (Fig. 10B-d
and C-d). In early summer, estuarine SiðOH Þ4 and PO4 pools are partly
replenished by the influx through the upper boundary in combination
with net PO4 in-situ production (Fig. 10B and C). Therefore, a net
SiðOH Þ4 and PO4 export to the coastal zone is again simulated after the
decline of the estuarine bloom. Yet, in the case of SiðOHÞ4 the estuarine
summer phytoplankton bloom develops at the time when the riverine
influx is low, resulting in another period of SiðOH Þ4 import from the
adjacent coastal zone in late summer (Fig. 10B-d and C-d). DIN fluxes
are less affected by the spring bloom, which results in a relatively
large DIN outflux to the coastal zone (Fig. 10A-d). Later in the year, the
DIN summer consumption significantly depletes the DIN inventory in
the estuarine domain and results in a decreasing net flux, which
reaches a minimum in late summer (Fig. 10A-b to -d).
The adjacent coastal zone is characterized by comparatively long
residence times that allow intense biogeochemical transformation
processes (Fig. 10-e to -g). The nutrient uptake during the early stage
of the spring phytoplankton bloom leads to a complete depletion of
the coastal SiðOH Þ4 and PO4 levels and reduces the DIN inventory
(Fig. 10A–C-e). This reduction triggers a net nutrient influx through
60
S. Arndt et al. / Journal of Marine Systems 84 (2011) 49–66
A) DIN
b
a
C) PO4
g
d
c
B) Si(OH)4
e
f
S. Arndt et al. / Journal of Marine Systems 84 (2011) 49–66
the open sea boundaries (Fig. 10A–C-g). However, the increased
availability of labile organic matter and biogenic silica after the
decline of the spring bloom fuels an intense nutrient recycling
(Fig. 10A–C-f). As a result, the coastal SiðOH Þ4 inventory slowly
increases, resulting in a net outflux of this nutrient (Fig. 10B-e to -g).
The intense consumption of SiðOHÞ4 during the coastal diatom bloom
in summer leads again to low levels (Fig. 10B-e to -g) and, thus, to an
influx through the open boundaries. In the case of DIN, the uptake by
Phaeocystis, combined with estuarine export fluxes, maintain low
coastal stocks (Fig. 10A-e). A net DIN influx is thus simulated during
the summer month, until the flux variability increases in autumn due
to the intensification of the meteorological forcing (Fig. 10A-g). PO4
transformations show a similar pattern (Fig. 10C-f). Yet, unlike DIN,
persistent PO4 influxes through the open sea boundaries and the
estuarine mouth result in an increase in PO4 stocks after the early
spring bloom (Fig. 10C-d, -e and -g).
The evolution of nutrient limitation factors provides important
insights into the functioning of the phytoplankton succession in the
coastal zone. Fig. 11 summarizes the seasonal evolution of nutrient
limitations for the dominant estuarine and coastal phytoplankton
groups. The timing of the respective phytoplankton concentration
maxima is indicated by dashed lines. Nutrient limitation in the
estuarine and coastal-zone sub-domains is assessed by calculating the
N
species-specific limitation factors, formulated as 1−
with
N + KN
N = PO4 ; DIN; SiðOH Þ4 , based on specific half-saturation constants
(KN, Table 1) and volume-averaged concentrations, N. Low values
indicate no or little nutrient limitation, while high values reflect a
strong limitation of primary production. Thus, they represent an
integrative measure of the control of nutrient availability on
phytoplankton succession. At its early stages, the spring diatom
bloom is mainly limited by PO4 availability. Yet, the exponential
phytoplankton growth almost completely depletes SiðOHÞ4 (Fig. 10B-e),
causing a shift towards a silica limitation around day 100 (Fig. 11-a).
Similarly, Van der Zee and Chou (2005) report, based on nutrient
measurements at different stations in the Belgian coastal zone, a PO4
limitation of the spring diatom production in the southern and offshore
parts of the coastal zone, co-limited by SiðOHÞ4 close to the estuarine
mouth. Previous modeling studies show similar trends even though the
predicted PO4 limitation was weaker (0.3 for 1995) for the spring
phytoplankton bloom and simulated spring SiðOHÞ4 concentrations
were generally overestimated (Lancelot et al., 2007; Gypens et al.,
2007). The intense coastal diatom bloom depletes the coastal PO4
inventory (Fig. 10C-e), which results in a strong PO4 limitation of
phytoplankton growth. The simulation predicts a PO4 limitation (0.4) of
the spring Phaeocystis colony bloom, which is in agreement with
previous modeling results (Gypens et al., 2007). However, the transient
simulation results reveal that this limitation is short-lived, due to the
very fast recovery of PO4 inventories by transport and in-situ
regeneration (Fig. 11-b). Therefore, the production of Phaeocystis
colonies, which are also able to use organic phosphorous for growth,
is virtually unaffected by low PO4 concentrations (Figs. 10C-e, 11-b). A
moderate DIN limitation (around 0.3) is simulated at the late stage of the
bloom (Fig. 11-b). Later in the year, the nutrient influxes through the
boundaries and the in-situ recycling increase the nutrient availability in
the coastal zone (Fig. 10A to C). The coastal summer diatom bloom is
thus weakly limited by nutrient availability (Fig. 11-a), indicating that
temperature, light conditions and grazing are more important controlling factors. In the estuary, the intruding coastal spring diatoms, which
have comparably low nutrient requirements, encounter excellent
growth conditions. Therefore, the estuarine spring bloom is characterized by a weak nutrient limitation factors close to 0.1 (Fig. 11-c). Salinity
61
stress, combined with the decreasing availability of light in the turbid
estuary put more severe environmental constraints on the intruding
coastal diatoms and limit their upstream propagation (Fig. 7, Van
Damme et al. (2005)). The freshwater diatoms, which are imported
from the tidal river during the summer months, are adapted to nutrientrich conditions and, therefore, are seriously limited by the decrease in
estuarine nutrient levels along the salinity gradient (Fig. 11-d). The
summer freshwater diatom bloom thus reveals a strong co-limitation of
SiðOH4 Þ and PO4 . DIN concentrations are less affected by dilution and
consumption within the estuarine domain and, as a consequence, the
limitation by this nutrient is weaker.
3.6. Patterns in ecosystem structure
The residual transport field and in-situ turnover rates control the
local nutrient availabilities and the emergence of distinct spatial
patterns in ecosystem structure throughout the productive period.
Fig. 12 illustrates the spatio-temporal evolution of the dominant
phytoplankton groups and bacteria along the estuary-coastal zone
continuum. They show a general pattern in agreement with previous
field observations (Nihoul and Hecq, 1984; Borges and Frankignoulle,
1999, 2002; Muylaert et al., 2006). This pattern is characterized by
high primary production rates in the estuarine plume, the development of zooplankton along the deflected plume and an intense
recycling around the coastal residual gyre (Nihoul and Hecq, 1984).
Simulation results indicate that the coastal spring phytoplankton
bloom is spatially constrained by the species-specific nutrient
requirements. Even though the early spring diatom bloom is initiated
in the southern part of the Belgian coastal zone, highest biomass
rapidly clusters around the nutrient-rich Scheldt plume, where
SiðOH Þ4 is available. The bloom thus follows the movement of the
estuarine plume along the coast line and maximum biomasses are
simulated between Vlissingen and Zeebrugge (Fig. 12-a, stage 1). The
exponential phytoplankton growth in late March progressively
depletes coastal SiðOH Þ4 and PO4 levels which ultimately become
limiting for diatoms (Fig. 11-a). At the same time, tidal pumping,
caused by the asymmetry between flood and ebb currents, transports
euryhaline diatoms upstream into the estuary, where they encounter
good growth conditions. The enhanced estuarine diatom production
significantly reduces the nutrient fluxes across the estuarine mouth
(Fig. 10-d), exerting a negative feedback on the coastal diatoms and
further limiting their production to the estuarine domain (Fig. 12-a,
stage 2). The release of labile organic matter during the decline of the
spring phytoplankton blooms stimulates bacterial growth and the
recycling of DIN and PO4 in the coastal zone (Fig. 10A and C).
Maximum bacterial biomass drifts with the residual gyre (Fig. 12-c,
stage 1). In addition, the residual transport field is characterized by a
decrease in velocities between Oostende and Zeebrugge, which leads
to an accumulation of organic matter close to Oostende (results not
shown) and the formation of remineralization patches of DIN and, to a
lesser extent, PO4 . Field observations are consistent with our
simulations and show a spring minimum in O2 saturation, which
coincides with maxima in phaeopigment concentrations, organic
carbon and pCO2 oversaturation in this area (Nihoul and Hecq, 1984;
Borges and Frankignoulle, 1999, 2002; Van der Zee and Chou, 2005).
Low DIN concentrations moderately limit Phaeocystis colony growth
in the coastal domain (Fig. 11-c). In early spring, the maximum in
Phaeocystis colony biomass is situated close to the open boundaries
(Fig. 12-b, stage 1), where they benefit from the inflow of nutrientrich waters from the English channel and, occasionally, from the Rhine
plume (Lacroix et al., 2004). Yet, the progressive nutrient depletion of
Fig. 10. Annual evolution of nutrient cycling along the estuary–coastal zone (A dissolved inorganic nitrogen, B silica, C phosphate) a—estuarine influx, b—instantaneous (grey) and
low-pass filtered (black) mass in the estuary, c—low-pass filtered transformation processes in the estuary, d—low-pass filtered flux at the estuarine mouth, e—instantaneous (grey)
and low-pass filtered (black) mass in the coastal zone, f—low-pass filtered net transformation processes in the coastal zone and g—low pass filtered net flux through the open sea
boundaries.
62
S. Arndt et al. / Journal of Marine Systems 84 (2011) 49–66
Fig. 11. Annual evolution of species-specific nutrient limitation functions 1− N +N KN with N = DIN; PO4 ; SiðOHÞ4 for a—diatom, and b—Phaeocystis production in the coastal zone and
the c—coastal diatom and d—estuarine diatom production in the estuary. The dashed lines indicate the onset and the termination of the respective blooms.
coastal waters, especially with respect to SiðOH Þ4 and PO4 (Figs. 10
and 11) ultimately favors the onset of a Phaeocystis colony bloom
along the coast (Fig. 12-b, stage 2). During this period, the estuarine
nutrient fluxes to the coastal zone are strongly reduced by the diatom
consumption in the lower estuarine reaches and DIN, as well as PO4
are mainly supplied by import through the open boundaries, in
combination with local DIN recycling (Fig. 10A and C). Therefore,
maximum Phaeocystis biomass clusters along the outer edge of the
estuarine bloom and oscillates between the area of enhanced nutrient
recycling close to Zeebrugge and the north-eastern part of the coastal
domain, which occasionally profits from the southward spreading of
the Rhine plume (Fig. 12-b, stage 2). Observational evidence confirms
the dominance of Phaeocystis colonies in the outer Scheldt plume,
while diatoms and Phaeocystis equally contribute to the Chl a spring
concentrations close to the estuarine mouth (Muylaert et al., 2006). At
the end of May, nutrient depletion leads to a decline of the spring
diatom and Phaeocystis blooms (Fig. 12). Nevertheless, Phaeocystis
colonies are still imported through the northern boundary (Fig. 12-b,
stage 3). During summer, microbially-mediated heterotrophic processes drive a nutrient recycling loop around the coastal residual gyre
(Figs. 10-f and 12-c, stage 2).
Concomitantly, a summer diatom bloom develops in the freshwater reaches of the Scheldt estuary (Muylaert et al., 2000; Meire et al.,
2005) and leads to a significant import of freshwater diatoms by the
residual estuarine downstream transport (Fig. 12-d). Yet, these species
are generally adapted to high nutrient and light conditions (Muylaert
et al., 2005) and phytoplankton growth is thus limited to the area
upstream of the Dutch/Belgian border (Fig. 12-d). The development of
this summer diatom bloom further reduces the summer nutrient
export fluxes that are already low due to the summer decrease in river
discharge (Fig. 10-a and -d). Therefore, the coastal summer diatom
bloom starts in the southern coastal zone, where low turbidities
(Fig. 6) and intense nutrient recycling favour production (Fig. 12-d,
stage 1). In late summer, the decline of the freshwater diatom bloom
and the associated increase in nutrient export fluxes leads to a shift of
maximum coastal diatom biomass towards the nutrient-rich estuarine
plume (Fig. 12-d, stage 2).
Simulation results on nutrient dynamics and ecosystem structure
reveal that the spatio-temporal variability in the distribution and
transformation of carbon and nutrients is strongly controlled by the
constraints set by the physical environment. The estuary is characterized by a highly-transient one-dimensional spatial structure, which is
controlled by the residual downstream transport. The land-derived
nutrient export to the coastal zone shows a strong seasonal variability,
which results from the evolution in the magnitude of riverine inputs and
estuarine biogeochemical transformations. In addition, large transient
fluctuations in estuarine nutrient inventories arising from the timelagged response of the scalar fields to upstream hydrological perturbations influence the timing of the nutrient delivery to the coastal zone.
The coastal zone, on the other hand, exhibits a complex twodimensional pattern. This pattern is mainly controlled by the Lagrangian
residual transport field, whose complex structure determines the
distribution of nutrient-rich estuarine waters and defines the residence
times. High primary production rates are observed in the nutrient-rich
estuarine plume and drive the development of bacteria and higher
trophic levels along the plume. The results show also that the Lagrangian
residual transport field is essentially unsteady and subject to continuous
hydrodynamic adjustments, which lead to large variations in the
spreading of the estuarine plume and, therefore, in the overall structure
of the coastal ecosystem over short periods of time. In addition, the
formation of a well-developed residual gyre increases the residence
times in the nearshore coastal zone and favors an enhanced nutrient
recycling in the vicinity of the gyre. The specific nutrient requirements of
the various phytoplankton groups further constrain the detailed spatial
S. Arndt et al. / Journal of Marine Systems 84 (2011) 49–66
structure of the ecosystem. Maximum diatom concentrations develop in
the immediate proximity of the estuarine mouth where nutrient
concentrations are generally higher than elsewhere in the coastal
zone. This bloom progressively depletes PO4 and SiðOH Þ4 at the interface
between the estuary and the coastal zone. This local depletion strongly
limits the development of Phaeocystis in the vicinity of the estuarine
mouth. Therefore, they preferentially develop in areas where enhanced
63
recycling rates or import fluxes through the open sea boundaries
provide suitable growth conditions.
4. Conclusions
A mechanistic model accounting for the coupled biogeochemical
transformation processes and fluxes of carbon and macro-nutrients
Fig. 12. Spatio-temporal dynamics of daily maximum phytoplankton biomasses and bacteria during the productive period in spring and summer 1995. a—spring diatoms,
b—Phaeocystis, c—bacteria, and d—summer diatoms.
64
S. Arndt et al. / Journal of Marine Systems 84 (2011) 49–66
Fig. 12 (continued).
along the entire mixing zone of the shallow, tidally-dominated
Scheldt estuary and adjacent coastal zone has been developed. The
spatial and temporal resolution of the model allow to include the
short-term transient forcings triggered by tides, freshwater inflows
and wind stress, thus providing a detailed representation of the
hydrodynamic and transport fields along a topographically-complex
system. The results give important insights into the processes and
environmental factors that control the spatial structure and evolution
of autotrophic and heterotrophic processes affecting the C, N, P and Si
cycles at the intra-annual timescale.
The land-derived nutrient export to the coastal zone is characterized by a strong seasonal variability, which results from the evolution
of riverine inputs as well as large transient fluctuations in estuarine
nutrient inventories and biogeochemical transformation processes.
The spatial pattern of the ecosystem in the coastal zone is strongly
controlled by the constraints set by the local physical environment.
The latter are mainly determined by the Lagrangian residual transport
field, whose complex structure control the location and residence
time of land-derived, nutrient-rich waters within the coastal zone
and, thus, areas of high primary production. Furthermore, the
unsteady nature of the Lagrangian residual transport field and its
continuous hydrodynamic adjustments lead to large variations in the
spreading of the estuarine plume and, therefore, in the overall
structure of the coastal ecosystem over short periods of time. The
formation of a residual gyre is another distinct feature of the local
physics. It increases the residence times in the nearshore coastal zone
and favors an enhanced nutrient recycling in the vicinity of the gyre.
The specific nutrient requirements of the various phytoplankton
species further control the spatial structure of the ecosystem.
The interplay between estuarine nutrient inputs and physical
constraints suggests that optimum phytoplankton growth conditions
are encountered near the dynamic, yet poorly-surveyed, estuarine
mouth. Marked spatial concentration gradients develop in this narrow
zone and episodically lead to a reversal of the material fluxes, from the
coast into the estuary. Therefore, the estuary does not operate
independently from processes in the coastal zone. Euryhaline coastal
diatoms, for instance, intrude far upstream into the saline estuary
where favorable growth conditions are encountered during distinct
episodes of the productive period. This intrusion reduces the
estuarine nutrient concentrations and export fluxes, thereby reinforcing the nutrient limitation in the coastal area. Furthermore, the
mixture of marine and estuarine phytoplankton may sustain an
intense carbon recycling loop. The interdependency of the biogeochemical dynamics in these two ecosystems and the high process
rates which occur at their transition strongly supports the continuum
approach adopted here. In the past, monitoring programs and
process-based laboratory studies have nevertheless largely restricted
their domain of investigation to either the estuary or the coastal zone.
As a consequence, the highly dynamic estuary–coastal zone interface
has been poorly surveyed and comprehensive data sets for the
transition zone are lacking. An improved description of the spatial
structure of coastal, estuarine and freshwater phytoplankton communities from the shallow, nutrient-rich estuary to the deeper,
nutrient-starved coastal zone is thus hampered by the lack of
monitoring strategies encompassing the entire estuary–coastal zone
ecosystem. Process-based experiments investigating the spatial
variability in rate constants and model parameters of the carbon
and nutrient cycles along the continuum are also essentially missing.
In the future, joint observational and monitoring programs could help
advance our understanding of the pronounced and fast temporal
variations in biogeochemical dynamics along the continuum.
The design and the parameterization of the coupled model mainly
build on the mechanistic understanding of the biogeochemistry in the
Scheldt estuary and the adjacent coastal zone. It could thus be applied
with minimum parameter adjustment to neighboring systems in the
North Sea or other systems with similar physical and climatological
characteristics. However, the set-up of the fully-coupled estuary–
coastal zone model requires a detailed bathymetry, transient wind
field maps and a comprehensive set of boundary conditions,
compromising therefore its application to poorly surveyed environments. Although the computational time (ca. 3 days per simulated
year) does not allow for an extensive sensitivity analysis, it allows for
simulations at inter-annual timescales if a data set is available to
constrain boundary conditions and external forcings over these
periods. Therefore, the coupled model is ideally suited to explore
the effect of global change and/or water quality management on
coastal eutrophication. It could also provide useful guidelines for the
incorporation of simplified representations of the land–ocean transition in biogeochemical models of the global coastal ocean.
Acknowledgements
This manuscript has benefited from the constructive and helpful
reviews of two anonymous reviewers. The presented work has been
financially supported by the Netherlands Organization for Scientific
Research (NWO) (RUBICON grant to SA and VIDI grant to PR) and by
the government of the Brussels-Capital region (Brains Back to Brussels
award to PR). This work is a contribution to the TIMOTHY (P6/13) and
AMORE III Projects funded respectively by the Interuniversity
Attraction Poles Programme and by the Science for Sustainable
S. Arndt et al. / Journal of Marine Systems 84 (2011) 49–66
Development Programme of the Belgian State-Belgian Science Policy.
Nathalie Gypens has a post-doctoral fellowship funded by the Belgian
ÎFonds National de la Recherche Scientifique-FNRS.
References
Allen, J.I., Blackford, J., Holt, J., Proctor, R., Asworth, M., Siddorn, J., 2001. A highly
spatially resolved ecosystem model for the North West European continental shelf.
Sarsia 86, 423–440.
Allen, J.I., Holt, J., Blackford, J., Proctor, 2007. Error quantification of a high-resolution
coupled hydrodynamic-ecosystem coastal-ocean model: Part 2. Chlorophyll-a,
nutrients and SPM. Journal of Marine Systems 68, 381–404.
Ariathurai, C.R., 1974. Finite element model for sediment transport in estuaries. Phd
thesis, University of California.
Arndt, S., Vanderborgth, J.-P., Regnier, P., 2007. Diatom growth response to physical
forcing in a macrotidal estuary: coupling hydrodynamics, sediment transport and
biogeochemistry. Journal of Geophysical Research 12 doi:10.1029/2006JC003581.
Arndt, S., Regnier, P., Vanderborght, J.P., 2009. Seasonally-resolved nutrient filtering
capacities and export fluxes in a macrotidal estuary. Journal of Marine Systems 78,
42–58.
Baretta, J.W., Ebenhoh, W., Ruardij, P., 1995. The European regional seas ecosystem
model, a complex marine ecosystem model. Netherlands Journal of Sea Research
33.
Billen, G., Garnier, J., 1997. The Phison River plume: coastal eutrophication in response
to changes in land use and water management in the watershed. Aquatic Microbial
Ecology 13, 3–17.
Borges, A.V., Frankignoulle, M., 1999. Daily and seasonal variations of the partial
pressure of Co2 in surface seawater along Belgian and Southern Dutch coastal areas.
Journal of Marine Research 19, 251–266.
Borges, A.V., Frankignoulle, 2002. Distribution and air–water exchange of carbon
dioxide in the Scheldt plume off the Belgian coast. Biogeochemistry 59, 41–67.
Brandt, L.E., 1984. The salinity tolerance of forty-six marine phytoplankton isolates.
Estuarine, Coastal and Shelf Science 18, 543–556.
Breton, E., Rousseau, V., Parent, J.-Y., Ozer, J., Lancelot, C., 2006. Hydroclimatic
modulation of diatom/Phaeocystis blooms in the nutrient-enriched Belgian coastal
waters (North Sea). Limnology and Oceanography 51, 1401–1409.
Cadee, G.C., Hegeman, J., 1991. Historical phytoplankton data of the Marsdiep.
Hydrobiological Bulletin 24, 111–118.
Chen M.S., 2003. Suspended matter and flocculation in the estuarine environment. Ph.D
Thesis, Vrije Universiteit Brussel.
Chen, M.S., Wartel, S., Van Eck, B., Van Maldegem, D.C., 2005. Suspended matter in the
Scheldt estuary. Hydrobiologia 540, 79–104.
Cloern, J.E., 2001. Our evolving conceptual model of the coastal eutrophication problem.
Marine Ecology Progress Series 210, 223–253.
Cugier, P., Billen, G., Guillaud, J.F., Garnier, J., Menesguen, A., 2005. Modelling of
eutrophication of the Seine Bight (France) under historical, present and future
riverine nutrient loading. Journal of Hydrology 304, 381–396.
de Galan, S., Elskens, M., Goeyens, L., Pollentier, A., Brion, N., Baeyens, W., 2004. Spatial
and temporal trends in nutrient concentrations in the Belgian continental area of
the North Sea during the period 1993–2000. Estuarine, Coastal and Shelf Science
61, 517–528.
Delhez, E.J.M., 1998. Macroscale ecohydrodynamic modeling on the northwest
European continental shelf. Journal of Marine Research 16, 171–190.
Delhez, E.J.M., Carabin, G., 2001. Integrated modelling of the Belgian coastal zone.
Estuarine, Coastal and Shelf Science 53, 477–491.
Desmit, X., Vanderborght, J.P., Regnier, P., Wollast, R., 2005. Control of primary
production by physical forcing in a strongly tidal, well-mixed estuary. Biogeosci. 2,
205–218.
Einstein, H.B., Krone, R.B., 1962. Experiments to determine modes of cohesive sediment
transport in salt water. Journal of Geophysical Research 67 (4).
Ekebjaerg, L., Justesen, P., 1991. An explicit scheme for advection–diffusion modelling in
two dimensions. Computer Methods in Applied Mechanics and Engineering 88, 3–8.
Fettweis, M., Van den Eynde, D., 2003. The mud deposits and the high turbidity in the
Belgian–Dutch coastal zone, Southern Bight of the North Sea. Continental Shelf
Research 23, 669–691.
Fettweis, M., Francken, F., Pison, V., Van den Eynde, D., 2006. Suspended particulate matter
dynamics and aggrgate sizes in a high turbidity area. Marine Geology 235, 63–74.
Gullentops, F., Moens, M., Ringele, A., Sengier, R., 1976. Geologische kenmerken van de
suspensie en de sedimenten. In: Nihoul, J., Gullentops, F. (Eds.), Project Zee- Projet
Mer., vol. 4. Science Policy Office, Brussles, Belgium, pp. 1–137.
Gypens, N., Lancelot, C., Borges, A.V., 2004. Carbon dynamics and CO2 air–sea exchanges
in the eutrophicated coastal waters of the southern bight of the North Sea: a
modelling study: a modelling study. Biogeosciences 1, 147–157.
Gypens, N., Lacroix, G., Lancelot, C., 2007. Causes of variability in diatom and Phaeocystis
blooms in Belgian coastal waters between 1989 and 2003: a model study. Journal of
Sea Research 57, 19–35.
Gypens, N., Borges, A.V., Lancelot, C., 2009. Effect of eutrophication on air–sea Co2 fluxes
in the coastal Southern North Sea: a model study of the past 50 years. Global
Change Biology 15, 1040–1056.
Horrevoets, A.C.H., Borges, H.H.G., Schuurman, J.N., Graas, S., 2004. The influence of river
discharge on tidal damping in alluvial estuaries. Journal of Hydrology 294, 213–228.
Hydes, D.J., Kelly-Gerreyn, B.A., Proctor, R., Thomson, S., 1996. An analysis of the
nitrogen flow through the Southern North Sea from a coupled 2-D physical/
biogeochemical model. In: Andersen, J., Karup, H., Neilsen, U.B. (Eds.), Proceedings
65
of the Scientific Symposium on the North Sea Quality Status Report 1993, pp. 54–61.
Ebletoft, Denmark.
Killops, S.D., Killops, V.J., 1993. An Introduction to Organic Geochemistry. Longman
Group, London.
Koeman, R., Rademaker, M., Buma, A., 1992. Biomonitoring van phytoplankton in de
Nederlandse zoute en brakke wateren 1991. Tech. rep. National Institute for Coastal
and Marine Management (RIKZ).
Kromkamp, J., Peepe, J., 1995. Possibility of net phytoplankton primary production in
the turbid Schelde estuary (SW Netherlands). Marine Ecology Progress Series 121,
249–259.
Lacroix, G., Ruddick, K., Ozer, J., Lancelot, C., 2004. Modelling the impact of the Scheldt
and Rhine/Meuse plumes on the salinity distribution in Belgian waters (Southern
North Sea). Journal of Sea Research 52, 149–163.
Lacroix, G., Ruddick, K., Gypens, N., Lancelot, C., 2007a. Modelling the relative impact of
rivers (Scheldt/Rhine/Seine) and western channel waters on the nutrient and
diatoms/Phaeocystis distributions in Belgian waters (Southern North Sea).
Continental Shelf Research 27, 1422–1446.
Lacroix, G., Ruddick, K., Park, Y., Gypens, N., Lancelot, C., 2007b. Validation of the 3-D
biogeochemical model MIRO & Co with field nutrient and phytoplankton data and
MERIS-derived surface chlorophyll a images. Journal of Marine Systems 64, 66–88.
Lancelot, C., 1995. The mucilage phenomenon in the continental coastal waters of the
North Sea. The Science of the Total Environment 165, 83ß102.
Lancelot, C., Billen, G., Sournia, A., Weisse, T., Colijn, F., Veldhuis, M.J.W., Davies, A.,
Wassman, P., 1987. Phaeocystis blooms and nutrient enrichment in the continental
coastal zones of the North Sea. Ambio 16, 38–46.
Lancelot, C., Keller, M.D., Rousseau, V., Smith, W.O., Mathot, S., 1998. Auto-ecology of
the marine haptophyte Phaeocystis. In: Anderson, D.M., Cembella, A.D., Hallegraeff,
G.M. (Eds.), Physiological ecology of harmfull algal blooms. NATO ASI Series.
Springer Verlag, Berlin.
Lancelot, C., Spitz, Y., Gypens, N., Ruddick, K., Becquevort, S., Rousseau, V., Lacroix, G.,
Billen, G., 2005. Modelling diatom and Phaeocystis blooms and nutrient cycles in the
Southern Bight of the North Sea: the MIRO model. Marine Ecology Progress Series
289, 63–78.
Lancelot, C., Gypens, N., Billen, G., Garnier, J., Roubeix, V., 2007. Testing an integrated
riverÐocean mathematical tool for linking marine eutrophication to land use: the
Phaeocystis-dominated Belgian coastal zone (Southern North Sea) over the past
50 years. Journal of Marine Systems 64, 216–228.
Leendertse, J.J., 1967. Aspects of a computational model for long water wave
propagation. Rand memorandum, RH-5299-PR. .
Luyten, P., Jones, K., Proctor, R., Tabor, P., Tett, P., Wild-Allen, K., 1999. Coherens
documentation: a coupled hydrodynamical–ecological model for regional and shelf
seas: user documentation. Tech. rep. MUMM.
Meire, P., Ysebaert, T., Van Damme, S., Van den Bergh, E., Maris, T., Struyf, E., 2005. The
Scheldt estuary: a description of a changing ecosystem. Hydrobiologia 540, 1–11.
Mills, D.K., Tett, P.B., Novarino, G., 1994. The spring bloom in the south western North
Sea in 1989. Netherlands Journal of Sea Research 33, 65–80.
Moll, A., 1998. Regional distribution of primary production in the North Sea simulated
by a three-dimensional model. Journal of Marine Systems 16, 151–170.
Muylaert, K., Sabbe, K., 1999. Spring phytoplankton assemblages in and around the
maximum turbidity zone of the estuaries of the Elbe (Germany), the Schelde
(Belgium/The Netherlands) and the Gironde (France). Journal of Marine Systems
22 (2–3), 133–149.
Muylaert, K., Koen, S., Vyverman, T., 2000. Spatial and temporal dynamics of
phytoplankton communities in a freshwater tidal estuary. Estuarine, Coastal and
Shelf Science 50, 673–687.
Muylaert, K., Tackx, M.L., Vyeverman, W., 2005. Phytoplankton growth rates in the
freshwater tidel reaches of the Sschelde estuary (Belgium) estimated using a
simple light-limited primary production model. Hydrobiologia 540, 127–140.
Muylaert, K., Gonzales, R., Franck, M., Lionard, M., Van der Zee, C., Cattrijsee, A., Sabbe,
K., Chou, L., 2006. Spatial variation in phytoplankton dynamics in the Belgian
coastal zone of the North sea studied by microscopy, HPLC-chemtax and underway
fluorescence recordings. Journal of Sea Research 55, 253–265.
Muylaert, K., Sabbe, K.L., Vyeverman, W., 2009. Changes in phytoplankton diversity and
community composition along the salinity gradient of the Schelde estuary
(Belgium/The Netherlands). Estuarine, Coastal and Shelf Science 82, 335–340.
Neal, C., Jarvie, H.P., Whitton, B.A., Gemmell, J., 2000. The water quality of the river
Wear, north–east England. The Science of the Total Environment 2, 251–252.
Nihoul, J., 1975. Effect of tidal stress on residual circulation and mud deposition in the
Southern Bight of the North Sea. Review of Pure and Applied Geophysics 113,
577–591.
Nihoul, J., Hecq, J.H., 1984. Influence of the residual circulation on the physico-chemical
characteristics of water masses and the dynamics of ecosystems in the Belgian
coastal zone. Continental Shelf Research 3, 167–174.
Partheniades E., 1962. A study of erosion and deposition of cohesive soils in salt water.
Ph.D. Thesis, University of California.
Patsch, J., Radach, G., 1997. Long-term simulation of the eutrophication of the North
Sea: temporal development of nutrients, chlorophyll and primary production in
comparison to observations. Journal of Sea Research 38, 275–310.
Pingree, R.D., Maddock, L., 1977. Tidal residual in the English channel. Journal of Marine
Biological Association U.K. 57, 339–354.
Press, W.H., Teukolsky, S.A., Vetterling, W.T., Flannery, B.P., 1992. Numerical Recipies in
C: The Art of Scientific Programming. Cambridge University Press, England.
Proctor, R., Holt, J., Harris, J.R.W., Tappin, A.D., Boorman, D., 2000. Modeling the Humber
estuary catchment and coastal zone. In: Spaulding, M., Butler, H.L. (Eds.), Estuarine
and Coastal Modeling: Proceedings of the 6th International Conference. American
Society of Civil Engineers, pp. 1259–1274.
66
S. Arndt et al. / Journal of Marine Systems 84 (2011) 49–66
Regnier, P., Steefel, C.I., 1999. A high resolution estimate of the inorganic nitrogen flux
from the Scheldt estuary to the coastal North Sea during a nitrogen-limited algal
bloom, spring 1995. Geochimica et Cosmochimica Acta 63 (9), 1359–1374.
Regnier, P., Wollast, R., Steefel, C.I., 1997. Long-term fluxes of reactive species in
macrotidal estuaries: estimates from a fully transient, multicomponent reactiontransport model. Marine Chemistry 58, 127–145.
Regnier, P., Mouchet, A., Wollast, R., Ronday, F., 1998. A discussion of methods for
estimating residual fluxes in strong tidal estuaries. Continental Shelf Research 18,
1543–1571.
Riegman, R., Noordeloos, A.A.M., Cadee, G.C., 1992. Phaeocystis blooms and eutrophication of the continental coast zones of the North Sea. Marine Biology 112,
479–484.
Rijstenbil, J.W., Bakker, C., Jackson, R.H., Merks, A.G.A., Visscher, P.R.M., 1993. Spatial
and temporal variation in community composition and photosynthetic characteristics of phytoplankton in the upper Westerschelde estuary (Belgium, SW
Netherlands). Hydrobiologia 269/270, 263–273.
Roubeix, V., Rousseau, V., Lancelot, C., 2008. Diatom succession and silicon removal
from freshwater in estuarine mixing zones: from experiment to modelling.
Estuarine, Coastal and Shelf Science 78, 14–26.
Rousseau V., 2000. Dynamics of Phaeocystis and diatom blooms in the eutrophicated
coastal waters of the Southern Bight of the North Sea. Ph.d. Thesis, Université Libre
de Bruxelles.
Rousseau, V., Becquevort, S., Parent, J.-Y., Gasparini, S., Daro, M.H., Tackx, M.L., Lancelot,
C., 2000. Throphic efficiency of the planktonic food web in a coastal ecosystem
dominated by Phaeocystis colonies. Journal of Sea Research 43, 357–372.
Rousseau, V., Leynaert, A., Daoud, N., Lancelot, C., 2002. Diatom succession, silification
and silicic acid availability in Belgian coastal waters (Southern North Sea). Marine
Ecology Progress Series 236, 61–73.
Ruddick, K., Park, Y., Nechad, B., 2003. MERIS imagery of Belgian coastal waters:
mapping of suspended particulate matter and chlorophyll-a. Meris Users
Workshop, vol. SP-549. ESA Special Publication, Frascati.
Sabbe, K., Muylaert, K., Kromkamp, J., Vyverman, W., 2006. Primaire productie op het
kruispunt van estuarium en kustzone. Symposium: The Vlakte van de Raan
revisited, vol. 35. VLIZ Special Publication, Oostende, pp. 73–84.
Salomon, J.C., Breton, M., 1993. An atlas of long-term currents in the channel.
Oceanologia Acta 16, 439–448.
Skogen, M.D., Moll, A., 2005. Importance of ocean circulation in ecological modelling: an
example from the North Sea. Journal of Marine Systems 57, 289–300.
Skogen, M.D., Soiland, H., 1998. A user's guide to Norwecom V2, the Norwegian
ecological model system. Tech. rep. Institute of Marine Research.
Skogen, M.D., Svendsen, E., Berntsen, J., Aksnes, D., Ulvestad, K.B., 1995. Modelling the
primary production in the North Sea using a coupled 3-dimensional physical
chemical biological ocean model. Estuarine, Coastal and Shelf Science 41, 545–565.
Smagorinsky, J., 1963. General circulation experiments with the primitive equations. i.
the basic experiment. Monthly Weather Review 91, 99–164.
Soetaert, K., Herman, P., 1995a. Nitrogen dynamics in the Westerschelde estuary (SW
Netherlands estimated by means of the ecosystem model MOSES. Hydrobiologia
311, 225–246.
Soetaert, K., Herman, P.M.J., 1995b. Carbon flows in the Westerschelde estuary (The
Netherlands) evaluated by means of an ecosystem model (MOSES). Hydrobiologia
311, 247–266.
Tappin, A.D., Harris, J.R.W., Uncles, R.J., 2003. The fluxes and transformations of
suspended particles, carbon and nitrogen in the Humber estuarine system (UK)
from 1994 to 1996: results from an integrated observation and modelling study.
The Science of The Total Environment 314–316, 665–713.
Turner, R.E., Rabalais, N.N., Justic, D., Dortch, Q., 2003. Global patterns of dissolved N, P
and Si in large rivers. Biogeochemistry 64, 297–317.
Van Damme, S., Struyf, E., Maris, T., Ysebaert, T., Dehairs, F., Tackx, M.L., Heip, C., Meire,
P., 2005. Spatial and temporal patterns of water quality along the estuarine salinity
gradient of the Scheldt estuary (Belgium and The Netherlands): results of an
integrated monitoring approach. Hydrobiologia 540, 29–45.
Van den Eynde, D., 2004. Interpretation of tracer experiments with fine-grained
dredging material at the Belgian continental shelf by the use of numerical models.
Journal of Marine Systems 48, 171–189.
Van der Zee, C., Chou, L., 2005. Seasonal cycling of phosphorus in the Southern Bight of
the North Sea. Biogeosciences 2, 27–47.
Vanderborght, J.P., Wollast, R., Loijens, M., Regnier, P., 2002. Application of a transportreaction model to the estimation of biogas fluxes in the Scheldt estuary.
Biogeochemistry 59, 207–237.
Vanderborght, J.P., Folmer, I., Aguilera, D.R., Uhrenholdt, T., Regnier, P., 2007. Reactivetransport modelling of a river-estuarine-coastal zone system: application to the
Scheldt estuary. Marine Chemistry 106, 92–110.
Vichi, M., Baretta, J.W., Baretta-Bekker, J.G., Ebenhoh, W., Kohlmeier, C., Ruardij, P.,
Pinardi, N., Zavatarelli, M., 2004. European regional seas ecosystem model III.
Review of the biogeochemical equations. Tech. rep., Documentation Release
Version: 3.4-Alpha Id:ERSEMIIIreport.lyx, v1.62004/07/2810:44:29mavExp.
Wirtz K.W., 1997. Modellierung von Anpassungsvorgaengen in der belebten Natur. Ph.d
Thesis, Universität Kassel.
Wirtz, K.W., Eckhardt, B., 1996. Effective variables in ecosystem models with an
application to phytoplankton succesion. Ecological Modelling 92, 33–53.