Case Study for Ruegen Area in Mike21

1
2D Modeling Exercise in Hydroinformatics

• Case Study for Ruegen Area (in Mike21)
• Name YONG,WANG
• Euroaquae 2007/08

2
Contents
1. Project description 2. Model setting up and
calibration 3. Gird editing and scenario
analysis 4. Conclusion
3
1. Project description

• A 2D coastal Hydrodynamic model was built using
  Mike21.Ruegen island water had been selected as
  the modeling body, the following figure shows the
  geographical description of the area in Baltic
  sea.

4
1. Project description

• Task and requirement
• In the past it happened that the island of
  Hiddensee fell into 2 parts. This situation is to
  be modeled by introducing a channel of 1 meter
  depth and 100 meter width at the given location.
  The impact of this situation is to be
  investigated with respect the discharge at
  locations 3 and 4. A qualitative analysis about
  sediment transport is to be made by using
  particle tracking at the entrance of the old
  navigation channel between Bock and Hiddensee.

5
1. Project description

• Procedures
• 1) Set up the model with original bathymetry
  (boundary, wind force).
• 2) Calibrate the model using measured data at
  Stralsund and Neuendorf gauge station by
  different combination of eddy viscosity and bed
  resistance.
• 3) After calibration, calculate the discharge at
  location 3 and 4 as the default data to be used
  to compare with the hypothesis scenarios
  calculation.
• 4) Edit the mesh by building the channel at given
  location, calculate the discharge at location 3
  and 4 meanwhile release particle at the entrance
  of the old navigation channel between Bock and
  Hiddensee.
• 5) Based on above calculation draw conclusion on
  the model calibration and the influence of
  splitting the Hiddensee island into two parts.

6
2. Model setting up and calibration

• 2.1 model set up
• Original bathymetry, boundary condition data and
  wind data were given. these data were used to set
  up the first model.
• Three months hourly data for the boundary, wind
  data and measured water level, however due to
  time and computation effort it is unrealistic to
  run the whole time period simulation every time,
  so typical time period like maximum and minimum
  water level has to be modeling in order to get a
  better representation of the nature.

Boundary Condition
Shorelines
HD Model
Eddy viscosity
Bathymetry
Wind force
7
2. Model setting up and calibration

• 2.2 statistic analysis for the water levels
• Three month data was analyzed and split up into
  three categories

8
2. Model setting up and calibration

• 2.3 model calibration
• Model calibration was the key procedures for
  modeling of water area, the idea is to compare
  the model results with the record data at given
  locations (Bodden, Ostee and Strasund station
  gauge).
• The parameter that can be modified are eddy
  viscosity and bed resistant.
• Our main goal for calibration is to get a better
  fit between calculation result and measured data.
• At least 5 calculations to be taken with two eddy
  viscosity and two bed frictions.
• 1st Default viscosity and bed resistant, 2nd
  Default viscosity and smaller bed resistant, 3rd
  Default viscosity and larger bed resistant

9
2. Model setting up and calibration

• 1st run results (default eddy viscosity and bed
  friction)

10
2. Model setting up and calibration

• 1st run results (default eddy viscosity and bed
  friction)
• From these three figures we can see that, the
  model gives more or less reasonable results, with
  the same pattern as the observed water level at
  three different gauging stations. But the peak
  values of the model results are less than the
  actual ones, for each gauging station. And the
  non-peak values of the model result are bigger
  than the actual values.
• The next step is to change the parameters of the
  model to get nearer results as the observed ones
  as possible.

11
2. Model setting up and calibration

• 2nd run results (default viscosity and larger
  friction)

12
2. Model setting up and calibration

• 2nd run results (default viscosity and larger
  friction)
• This combination of the model parameter gives a
  little better result but not differ much as the
  first two runs. Without eddy viscosity the whole
  water level shifts to the right a certain time.
• The model acts as some kind of average the water
  level in the whole time domain. Another reason
  might be the manually change in the mesh file we
  delete all the bathymetry larger than 5 meter.
• 3rd run (default viscosity and smaller friction)
• 4th run (smaller viscosity and default friction)
• 5th run (larger viscosity and default friction)
• Results not be present here, they looks very
  similar to each other

13
2. Model setting up and calibration
Does this graph looks better?

• Total period simulation!

14
2. Model setting up and calibration

• Total period simulation!
• The figure is the model results and gauging
  station record comparison from the figure we can
  see that the full-time simulation does not get
  better result than the smaller time simulation.
  (Visually, the figures give a better model
  results, but actually, it looks better only
  because we plot a long time series not because
  the results itself became better) This indicates
  the problem is not at the simulation time.
•  
• So the not fitting with measured data is probably
  due to the not well representative mesh. (Cutting
  the bathymetry to 5m)

15
2. Model setting up and calibration

• Conclusion for calibration
• After many calculations, a conclusion can be
  drawn that, for this case by averaging the
  bathymetry, it is difficult to get awell
  calibrated model due to the incorrect
  representation of the mesh.
• However, experiences about how bed friction
  and eddy viscosity influence the system were
  obtained. if high bed friction number is used in
  the model, it will smooth the calculation result.
  
• Eddy viscosity act as the phase shift of the
  calculation results (but its affect is not as
  sensitive as bed friction)

16
3. Gird editing and scenario analysis

• 3.1 calculate discharge and particle tracking
  before modify the gird (prepare for compare the
  effects)

Accumulated discharge !
17
3. Gird editing and scenario analysis

• 3.1 particle tracking before modify the gird
  (prepare for compare the effects)

18
3. Gird editing and scenario analysis

• 3.1 particle tracking before modify the gird

19
3. Gird editing and scenario analysis

• 3.2 Gird editing
• A channel of 100 meter width was built in
  Hiddensee

20
3. Gird editing and scenario analysis

• 3.2 calculation result after grid editing

21
3. Gird editing and scenario analysis

• 3.2 calculation result after grid editing

22
3. Gird editing and scenario analysis

• 3.2 calculation result after grid editing
• Form these three discharge figures we can see
  that, the splitting of Hiddensee has some
  influence on water discharge at section 2 and 3,
  when the Hiddensee island fall into two parts,
  the water discharge at section 2 and 3 become
  smaller than the situation without changing.
• This is what can be understandable and expected
  to happen. Also, part of the water flow through
  the opening channel, this result into the water
  volume decrease at nearby location. This can also
  be seen from the vectors of the calculation
  indication the water flow direction. However, it
  does not affect section4 so much this may
  probably due to it is far away from the small
  channel.

23
3. Gird editing and scenario analysis

• 3.2 calculation result after grid editing
  (particle tracking)

24
4. Conclusion

• Setting up and analysis of Mike21 Hydrodynamic
  module and particle tracking module.
• Tackle with the limitation of number of nodes and
  elements in 2D modeling.
• In the calibration step, high friction smooths
  the calculation results, eddy viscosity shift the
  water level phase.
• Due to poor representation of the bathymetry
  (limitation on element number), good calibration
  results can not be reached only by changing the
  bed friction and eddy viscosity.
• After cutting off the Hiddensee Island, water
  discharge at section 2 and 3 was influenced and
  decreased the volume of discharge, section3 had
  not much change as before.
• Sediment transport was slightly affected by the
  Hiddensee cutting off.

25
Thanks! QA