Geomorphology 191 (2013) 23–34

Contents lists available at SciVerse ScienceDirect

Geomorphology
journal homepage: www.elsevier.com/locate/geomorph

Modelling the effects of tidal range and initial bathymetry on the

morphological evolution of tidal embayments
B. van Maanen a, b,⁎, Giovanni Coco a, 1, K.R. Bryan b
a
National Institute of Water and Atmospheric Research (NIWA), PO Box 11-115, Hamilton, New Zealand
b
Department of Earth and Ocean Sciences, University of Waikato, Private Bag 3105, Hamilton, New Zealand

a r t i c l e i n f o a b s t r a c t

Article history: Tidal embayments are characterized by a wide variety of landscape features, often including either complex
Received 16 October 2012 tidal channel networks or extensive flood-tidal deltas. The origin of these features and the influence of hydro-
Received in revised form 24 January 2013 dynamic drivers and initial geological setting on their long-term characteristics are essentially unexplored. A
Accepted 24 February 2013
model was applied to simulate the long-term morphological evolution of tidal embayments, with the pur-
Available online 4 March 2013
pose of providing insight into the environmental conditions that lead to the differences in tidal embayment
Keywords:
morphology. Numerical simulations indicated that the interaction between hydrodynamics, sediment trans-
Tidal embayments port, and the evolving topography gives rise to the formation of channel networks. The tidal range and the
Channel networks depth of the initially unchannelized tidal basin controlled the way in which the morphology evolved and de-
Morphodynamics termined the timescale over which channels and intertidal areas developed. Channel network formation oc-
Numerical modelling curred more rapidly when the tidal range increased and/or when the initial basin depth decreased. Tidal
basins with a large initial depth showed the development of a flood-tidal delta and for these deep basins
channel incision could remain absent over long timescales. Both tidal range and initial bathymetry affected
final basin hypsometry and channel network characteristics, including the channel density and the fraction
of the basin occupied by the channels. All the simulated morphologies, with different combinations of the
tidal range and depth of the basin, evolved towards a state of less morphodynamic activity for which the rel-
ative intertidal area was proportional to the ratio of tidal amplitude to basin depth.
© 2013 Elsevier B.V. All rights reserved.

1. Introduction
Tidal embayments are amongst the most productive ecosystems
in the world (Mitsch and Gosselink, 2007) and understanding their
behavior has become increasingly important, especially in the context
of climate change. These environments are characterized by a wide
variety of landscape features and morphological patterns, with tidal
channel networks arguably being the most striking example (Fig. 1a).
The channel networks have a strong control on the hydrodynamics
and sediment transport (Fagherazzi et al., 1999) and affect therefore
both the short- and long-term morphological evolution of tidal environ-
ments (Dronkers, 2005). Because of the influential character of channel
networks, numerous studies (see de Swart and Zimmerman, 2009)
have been performed on their initial formation and subsequent evolution.
These advances are largely based on laboratory studies and numerical
models, because obtaining detailed measurements of channel network
⁎ Corresponding author at: University of Bordeaux 1, EPOC Laboratory, UMR CNRS 5805,
Avenue des Facultés, 33405 Talence, France. Tel.: +33 5 40 00 88 49; fax: +33 5 56 84 08
48.
E-mail address: b.vanmaanen@epoc.u-bordeaux1.fr (B. van Maanen).
1
Now at Environmental Hydraulics Institute, “IH Cantabria”, Universidad de Cantabria,
Avda. de los Castros s/n, Santander 39005, Santander, Spain.
dynamics in reality is a daunting task due to the large spatial and tempo-
ral scales involved.
Laboratory experiments have been conducted to explore the dy-
namics of single tidal channels (Tambroni et al., 2005) as well as the de-
velopment of complete tidal networks (Stefanon et al., 2010; Iwasaki et
al., 2011; Vlaswinkel and Cantelli, 2011). For example, Stefanon et al.
(2010) conducted an experiment starting from a plane horizontal
tidal flat subject to tidal forcing and showed how the headward growth
of the initiated channels and tributary addition drove expansion of the
network and shaped the intertidal morphology. In this experiment and
in the laboratory study described by Vlaswinkel and Cantelli (2011),
the networks which formed displayed geomorphic features, such as the
width-to-depth ratios and the seaward widening of channels, which re-
semble those of natural networks. Such laboratory approaches are mean-
ingful, but up-scaling the results to the time and space scales over which
real-world networks develop is not straightforward.
Simplified modelling approaches, in which processes are abstracted,
can provide useful insights into the key processes that control the forma-
tion of channel patterns and the evolution on larger scales. Schuttelaars
and de Swart (1999), for example, used stability analysis to study the ini-
tial formation of channels and shoals, and demonstrated that bottom
friction plays a crucial role in the growth of bedforms. Later, D'Alpaos
et al. (2005) applied a simplified hydrodynamic model and relationships

0169-555X/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.geomorph.2013.02.023
24 B. van Maanen et al. / Geomorphology 191 (2013) 23–34

a multi-inlet tidal system. Simulations were carried out with different

initial bathymetries (a sloping bed inside the basin as well as a flat
bed at different depths) and the model results followed empirical equi-
librium relations suggested in the literature.
These laboratory and modelling experiments have provided useful
insight into the processes governing channel network formation and
have highlighted the importance of the feedbacks between hydrody-
namics, sediment transport, and the evolving morphology. However,
there are many landscape features that characterize natural tidal basins
which are not reproduced by these network-forming experiments. In
some tidal basins, the flood-tidal delta may be the dominant feature
(de Swart and Zimmerman (2009) and see also Fig. 1b), with the net-
work being a minor or even an absent characteristic. Such tidal basins
often lack not only the channel network, but also the extensive adjacent
intertidal areas and it is especially these intertidal areas that can provide
valuable ecological habitats (Mitsch and Gosselink, 2007). If models are
ultimately to be of theoretical as well as of practical use, then reproduc-
ing the different landscape characteristics is critical. This study therefore
aims to provide insight into the environmental conditions that lead to
the observed differences in tidal embayment morphology. Tidal basins
evolve in sedimentary environments, and their landscapes may range
widely because of the large variety in tidal conditions and in the under-
lying shape of the basin. We hypothesize that the balance between the
initial depth of the basin and the strength of the tidal currents that redis-
tribute the sediment ultimately control the type of basin that eventuates.
Here we apply a numerical model based on the morphodynamic interac-
tions to perform controlled experiments on tidal embayment evolution
by varying both tidal range and initial depth to test this hypothesis.

2. Numerical modelling

2.1. Model description

The numerical model developed and used throughout this study

simulates morphological change as a result of the interactions between
Fig. 1. Aerial view of a (a) well developed (Bassin d'Arcachon, France) and (b) underdevel- hydrodynamics, sediment transport, and the evolving morphology.
oped (Shinnecock Inlet, USA) tidal channel network. Photo courtesy of Aquitaine Coast Fluid flow was simulated using ELCOM (Estuary and Lake Computer
Observatory/SIBA and “Inlets Online”, Coastal Inlet Research Program (CIRP), U.S. Army Model; Hodges et al., 2000) which is a 3D hydrodynamic model based
Corps of Engineers.
on the unsteady Reynolds-averaged Navier–Stokes equations. Although
a 3D-model was applied we only used the horizontal velocity compo-
describing channel characteristics to simulate tidal network initiation nents for the computation of sediment transport. Influences of Coriolis
and its progressive headward extension within tidal flats. This model force, density differences, wind and waves were neglected. The continu-
was capable of reproducing statistical properties of an observed network ity and the horizontal momentum equations in this case read:
in the Venice Lagoon (D'Alpaos et al., 2007). Di Silvio et al. (2010) pro-
posed a modelling approach in which deposition and erosion rates de- ∂u ∂v ∂w
þ þ ¼ 0; ð1Þ
pend on the difference between the local transport concentration and ∂x ∂y ∂z
the local equilibrium concentration. The model was used to simulate
the formation of a channel network in a tidal lagoon and results indicated 2 2 2
∂u ∂u ∂u ∂u ∂η ∂ u ∂ u ∂ u
that morphodynamic changes are rapid when the system is far from þu þv þw ¼ −g þ νx 2 þ νy 2 þ νz 2 ; ð2Þ
∂t ∂x ∂y ∂z ∂x ∂x ∂y ∂z
equilibrium but slow down once a network has developed.
A different type of modelling approach involves the coupling of
∂v ∂v ∂v ∂v ∂η ∂2 v ∂2 v ∂2 v
comprehensive hydrodynamic models and commonly used sediment þ u þ v þ w ¼ −g þ νx 2 þ νy 2 þ νz 2 ; ð3Þ
∂t ∂x ∂y ∂z ∂y ∂x ∂y ∂z
transport formulations. Marciano et al. (2005), for example, simulat-
ed morphological evolution by coupling the solution of the unsteady
where u, v, and w are the velocity components in the horizontal x-, y- and
depth-averaged shallow water equations with a sediment transport
in the vertical z-direction, respectively. t is time, g is the gravitational ac-
module based on the formulation of Engelund and Hansen (1967).
celeration, υx, υy, and υz are eddy viscosity coefficients, and η is the water
To facilitate the execution of long-term simulations and to overcome
level which can be described as:
the issue related to the difference in time scale over which hydrody-
namic and morphodynamic processes occur, Marciano et al. (2005) η
! η
!
accelerated morphological change by extrapolating the bathymetric ∂η ∂ ∂
þ ∫ udz þ ∫ vdz ¼ 0; ð4Þ
changes as the result of residual sediment transports over a single ∂t ∂x −h
∂y −h
tide. The model simulated the evolution of branching channel patterns
in a short idealized basin. The resulting channel network displayed a where h is the water depth. The solution grid uses rectangular Cartesian
branching behavior similar to patterns observed in natural basins. cells, enabling the application of a simple, efficient finite-difference/
Dastgheib et al. (2008) used a similar type of model to explore channel finite-volume scheme on a staggered grid. ELCOM applies a z-coordinate
network formation and the interaction between adjacent tidal basins in system in the vertical direction. The numerical scheme for computing the
 B. van Maanen et al. / Geomorphology 191 (2013) 23–34
temporal evolution of velocity is a semi-implicit solution of the governing
equations using a hybrid discretization of advective terms. ELCOM ac-
counts for the flooding and drying of grid cells.
Depth-averaged flow velocities, computed at every time-step, were
used to obtain instantaneous sediment transport rates according to the
formula developed by Engelund and Hansen (1967) which accounts for
both bed load and suspended load:
0:05U 5
Sflow ¼ pffiffiffi 3 2 ; ð5Þ
g C Δ D50
where Sflow is the sediment transport flux, U is the magnitude of flow ve-
locity, C is the Chézy coefficient, Δ is the relative density (ρs − ρw)/ρw,
ρs is the sediment density, ρw is the water density, and D50 is the median
grain size. The slope-driven sediment transport term Sslope is difficult to
parameterize. Here we follow the approach initially described by, for
example, Parker (1984) and Murray and Paola (2003), using the formu-
lation proposed by Kirwan and Murray (2007):
Sslope ¼ αb; ð6Þ
where α is a dimensional constant (m2/s) and b is the slope towards the
neighboring grid cell. Here slope-driven sediment transport only oc-
curred when grid cells were under water and when b exceeded a specif-
ic threshold. A threshold was used so that channel initiation was not
hindered and it was set to a value of 0.01 which resulted in realistic
channel development. Gradients in sediment transport rate yield bed
level changes due to conservation of sediment mass:
! phological evolution of the basins had slowed down after 1000 years and
 ∂z  ∂Sx ∂Sy
1−εpor b
þ þ ¼ 0; ð7Þ
∂t ∂x ∂y
where εpor is the bed porosity (kept fixed and equal to 0.4), and Sx and Sy
are the total sediment transport fluxes in the x- and y-direction, respec-
tively. This type of model that simulates the long-term morphological
evolution of tidal systems (e.g., Cayocca, 2001; Hibma et al., 2003;
Marciano et al., 2005; Dastgheib et al., 2008; van der Wegen and
Roelvink, 2008; Dissanayake et al., 2009) adopts a strategy to increase
the rate of morphological change (see Roelvink (2006) for an overview
of different methodologies). For the morphodynamic model developed
here, bed level changes were integrated over one tidal cycle and then
multiplied by an integer factor (i.e., a specific number of tidal cycles
which equals the morphodynamic time-step) before the updated mor-
phology was fed back into the hydrodynamic model. This factor was
optimized and determined at every iteration so that it could vary
throughout the morphological evolution. The factor was computed
by increasing it until bed level changes at any single grid-cell within
the domain exceeded 10% of the local water depth at high tide. Subse-
quently, the factor was applied to all grid-cells to compute the morpho-
logical change over the morphodynamic time-step. This allowed us to
reduce numerical effort and facilitated the execution of long-term sim-
ulations. Further details concerning model development and the results
of a sensitivity analysis (e.g., the sensitivity of model outcome to the
morphodynamic time-step) can be found in Van Maanen et al. (2011).
2.2. Model set-up
Numerical simulations were undertaken using idealized initial ba-
thymetries (Fig. 2). The configuration and the dimensions of the topog-
raphy were somewhat based on those as reported in Marciano et al.
(2005), who were amongst the first to perform this type of channel
network forming simulations. Marciano et al. (2005) applied model
parameters which corresponded with those of the tidal basins in the
Dutch Wadden Sea. The initial topographies applied here covered an
area of 17 × 17 km and were composed of an offshore area, inlet, and
basin. The offshore area was separated from the tidal basin by the
25
presence of land regions which formed impermeable and non-
erodible barriers. Random perturbations of −1.5 to +1.5 cm were
added to the bed level in the tidal basin to trigger channel pattern devel-
opment. Initial bathymetric conditions were varied by having a differ-
ent initial depth in the inlet of the tidal basin of 2, 4 or 6 m (Fig. 2).
The model was forced with a semidiurnal tide. Simulations were
performed with different combinations of initial depth and tidal
range. The geometry of the tidal basins and the tidal ranges were chosen
such that, even at high tide, a part of the basin remained dry (this is in
contrast for example with Marciano et al. (2005) and Dissanayake et
al. (2009)) so that the tidal channels could expand towards the shallow
areas in the upper part of the basin. A grid size of 100 m in both x- and
y-directions was used. With respect to the z-coordinate system within
the hydrodynamic model, the thickness of the horizontal layers was
kept at a uniform value of 40 cm. The hydrodynamic time-step was
set to 2 min. The sediments had a mean grain size diameter of
0.12 mm and a value of 65 m0.5/s has been used for the Chézy
coefficient. Slope-driven sediment transport was calculated using
α = 1.16 × 10−5 m 2/s.
3. Results
Model results indicated that morphodynamic interactions cause
channel initiation and give rise to channel pattern development (Fig. 3).
Fig. 4 shows simulated morphologies after 1000 years for different com-
binations of the basin's initial depth and tidal range (the computational
time to perform these long-term simulations was in the order of weeks
on a 3.0 GHz, 8.0 Gb RAM PC). As discussed in more detail later, the mor-
the basins thus had reached a more steady state. A large variability in
morphology can be observed, indicative of the crucial role that initial
bathymetry and tidal conditions play in the large scale evolution of tidal
embayments. The formation of a channel network was enhanced when
the initial depth of the basin decreased and/or the tidal range increased
(Fig. 4). For an initially unchannelized tidal basin with a depth of 2 m
and a tidal range of 2 m (the conditions corresponding to the evolution
as shown in Fig. 3), the tide residual sediment transport pattern
(Fig. 5a) shows a morphodynamically active situation with large gra-
dients in the sediment transport rate. As a consequence, large (of the
order of meters in the vertical direction) bathymetric changes occurred
within the first few years of morphological evolution (Fig. 3a). Two
deep channels rapidly developed in the inlet (area between the barrier
islands) of the tidal system. During the ebb-tide, large volumes of sedi-
ment were transported offshore where decelerating flows caused
sediment deposition. Consequently, a large ebb-tidal delta was formed
which, after 5 years, extended about 3 km offshore. During the first
5 years of morphological evolution (including the morphological en-
hancement factor), the tidal basin experienced a net sediment import
(solid red line in Fig. 6). The sediment transported into the basin origi-
nated from the inlet where accelerating flows caused scouring of the
sediment (Fig. 5a). In the basin, small-scale channels developed and
intertidal areas emerged. During subsequent morphological evolution
(Fig. 3b–d) the channels through the inlet continued to deepen and
reached a maximum depth of about 10 m after 1000 years. The
ebb-tidal delta increased in size and attained a slightly irregular
shape. The tidal channels in the basin continued to branch; a pro-
cess that ultimately resulted in the formation of a complex tidal
channel network.
The formation of channels and intertidal areas affects the duration of
rising and falling tide and this tidal asymmetry is a dominant factor in
driving tidally-averaged sediment transport (de Swart and Zimmerman,
2009). Flood dominance is likely where friction-induced distortion of
the tidal wave occurs. On the other hand, large intertidal areas favor
ebb-dominant sediment transport (Friedrichs and Aubrey, 1988;
Dronkers, 2005). The development of the channel network (and the
associated formation of intertidal areas) was then also accompanied
26 B. van Maanen et al. / Geomorphology 191 (2013) 23–34

2 2
a b
0 0
5

Distance (km)
−2 −2

10 −4 −4

−6 −6
15

−8 −8
2
c
2
0 d

Bed elevation (m)

5 0
Distance (km)

−2
−2

10 −4 −4

−6
−6
15 −8
5 10 15
−8 Distance (km)
5 10 15
Distance (km)

Fig. 2. Plan view of the initial morphologies used throughout this study. The initial depth of the tidal basin amounted to (a) 2, (b) 4, and (c) 6 m. Grey areas represent land regions
which form impermeable and non-erodible barriers. (d) Longitudinal profiles of the initial morphologies. Locations of the longitudinal profiles are indicated by the black lines in (a), (b),
and (c).

by a transition from sediment import to sediment export (solid red line affected the hypsometry of the tidal basin and over time the hypsomet-
in Fig. 6). A potential switch to ebb dominance when sediment-infilling ric curve tends to attain a shape which is similar to the one of a typical
becomes advanced was previously suggested by Boon and Byrne (1981). inlet system in the Dutch Wadden Sea (grey dots in Fig. 7). Further-
Large scale bathymetric changes related to the formation of the network more, the overall flow pattern had changed after 1000 years and a

2 2
a b
0 0
5
Distance (km)

−2 −2

10 −4 −4

−6 −6
15

<−8 <−8
2 2
c d
0 0
5
Distance (km)

−2 −2

10 −4 −4

−6 −6
15

<−8 <−8
5 10 15 5 10 15
Distance (km) Distance (km)

Fig. 3. Simulation of morphological evolution for a tidal range of 2 m and an initial depth of the basin of 2 m. Morphologies after (a) 5, (b) 30, (c) 80, and (d) 250 years. Black line
represents the 0 m contour line.
 B. van Maanen et al. / Geomorphology 191 (2013) 23–34 27

Bed elevation (m) 1

<−8 −6 −4 −2 0 2
3.0

2 3

2.5
Tidal range (m)

4 5 6
Distance (km)

5
2.0
10

15

7 8 9
Distance (km)

5
1.5
10

15

10 11 12
Distance (km)

5
1.0
10

15
5 10 15 5 10 15 5 10 15
Distance (km) Distance (km) Distance (km)

2.0 4.0 6.0

Initial depth (m)
Fig. 4. Simulated morphologies after 1000 years for different combinations of the basin's initial depth and tidal range. The colored dots in the left bottom corner of each morphology
plot are used in Fig. 9 to indicate the tidal range and initial depth of each simulation.

strong reduction in the tide residual sediment transport had occurred
(Fig. 5b).
When the initial depth of the basin was increased to 6 m (while the
tidal range was kept at 2 m; similar to the case previously discussed),
the morphology of the tidal system evolved differently (Fig. 8). The
scouring of the inlet and the import of sediment occurred at a slower
pace (Fig. 5c and dotted red line in Fig. 6). Also, because of the large
water depths, the imported sediment formed an unchannelized flood-
tidal delta (Fig. 8a). Over time, the flood-tidal delta expanded landward
and built up vertically and eventually became incised by channels. Si-
multaneously, small-scale channels were initiated in the shallow area
in the upper part of the basin. After 1000 years this initial formation
of channels could be detected (Fig. 8b). A net sediment import occurred
continuously throughout this period of morphological evolution (Fig. 6)
and the bed level in the inlet reached a maximum depth of more than
16 m. However, intertidal areas dissected by channels still had not
formed after 1000 years. The hypsometric curve for the basin after
1000 years (red stars with dotted line in Fig. 7) also indicates that the
channel network and intertidal areas were still under-developed. The
tidal basin continued to gain sediment and currents redistributed sedi-
ment from the deep areas to the shallow areas, ultimately resulting in
the formation of intertidal areas and a complete tidal channel network
(see Fig. 8d and the hypsometric curve represented by the light shaded
red triangles in Fig. 7).
In the case of a tidal basin with an initial depth of 6 m, an increase
in tidal range to 3 m resulted in enhanced tidal flow and flow acceler-
ation through the inlet because of the increase in tidal prism. The im-
port of large volumes of sediment (more than 7 million m 3 after just
200 years; dotted black line in Fig. 6) occurred therefore at the begin-
ning of morphological evolution. The flood-tidal delta, which formed
in the basin, started to become channelized after only 30 years. After
200 years, the volume of sediment present in the tidal basin remained
relatively constant and the sediment was mainly redistributed within
the basin such that a complete channel network had developed after
28 B. van Maanen et al. / Geomorphology 191 (2013) 23–34

1.0x10−4 m2/s 5.0x10−7 m2/s

a b

2.0x10−5 m2/s 2.0x10−7 m2/s

c d

1.0x10−4 m2/s 5.0x10−7 m2/s

e f

2.0x10−6 m2/s 2.0x10−7 m2/s

g h

Fig. 5. Tide residual sediment transport patterns over the initial morphologies for the area around the inlet when (a) the tidal range (TR) is 2 m and the basin's initial depth (ID) is
2 m, (c) TR = 2 m and ID = 6 m, (e) TR = 3 m and ID = 6 m, and (g) TR = 1 m and ID = 6 m. Subplots (b), (d), (f), and (h) show the corresponding sediment transport patterns over
the simulated morphologies after 1000 years.

1000 years (Fig. 4). The hypsometric curve (black stars in Fig. 7) is
somewhat comparable to the curve obtained for the simulated mor-
phology generated after 10,000 years with an initial depth of 6 m and
a tidal range of 2 m (light shaded red triangles). Decreasing the tidal
range to 1 m had the opposite effect. The ability of the currents to scour
the sediment from the inlet and transport it into the basin was reduced
(Fig. 5g and dotted blue line in Fig. 6) and there was no sign of channel
initiation after 1000 years (Fig. 4). Large-scale bathymetric changes
and adaptation in the hypsometry progressed more slowly and the
hypsometric curve associated with the simulated bathymetry after
10,000 years (light shaded blue triangles in Fig. 7) did not display
the typical sigmoidal shape, indicating that the channel network was
still under-developed.
4. Discussion
The basin's initial depth and tidal range were varied in our model
experiments to allow for a comparison with other studies that character-
ized tidal systems by evaluating the tidal amplitude to channel depth
ratio a/h (e.g., Friedrichs and Aubrey, 1988; Dronkers, 2005). Dronkers
(2005), for example, defined equilibrium conditions in terms of a/h
and the relative intertidal area, and found that the two were positively
related, since intertidal areas are needed to counterbalance the flood
dominance caused by the difference in tidal wave propagation between
high and low water and by friction-induced tidal distortion. For the sim-
ulations considered here, the magnitude of the mean residual sediment
transport decreased by several orders of magnitude during morphological
 B. van Maanen et al. / Geomorphology 191 (2013) 23–34 29

6
x 10 have a tendency towards ebb dominance (Friedrichs and Aubrey,
8
1988; Dronkers, 2005) and it is therefore interesting to note that the
simulated basins which are exporting sediment after 1000 years (see
6
Fig. 6) are all associated with a ratio (Aint/AHW)/(a/hbasin) above 1
(runs 1, 2, 4, 5, 7 in Fig. 4 and the corresponding colored dots in
Sediment volume (m3)

4
Fig. 9). For the morphologies simulated here, the 1:1 line in Fig. 9e
thus distinguishes between exporting and importing systems.
2
Dronkers (1998) argued that morphodynamic equilibrium can only
0
be present when ebb and flood durations are approximately equal and a
theoretical analysis showed that, under equilibrium conditions, the
−2
basin geometry satisfies the following relationship:
TR=3.0 ID=6  
h þ a 2 AHW
−4 TR=2.5 ID=6 TR=2.5 ID=4 ¼ ; ð8Þ
TR=2.0 ID=6 TR=2.0 ID=4 TR=2.0 ID=2 h−a ALW
TR=1.5 ID=6 TR=1.5 ID=4 TR=1.5 ID=2
−6
TR=1.0 ID=6 TR=1.0 ID=4 TR=1.0 ID=2
where h + a and h − a are the average channel depths at high and low
0 200 400 600 800 1000 water, and ALW is the basin area inundated at low tide. Measurements of
Time (years) morphological and tidal characteristics of Dutch tidal basins agreed with
this theoretical relationship (Dronkers, 1998). Following Dronkers
Fig. 6. Changes in sediment volume within the basin. Sediment import to (positive
numbers) and export from (negative numbers) the tidal basin for different combinations (1998), channel depth at low water was computed here as:
of the basin's initial depth and tidal range.
h−a ¼ V LW =ALW ; ð9Þ

evolution (Fig. 5). The rate at which sediment was transported to or from
the basin also decreased over time (Fig. 6). The tidal basins thus reached a
configuration for which morphological changes occurred at a slower
pace. Following the ideas suggested by Dronkers (2005), we assessed
how the amplitude-to-depth ratio and the relative intertidal area
changed during morphological evolution. The relative intertidal
area was defined here as the ratio of the intertidal surface area Aint
and the total area inundated at high tide AHW. The tidal amplitude
a was taken at the offshore boundary and we adopted the average
depth in the tidal basin at mean sea level hbasin as the characteristic
depth. Fig. 9 shows how the relationship between Aint/AHW and a/hbasin
changed over time and indicates that the morphology tends to evolve
such that Aint/AHW and a/hbasin become equal. Aint/AHW increases thus
linearly with a/hbasin and the ratio of the two approximate values of 1
for relatively stable morphologies. For the simulations with a large am-
plitude to basin depth ratio, (Aint/AHW)/(a/hbasin) was initially smaller
than 1 (Fig. 9a). (Aint/AHW)/(a/hbasin) increased during morphological
evolution because of the deepening of channels and the development
of intertidal areas. For small amplitude to basin depth ratios, morpho-
logical change slowed down so that (Aint/AHW)/(a/hbasin) remained rela-
tively stable through time. Basins with a large relative intertidal area
4 work by Passalacqua et al. (2010). This technique includes nonlinear
0 contours is then calculated. Channelized areas are characterized by pos-
Elevation (m)
where VLW is the basin volume at low water. Channel depth at high
water then directly results from:
h þ a ¼ ðh−aÞ þ 2a: ð10Þ
Model results indicated that the simulated tidal basins developed a
geometry that approached Dronker's (1998) equilibrium condition
(see Fig. 9b, d, and f in which the black line represents Eq. (8)). The ge-
ometry of the simulated basins continued to evolve after 1000 years
(situation as shown in Fig. 9f), resulting in an even closer agreement
with Eq. (8) and the field data.
To quantify tidal network properties, Rinaldo et al. (1999) applied
a method developed by Fagherazzi et al. (1999) to automatically extract
networks from digital maps so that these networks could be studied in
detail. One of the measures used by Rinaldo et al. (1999) to characterize
the networks was the channel area proportion, defined as the ratio of
channeled area (the number of channelized grid cells) to the total
basin area (the total number of grid cells), representing thus the frac-
tion of the tidal embayment occupied by the channels. We followed
this approach to quantify the differences in channel network evolution
as observed in Fig. 4. To extract the channel networks from the simulat-
ed morphologies in our study, we adopted a technique based on the
geometric filtering of the topography to enhance features that are crit-
ical to the network extraction. The geometric curvature of the isoheight
itive curvature and channels are defined as areas where a sudden
change in the statistical signature of the landscape occurs. We refer

the reader to Passalacqua et al. (2010) for further details. This tech-
−4
nique, similar to the one developed by Fagherazzi et al. (1999), deter-
observed (TR=2.25) mines which grid cells are part of the channel network so that the
TR=2.0 ID=2 year 0 channel area proportion can be evaluated over time (Fig. 10). After
−8 TR=2.0 ID=2 year 1000 50 years, the channel area proportion is larger than 0.2 only in the
TR=2.0 ID=6 year 0 case of a shallow basin with an initial depth of 2 m and a tidal range
TR=2.0 ID=6 year 1000 of 2 m. Except for when the tidal range amounts to 3 m, less than 5%
TR=2.0 ID=6 year 10000
of the basin has become channelized when the initial depth is 6 m
−12 TR=1.0 ID=6 year 10000
(Fig. 10a). An increase in channel area proportion can be observed for
TR=3.0 ID=6 year 1000
increasing values of tidal range and for shallower basins. The continuing
0 0.2 0.4 0.6 0.8 1
development of the channel network is reflected by an ongoing increase
Proportion of basin above elevation in the fraction of the basin occupied by channels (Fig. 10b–d). The shal-
Fig. 7. Hypsometric curves for some of the simulated morphologies. The observed
low basins have become channelized for about 30% after 1000 years
hypsometry represents the Ameland inlet in the Dutch Wadden Sea (reproduced from and by then 8 out of the 12 simulations have a channel area proportion
Marciano et al. (2005)). which is larger than 0.2 (Fig. 10d).
30 B. van Maanen et al. / Geomorphology 191 (2013) 23–34
a b
5
Distance (km)
10
15
c d
5
Distance (km)
10
15
5 10 15
Distance (km)
Fig. 8. Simulation of morphological evolution for a tidal range of 2 m and an initial depth of the basin of 6 m. Morphologies after (a) 60, (b) 1000, (c) 3000, and (d) 10,000 years.
Black line represents the 0 m contour line.
Figs. 4 and 10 clearly show that both the forcing (tidal range) and
initial conditions (basin's initial depth) have a strong control on the mor-
phological evolution of the tidal basin. Recently, Perron and Fagherazzi
(2012) highlighted that the role of initial conditions in determining land-
scape evolution in general has received relatively little attention. They
emphasized that the initial topographic surface affects landscape evolu-
tion as well as the overall landform characteristics. The initial topography
also influences the characteristics of the tidal embayments simulated
here by affecting the fraction of the basin that becomes channelized.
The channel area proportion remains stable once a well-developed chan-
nel network has formed. In the case of an initial depth of 2 m and a tidal
range of 1 m, for example, the channel area proportion amounts to
0.29 after 4000 years which is similar to the proportion observed
after 1000 years (Fig. 10d). Although small differences in channel
area proportion should be treated with care, Fig. 10d indicates that
the channelized proportion for basins with a well-developed network
decreases for an increase in the initial depth. When the initial depth is
2 m, a well-developed network has formed after 1000 years for all the
tidal ranges considered here (Fig. 4) and the channel area proportion
is then between 0.29 and 0.30 (Fig. 10d). This proportion is between
0.25 and 0.26 for well-developed networks when the basin's initial
depth is 4 m (Fig. 10d). In the case of an initial depth of 6 m, a complete
tidal channel network has only formed for a tidal range of 3 m (Fig. 4)
and for this situation the channelized proportion has decreased further
to a value of 0.23 (Fig. 10d). In addition, the channel area proportion of
the tidal embayment as shown in Fig. 8d, which represents the situation
after 10,000 years when the initial depth is 6 m and the tidal range is
2 m, is 0.22.
To further explore the effects of forcing and initial conditions on
channel network characteristics, we used the extracted networks to
determine the number of channels located at a specific distance from
the coastal inlet. Fig. 11a shows the number of channels as a function
of distance for three different combinations of tidal range and initial
depth. The number of channels initially increases in the landward direc-
tion because of the ongoing branching of the channels. Not all the tidal
2 2
0 0
−2 −2
−4 −4
−6 −6
<−8 <−8
2 2
0 0
−2 −2
−4 −4
−6 −6
<−8 <−8
5 10 15
Distance (km)
channels expand equally far which causes the abrupt decline in channel
occurrence in the upper reaches of the basin. Both the tidal range and
the overall shape of the embayment affect the part of the basin that be-
comes inundated and thus directly control the landward extension of
channels. The landward extent to which the channel network expands
thus differs for each simulation and increases for larger initial depths
and decreases for smaller tidal ranges (Fig. 11a). Wang (1992) performed
a theoretical analysis (also presented by Marciano et al., 2005) which sug-
gested that larger water depths correspond to larger and more widely
spaced channels. As can be seen in Fig. 11a, a change in the initial depth
from 2 m to 6 m in the case of a tidal range of 2 m is associated with a
decrease in channel occurrence up to ~5.5 km from the inlet (at this
distance some tidal channels in the shallow basin already stopped
expanding). Furthermore, the channel drainage width, which is
evaluated by dividing the length of the semi-circle that connects
the points located at a specific distance from the inlet by the correspond-
ing number of channels, has indeed increased (Fig. 11b). The basin area
close to the inlet becomes shallower (compare the dotted green and
blue lines in Fig. 11a) and channel spacing decreases for a smaller tidal
range (Fig. 11b). This overall increase in bed elevation occurs because a
smaller tidal range leads to a reduction in the tidal prism flowing through
the channels which results in a decrease in the cross-sectional area of the
channels so that they become less deep and wide. The initial landward
decrease in channel drainage width, which can be observed for the
tidal networks simulated here (Fig. 11b), is in agreement with Wang's
(1992) analysis and also agrees with the modelling work of Marciano
et al. (2005) who showed that the drainage width increases with an in-
crease in the channel order as defined by Horton's hierarchical analysis
(according to Horton's theory, the confluence of two channels can define
a higher-order channel). Fig. 11c shows how the channel density, which
is the inverse of channel drainage width, varies with bed elevation. A
maximum density of 1.9 channels per km is reached at a bed elevation
of −0.22 m for the embayment with an initial depth of 2 m and a tidal
range of 1 m. Also, for a tidal range of 2 m the channel density in the
deep basin starts to decrease at an elevation of −0.74 m, while the
 B. van Maanen et al. / Geomorphology 191 (2013) 23–34 31

1.0 5
a b
4
0.75

AHW/ALW
Aint/AHW
3
0.5
2

0.25
1

0 0
1.0 5
c d
4
0.75

AHW/ALW
Aint/AHW

3
0.5
2

0.25
1

0 0

1.0 5
e f
4
0.75
AHW/ALW
Aint/AHW

3
0.5
2

0.25
1

0 0
0 0.25 0.5 0.75 1.0 0 5 10 15 20 25
a/hbasin ((h+a)/(h−a))2

Fig. 9. Relation between Aint/AHW and a/hbasin for the morphologies after (a) 0, (c) 100, and (e) 1000 years. Relation between AHW/ALW and ((h + a)/(h − a))2 for the morphologies
after (b) 0, (d) 100, and (f) 1000 years. Each dot represents a simulation with a specific tidal range and initial depth and the type of simulation is indicated by the corresponding
color of the dots in Fig. 4. The lines represent the line of equality. The grey dots in (b), (d), and (f) represent measurements of the Dutch tidal basins (Dronkers, 1998).

3.0 a 0.061 3.0 b 0.11

Tidal range (m)

2.5 0.11 0.047 2.5 0.17 0.063

2.0 0.22 0.071 0.046 2.0 0.25 0.11 0.048

1.5 0.18 0.048 0.044 1.5 0.23 0.062 0.043

1.0 0.084 0.043 0.034 1.0 0.12 0.048 0.036

3.0 c 0.20 3.0 d 0.23

Tidal range (m)

2.5 0.24 0.12 2.5 0.25 0.21

2.0 0.27 0.19 0.061 2.0 0.29 0.26 0.14

1.5 0.26 0.10 0.046 1.5 0.30 0.21 0.064

1.0 0.19 0.058 0.040 1.0 0.29 0.11 0.047

2.0 4.0 6.0 2.0 4.0 6.0
Initial depth (m) Initial depth (m)

Fig. 10. Channel area proportion (defined as the ratio of channeled area to the total basin area) after (a) 50, (b) 100, (c) 250, and (d) 1000 years for different combinations of the
basin's initial depth and tidal range.
32 B. van Maanen et al. / Geomorphology 191 (2013) 23–34

30 3
depth of the basin and the associated sediment availability and export
a constrain the hypsometrical evolution. Under similar tidal conditions,
the initial depth of the basin governs the bed level distribution and deter-
Number of channels

mines whether a deep or shallow tidal basin eventuates (compare the

Elevation (m)
20 0
red stars with solid line and light shaded red triangles in Fig. 7). Never-
theless, we observed that, despite the different hypsometric characteris-
tics, all the simulated basins reach a state of reduced morphodynamic
10 −3 activity. As argued by Dronkers (2005), the hypsometry of both shallow
basins with large intertidal areas and deep basins with small intertidal
areas can result in equal ebb and flood durations and thus be consistent
with morphologic equilibrium.
0 −6 Moore et al. (2009) used the shape of the hypsometric curve as an
3 4 5 6 7
indicator of the evolutionary stage of an estuarine system (see also
Distance from coastal inlet (km)
Boon and Byrne (1981)). An immature estuary has undergone little
2
infilling while a more mature and sediment-filled estuary is likely to
b
Channel drainage width (km)

have a greater extent of tidal flats and thus a different hypsometric

1.5
curve. The hypsometric curve of the estuarine system studied by
Moore et al. (2009) was evaluated in terms of the non-dimensionalized
basin area ahyp/Ahyp and the non-dimensionalized basin elevation
1 hhyp/Hhyp. ahyp is the basin area lying above a certain height above
the minimum basis elevation, Ahyp is the maximum basin area, hhyp
is the height above the minimum bed elevation, and Hhyp is the height
0.5 interval between maximum and minimum basin elevations. In addition
to the hypsometric curves shown in Fig. 7, we plotted these curves by
adopting the non-dimensional terms (Fig. 12a). The hypsometric curves
0 for the initial bathymetries with a depth of 2 m and 6 m are then
3 4 5 6 7
Distance from coastal inlet (km) somewhat similar, both showing almost a linear relationship be-
2
tween ahyp/Ahyp and hhyp/Hhyp (dark shaded red dots in Fig. 12a). De-
spite differences in the initial depth, basins with a well-developed
c
channel network also have comparable curves. A nonlinear relationship
Channel density (km−1)

1.5 between ahyp/Ahyp and hhyp/Hhyp exists for these channelized basins
(red stars with solid line, black stars, and light shaded red triangles in
Fig. 12a). Boon and Byrne (1981) and Moore et al. (2009) used the fol-
1 lowing formula to describe the hypsometrical relationship:

ahyp G
¼ with ð11Þ
0.5 Ahyp r þ Gð1−r Þ

0
−6 −4 −2 0 2
Elevation (m)
Fig. 11. (a) Number of channels versus distance from the coastal inlet, (b) channel drain-
age width versus distance from the coastal inlet, and (c) channel density versus the aver-
aged bed elevation at a specific distance from the coastal inlet after 4000 years when the
tidal range (TR) is 1 m and the basin's initial depth (ID) is 2 m (solid green), after
1000 years when TR = 2 m and ID = 2 m (solid blue), and after 10,000 years when
TR = 2 m and ID = 6 m (solid red). The dotted lines in (a) represent the averaged bed
elevation.
density peaks at an elevation of 0.30 m for the shallow basin. This causes
a large difference in channel density at a bed elevation of around 0 m
which is related to the change in geometry of the basins. That is, the un-
derlying bed profile is steeper for a deep basin (Fig. 2d and the red dotted
line in Fig. 11a) so that a smaller volume of water is stored by the inter-
tidal area in the upper part of the basin. This reduces the local tidal prism
which hinders further expansion of the tidal channels.
Forcing and initial conditions also influence the hypsometry of the
tidal basin. The sediments are transported towards higher elevations
under large tidal range conditions which has obviously an effect on
the distribution in bed levels (compare the black stars and light shaded
red triangles in Fig. 7). The basin's initial depth, on the other hand, has
an influence on the hypsometry by determining the overall amount of
sediment available within the tidal basin. It should be emphasized
that our simplified case is a tidal system where sediment input from ex-
ternal sources (e.g., alongshore drift, rivers) is absent so that the initial
 γ
G ¼ 1−hhyp =Hhyp ; and ð12Þ
A min
r¼ ; ð13Þ
Ahyp
where Amin is the minimum basin area. Following Boon and Byrne
(1981), values for the parameters r and γ were tuned to obtain the best
fit between Eq. (11) and the hypsometric curves for the simulated basins
(Fig. 12b). The values r = 0.04 and γ = 1.47 were found for the tidal
basin with a well-developed channel network (green stars in Fig. 12b).
These parameter values correspond to the ones found by Boon and
Byrne (1981) who analyzed a mature and sediment filled tidal basin
(r = 0.01 and γ = 1.8 in their case). The γ-value increases for less ma-
ture basins (γ = 3.5–5.0 for an immature system (Moore et al., 2009))
and γ = 2.19 for the basin that has still an underdeveloped channel net-
work after 10,000 years (green triangles in Fig. 12b). For this situation,
however, Eq. (11) agrees less well with the hypsometric curve. A good
match with the initial bathymetry is possible (green dots in Fig. 12b)
only when r is set to an unphysical value (the best match is obtained
for r > 1 which is unphysical because r = Amin/Ahyp so that it cannot ex-
ceed a value of 1).
The simulated morphological evolution described here is in agree-
ment with the results of modelling studies focusing on the initial forma-
tion of channel patterns. The stability analysis performed by Schuttelaars
and de Swart (1999), for example, showed that flow concentration and
acceleration are key factors in the formation and expansion of tidal chan-
nels. They demonstrated that bottom friction causes the velocity to
 B. van Maanen et al. / Geomorphology 191 (2013) 23–34 33

1 gaining maximum insight into the role of initial bathymetry and tidal
a range on the long-term evolution of tidal embayments.

0.8
5. Conclusions

We applied a model which is capable of simulating the long-term

0.6 morphological evolution of tidal embayments. The model was used to
hhyp/Hhyp

explore the effects of tidal range and initial bathymetry on landscape

evolution. A shallow initial basin and a large tidal range allow the
rapid (order of 10 1–10 2 years for the morphologies simulated here)
0.4
TR=2.0 ID=2 year 0 formation of channels and intertidal areas. When the basin is deeper,
TR=2.0 ID=2 year 1000 large tidal range conditions facilitate the transport of large volumes of
TR=2.0 ID=6 year 0 sediment into the basin. A flood-tidal delta forms which becomes chan-
0.2 TR=2.0 ID=6 year 1000 nelized and a complete channel network eventually develops (after
TR=2.0 ID=6 year 10000 103–10 4 years). However, when the tidal basin is deep and the tidal
TR=1.0 ID=6 year 10000 range is small, large scale bathymetric changes are slowed down and
TR=3.0 ID=6 year 1000 an underdeveloped channel network remains present over long time-
0 scales (>104 years). Based on the model results, we can draw the fol-
1 lowing conclusions:
b
• An increase (decrease) in the tidal range and a decrease (increase)
in the basin's initial depth result in a more (less) rapid channel net-
0.8
work formation.
• The large initial water depths in deep tidal basins hinder the forma-
tion of channels and a flood-tidal delta forms instead.
0.6 • The shape of the hypsometric curves resulting from numerical sim-
hhyp/Hhyp

ulations is consistent with existing observations.

0.4
TR=2.0 ID=2 year 0
γ=0.72 r=1.31
TR=2.0 ID=2 year 1000
0.2
γ=1.47 r=0.04
TR=1.0 ID=6 year 10000
γ=2.19 r=0.04
0 a concurrent decrease (increase) in the landward extent of the chan-
0 0.2 0.4 0.6 0.8 1
ahyp/Ahyp
Fig. 12. (a) Hypsometric curves for some of the simulated morphologies in terms of the
non-dimensionalized basin area ahyp/Ahyp and the non-dimensionalized basin elevation
hhyp/Hhyp. (b) Hypsometric curves and the results from the curve-fitting technique apply-
ing Eq. (11).
become smaller above shoals and larger in the channels, causing sedi-
ment to be transported from the channels towards the shoals giving
rise to a positive feedback mechanism which results in further deepening
of the channels and shallower shoals. Our simulations indicate that when
bottom friction is not affecting the flow in deep tidal basins, the positive
feedback mechanism is not effective and channels do not develop
(see for example the runs 9 and 12 in Figs. 4 and 8a). Alternatively,
a flood-tidal delta forms which grows and becomes shallower over
time until the effects of bottom friction start to influence flow veloc-
ities and determine preferential flow paths, ultimately causing chan-
nel initiation and channel network development.
Finally, the simplifications concerning the description of hydro-
dynamic and sediment transport processes that have been made
throughout model development should be mentioned. The flow is
solely tidally-driven (no waves, wind, or density-driven effects) and
spring-neap effects or variations in the mean water level have been ig-
nored. Furthermore, a simple sediment transport formulation (although
commonly applied) which includes a single grain size diameter and only
allows modelling of non-cohesive sediment has been used. We inten-
tionally included here only a minimum number of processes so that
model results are as transparent as possible and with the purpose of
• The tidal basins evolve towards a state of less morphodynamic ac-
tivity for which the relative intertidal area equals the ratio of tidal
amplitude and basin depth.
• Both initial depth and tidal range influence the hypsometry of the
basins by determining the overall sediment availability and affecting
the bed elevation towards which sediment can be redistributed, re-
spectively.
• For well-developed channel networks, a decrease (increase) in the
tidal range leads to an increase (decrease) in channel occurrence and
nels.
• An increase (decrease) in the basin's initial depth leads to a channel
network with larger (smaller) and more (less) widely spaced channels
and results in an overall decrease (increase) in channel area propor-
tion. Furthermore, there is a decrease (increase) in the ability of the
channels to expand into the upper part of a deep (shallow) basin be-
cause of a decrease (increase) in the local tidal prism which is related
to the overall shape of the basin.
Future challenges in this area of research will involve not only
increasing the complexity of the hydrodynamic forcing (e.g., wave
action), but also adding the effects of biological components (both
vegetation and benthic fauna) so that the role of the balance be-
tween biotic and abiotic processes in the long-term evolution of
these tidal systems can be explored.
Acknowledgments
B.v.M. and G.C. were funded by the (New Zealand) Foundation for
Research, Science and Technology. G.C. also acknowledges funding
from the “Cantabria Campus Internacional, Augusto Gonzalez Linares
Program”. Thanks to the Centre for Water Research (University of
Western Australia) for providing the hydrodynamic solver (ELCOM).
The National Center for Earth-surface Dynamics is acknowledged
for providing the software to extract channel networks. We thank
two anonymous reviewers for providing detailed and constructive
feedback.
34 B. van Maanen et al. / Geomorphology 191 (2013) 23–34
References
Boon, J.D., Byrne, R.J., 1981. On basin hypsometry and the morphodynamic response of
coastal inlet systems. Marine Geology 40, 27–48.
Cayocca, F., 2001. Long-term morphological modeling of a tidal inlet: the Arcachon
Basin, France. Coastal Engineering 42, 115–142.
D'Alpaos, A., Lanzoni, S., Marani, M., Fagherazzi, S., Rinaldo, A., 2005. Tidal network
ontogeny: channel initiation and early development. Journal of Geophysical Re-
search 110, F02001.
D'Alpaos, A., Lanzoni, S., Marani, M., Bonometto, A., Cecconi, G., Rinaldo, A., 2007. Spon-
taneous tidal network formation within a constructed salt marsh: observations
and morphodynamic modelling. Geomorphology 91, 186–197.
Dastgheib, A., Roelvink, J.A., Wang, Z.B., 2008. Long-term process-based morphological
modelling of the Marsdiep Tidal Basin. Marine Geology 256, 90–100.
De Swart, H.E., Zimmerman, J.T.F., 2009. Morphodynamics of tidal inlet systems. Annual
Review of Fluid Mechanics 41, 203–229.
Di Silvio, G., Dall'Angelo, C., Bonaldo, D., Fasolato, G., 2010. Long-term model of pla-
nimetric and bathymetric evolution of a tidal lagoon. Continental Shelf Re-
search 30, 894–903.
Dissanayake, D.M.P.K., Roelvink, J.A., van der Wegen, M., 2009. Modelled channel patterns
in a schematized tidal inlet. Coastal Engineering 56, 1069–1083.
Dronkers, J., 1998. Morphodynamics of the Dutch Delta. In: Dronkers, J., Scheffers,
M.B.A.M. (Eds.), Physics of Estuaries and Coastal Seas. Balkema, Rotterdam, Netherlands,
pp. 297–304.
Dronkers, J., 2005. Dynamics of Coastal Systems. World Scientific, Singapore (519 pp.).
Engelund, F., Hansen, E., 1967. A Monograph on Sediment Transport in Alluvial Streams.
Teknisk Forlag, Copenhagen (62 pp.).
Fagherazzi, S., Bortoluzzi, A., Dietrich, W.E., Adami, A., Lanzoni, S., Marani, M., Rinaldo, A.,
1999. Tidal networks 1. Automatic network extraction and preliminary scaling fea-
tures from digital terrain maps. Water Resources Research 35, 3891–3904.
Friedrichs, C.T., Aubrey, D.G., 1988. Non-linear tidal distortion in shallow well-mixed
estuaries: a synthesis. Estuarine, Coastal and Shelf Science 27, 521–545.
Hibma, A., De Vriend, H.J., Stive, M.J.F., 2003. Numerical modelling of shoal pattern formation
in well-mixed elongated estuaries. Estuarine, Coastal and Shelf Science 57, 981–991.
Hodges, B.R., Imberger, J., Saggio, A., Winters, K.B., 2000. Modeling basin-scale internal
waves in a stratified lake. Limnology and Oceanography 45, 1603–1620.
Iwasaki, T., Shimizu, Y., Kimura, I., 2011. Experimental and numerical studies on develop-
ment of tidal creek network. Proceedings of the 7th IAHR Symposium on River, Coastal
and Estuarine Morphodynamics, pp. 2091–2100 (Beijing, China).
Kirwan, M.L., Murray, A.B., 2007. A coupled geomorphic and ecological model of tidal
marsh evolution. Proceedings of the National Academy of Sciences of the United
States of America 104, 6118–6122.
Marciano, R., Wang, Z.B., Hibma, A., de Vriend, H.J., Defina, A., 2005. Modeling of chan-
nel patterns in short tidal basins. Journal of Geophysical Research 110, F01001.
Mitsch, W.J., Gosselink, J.G., 2007. Wetlands. John Wiley & Sons, New York (582 pp.).
Moore, R.D., Wolf, J., Souza, A.J., Flint, S.S., 2009. Morphological evolution of the Dee Estuary,
Eastern Irish Sea, UK: a tidal asymmetry approach. Geomorphology 103, 588–596.
Murray, A.B., Paola, C., 2003. Modelling the effect of vegetation on channel pattern in
bedload rivers. Earth Surface Processes and Landforms 28, 131–143.
Parker, G., 1984. Lateral bed load transport on side slopes. Journal of Hydraulic Engi-
neering 110, 197–199.
Passalacqua, P., Do Trung, T., Foufoula-Georgiou, E., Sapiro, G., Dietrich, W.E., 2010. A
geometric framework for channel network extraction from lidar: nonlinear diffusion
and geodesic paths. Journal of Geophysical Research 115, F01002.
Perron, J.T., Fagherazzi, S., 2012. The legacy of initial conditions in landscape evolution.
Earth Surface Processes and Landforms 37, 52–63.
Rinaldo, A., Fagherazzi, S., Lanzoni, S., Marani, M., 1999. Tidal networks 2. Watershed
delineation and comparative network morphology. Water Resources Research
35, 3905–3917.
Roelvink, J.A., 2006. Coastal morphodynamic evolution techniques. Coastal Engineering
53, 277–287.
Schuttelaars, H.M., de Swart, H.E., 1999. Initial formation of channels and shoals in a
short tidal embayment. Journal of Fluid Mechanics 386, 15–42.
Stefanon, L., Carniello, L., D'Alpaos, A., Lanzoni, S., 2010. Experimental analysis of tidal
network growth and development. Continental Shelf Research 30, 950–962.
Tambroni, N., Bolla Pittaluga, M., Seminara, G., 2005. Laboratory observations of the
morphodynamic evolution of tidal channels and tidal inlets. Journal of Geophysical
Research 110, F04009.
Van der Wegen, M., Roelvink, J.A., 2008. Long-term morphodynamic evolution of a tidal
embayment using a two-dimensional, process-based model. Journal of Geophysical
Research 113, C03016.
Van Maanen, B., Coco, G., Bryan, K.R., 2011. A numerical model to simulate the formation
and subsequent evolution of tidal channel networks. Australian Journal of Civil Engi-
neering 9, 61–71.
Vlaswinkel, B.M., Cantelli, A., 2011. Geometric characteristics and evolution of a tidal channel
network in experimental setting. Earth Surface Processes and Landforms 36, 739–752.
Wang, Z.B., 1992. Fundamental considerations on morphodynamic modeling in tidal
regions. Rep. Z331, Part I. Delft Hydraul, Delft, Netherlands.