Modeling of Coastal Morphological Processes 28.4 Future Directions 629
Part C | 28.4
the previous sections. The major reasons may be that
wave forcing is relatively limited due to short fetch
length and a shallow and wind-sheltered environment.
Also, estuarine channel-shoal patterns develop by tideresidual sediment transports that are about an order of
magnitude smaller than the maximum ebb, flood, or
river-induced sediment transports. As a result, data sets
for calibration and validation of these models should
cover decades and (preferably) longer time spans. This
makes case studies rare and often limited to estuaries
with considerable human impact such as environments
of ports.
28.4 Future Directions
28.4.1 Ensemble Modeling
Morphodynamic forecasts are never the result of a single run. Modeling practice shows the need for considerable testing of model parameter settings and model
behavior until the results can be presented with confidence. The testing is sometimes caused by model
shortcomings or bugs, but another important reason is
that model input parameters are unknown beforehand.
When we take process-based modeling as an example, uncertainty exists with respect to:
1. Process definitions in the model itself: Are sediment
transport formulae described adequately? Is there
a need for including more processes? Can we apply
a morphological factor?
2. Input parameter values: What should be the values
for the viscosity/diffusivity coefficient, and sediment characteristics such as the critical bed shear
stress or fall velocity? How many sediment classes
should be defined?
3. Initial conditions: What is the initial sediment composition in the bed? What should be the initial SSC
or salinity field in the water column?
4. Forcing conditions: Is it possible to schematize the
wind climate? Can we apply a morphological tide?
What is the impact of extreme events? What are expected rates in the sea-level rise?
5. The modeled system itself: Does the modeled system itself reflect (deterministic) chaos or is it insensitive to slightly different model input?
Until now, the literature shows only limited attention for a systematic analysis on morphodynamic model
uncertainties. Baart et al. [28.83] point to confidence
intervals around morphological forecasts as the result of
the propagation of uncertainty through a chain of models from weather (wind and pressure) to hydrodynamic
(waves and tides) and finally morphodynamic predictions. Vreugdenhil [28.84] and Pinto et al. [28.75]
present examples on the uncertainty associated with different sediment transport formulations, whereas Fortunato et al. [28.76] and Ruessink and Kuriyama [28.82]
explore the development of uncertainty over time. Further, Ruessink and Kuriyama [28.82] and Van der Wegen
et al. [28.79] show that ensemble runs in morphodynamic predictions can be used to discriminate dominant
forcing from secondary processes.
There lies a challenge in future research to further explore the conditions in which morphodynamic
models function well, what the quality of model input
parameters should be to come to trustworthy forecasts
and what the quality of model output is in the light of
uncertain model input parameters. Ensemble runs that
include a systematic variation of (unknown) input parameters as well as model output presented in terms of
a probability range may be a promising way to explore
the value of models under uncertain conditions.
28.4.2 Biogeomorphology
In the last two decades, recognition has grown on
the importance of the interaction between geomorphological and biological processes, and a new field
of biogeomorphology has sprung up. In the coastal
context, important areas where such interactions are
dominant are mudflats, salt marshes, mangroves, coral
reefs, and floodplains. Erosion and sedimentation of the
estuarine bed may be subject to biological processes.
Bed developments are influenced by benthic flora and
fauna and can be a function of biological processes
like the development of bio-films or the presence of
bio-engineers [28.85–87]. Salt marshes eventually may
be the most dramatic example where vegetation cover
governs flow patterns and sedimentation trapping rates.
D’Alpaos et al. [28.88] and Temmerman et al. [28.89]
present process-based modeling efforts including the
feedback process of vegetation on morphodynamics.
In this way, salt marsh dynamics and its feedback on
channel shoal morphodynamics can be included in the
modeling effort.
Bed stabilization may occur due to bio-films or
vegetation roots, whereas surface deposit feeders destabilize the upper bed layer during feeding activities. All
these processes may be a function of time so that they
change considerably over months, seasons, or even over
decades.
Part C | 28.4
the previous sections. The major reasons may be that
wave forcing is relatively limited due to short fetch
length and a shallow and wind-sheltered environment.
Also, estuarine channel-shoal patterns develop by tideresidual sediment transports that are about an order of
magnitude smaller than the maximum ebb, flood, or
river-induced sediment transports. As a result, data sets
for calibration and validation of these models should
cover decades and (preferably) longer time spans. This
makes case studies rare and often limited to estuaries
with considerable human impact such as environments
of ports.
28.4 Future Directions
28.4.1 Ensemble Modeling
Morphodynamic forecasts are never the result of a single run. Modeling practice shows the need for considerable testing of model parameter settings and model
behavior until the results can be presented with confidence. The testing is sometimes caused by model
shortcomings or bugs, but another important reason is
that model input parameters are unknown beforehand.
When we take process-based modeling as an example, uncertainty exists with respect to:
1. Process definitions in the model itself: Are sediment
transport formulae described adequately? Is there
a need for including more processes? Can we apply
a morphological factor?
2. Input parameter values: What should be the values
for the viscosity/diffusivity coefficient, and sediment characteristics such as the critical bed shear
stress or fall velocity? How many sediment classes
should be defined?
3. Initial conditions: What is the initial sediment composition in the bed? What should be the initial SSC
or salinity field in the water column?
4. Forcing conditions: Is it possible to schematize the
wind climate? Can we apply a morphological tide?
What is the impact of extreme events? What are expected rates in the sea-level rise?
5. The modeled system itself: Does the modeled system itself reflect (deterministic) chaos or is it insensitive to slightly different model input?
Until now, the literature shows only limited attention for a systematic analysis on morphodynamic model
uncertainties. Baart et al. [28.83] point to confidence
intervals around morphological forecasts as the result of
the propagation of uncertainty through a chain of models from weather (wind and pressure) to hydrodynamic
(waves and tides) and finally morphodynamic predictions. Vreugdenhil [28.84] and Pinto et al. [28.75]
present examples on the uncertainty associated with different sediment transport formulations, whereas Fortunato et al. [28.76] and Ruessink and Kuriyama [28.82]
explore the development of uncertainty over time. Further, Ruessink and Kuriyama [28.82] and Van der Wegen
et al. [28.79] show that ensemble runs in morphodynamic predictions can be used to discriminate dominant
forcing from secondary processes.
There lies a challenge in future research to further explore the conditions in which morphodynamic
models function well, what the quality of model input
parameters should be to come to trustworthy forecasts
and what the quality of model output is in the light of
uncertain model input parameters. Ensemble runs that
include a systematic variation of (unknown) input parameters as well as model output presented in terms of
a probability range may be a promising way to explore
the value of models under uncertain conditions.
28.4.2 Biogeomorphology
In the last two decades, recognition has grown on
the importance of the interaction between geomorphological and biological processes, and a new field
of biogeomorphology has sprung up. In the coastal
context, important areas where such interactions are
dominant are mudflats, salt marshes, mangroves, coral
reefs, and floodplains. Erosion and sedimentation of the
estuarine bed may be subject to biological processes.
Bed developments are influenced by benthic flora and
fauna and can be a function of biological processes
like the development of bio-films or the presence of
bio-engineers [28.85–87]. Salt marshes eventually may
be the most dramatic example where vegetation cover
governs flow patterns and sedimentation trapping rates.
D’Alpaos et al. [28.88] and Temmerman et al. [28.89]
present process-based modeling efforts including the
feedback process of vegetation on morphodynamics.
In this way, salt marsh dynamics and its feedback on
channel shoal morphodynamics can be included in the
modeling effort.
Bed stabilization may occur due to bio-films or
vegetation roots, whereas surface deposit feeders destabilize the upper bed layer during feeding activities. All
these processes may be a function of time so that they
change considerably over months, seasons, or even over
decades.