Models and modellers for coastal flooding Presented by Dr. Ole Svenstrup Petersen,
[email protected]
Head of Research Coastal and Estuarine Engineering, DHI And colleagues at DHI
Agenda • • • •
Introduction to DHI and DHI’s models Examples of model applications and validation relevant for coastal flooding Examples of model coupling Discussion of points of interest
© DHI
22 May, 2013
#2
About DHI • • • •
Founded 1963 as Coastal Lab from Danish Technical University Independent, not-for-profit foundation All about water – Marine, coastal, inland, urban and industrial 30 % Research - 20 % Software – 50 % international consulting
© DHI
22 May, 2013
#3
DHI OFFICES
Offices Head office RD Centre © DHI 2013
1100 Staff, 700 outside DK in 36 offices RD centers in Denmark and Singapore
A modeling timeline – the history of DHI’s models • 1958-59 Thyborøn Inlet, Lundgren and Sørensen (DHI’s first director) develop a 14-box diffusivewave storm surge model of the Limfjord on DASK
© DHI
22 May, 2013
#5
A modeling timeline – the history of DHI’s models • • • • • • • • • •
1959 Limfjord model on DASK 1962 Mike Abbott, IHE, initiates DHI’s modeling team 1983 MIKE 21 commercial software products 1986 MIKE 3 1998 Start of FM development 2000 Generalized Wave equation, 2D and 3D 2004 Finite Volume 2D and 3D with flooding and drying 2004 First Spectral Wave model based on Flexible Mesh 2012 MIKE 3/21 on 1000 cores 201X GPU, non-hydrostatic, BW ……
© DHI
MIKE 11 Limfjord model 22 May, 2013
#6
DHI Modeling systems
DHI’s modeling platforms - MIKE by DHI
• A platform for customised software building on top of the modeling engines in order to make results more accessible and useful for the public •
http://www.dhigroup.com/MIKECUSTOMISEDbyDHI.aspx
MIKE by DHI Marine models
http://mikebydhi.com/Products/CoastAndSea.aspx
© DHI
22 May, 2013
#9
MIKE 3/21 basics • Shallow water equations (hydrostatic) in cartesian or radial projections • Fractional step with Partial mode split • Explicit time integration – strict conditions for Courant numbers • Adaptive time stepping – global (for HD) or local (for SW) • 3 step flooding and drying • Finite volume method • Unstructured horizontally • Sigma or combined sigma-z vertically
Know more ? http://mikebydhi.com/~/media/Microsite_MIKEbyDHI/Publications/PDF/Short%20descriptions/MIKE213_FM_HD_Short_Description.ashx
Vertical Discretization in MIKE 3 FM MIKE 3/21 Spatial Combined sigma and z-layers Improved description along steep slopes
Better description of stratified layers Accurate description of heat exchange
-level
Z-level
MIKE 21 Spectral Wave • • • • •
Fully spectral (FS) and directionally decoupled parametric (DS) formulations Instationary and quasi-stationary formulations with adaptive and local timestepping Source functions based on state-of-the-art formulations (WAM cycle 4) Unstructured mesh Dynamic coupling with hydrodynamic flow model for modeling of wave-current interaction and time-varying water depth
Know more ? http://mikebydhi.com/~/media/Microsite_MIKEbyDHI/Publications/PDF/Short%20descriptions/MIKE21_SW_FM_Short_Description.ashx © DHI
22 May, 2013
#12
MIKE 21 BW (BOUSSINESQ WAVE) • Time domain,deterministic free surface model • Calculation and analysis of short- and longperiod waves in ports, harbours and coastal areas. • Solves a set of enhanced Boussinesq type equations • Includes diffraction, refraction, breaking, runup, overtopping, non-linear, random directional seas • Wave agitation in harbours and wave assessments in the coastal zone
© DHI 2013
MIKE 21 BW (BOUSSINESQ WAVE)
Know more ? http://mikebydhi.com/~/media/Microsite_MIKEbyDHI/Publications/PDF/Short%20descriptions/MIKE21_BW_short_description.ashx © DHI 2013
Sand Transport Processes – MIKE 21 STP • Building mainly on Fredsøe and Dejgaard’s work • Uses a profile ”intra-wave parameterisation” to calculate stress, turbulence and transport in wavecurrents • Can use the wave and flow models to calculate forces • Feed back on bathymetry
Know more ? http://www.mikebydhi.com/upload/dhisoftwarearchive/shortdescriptions/marine/SandTransportModuleST.pdf
© DHI
22 May, 2013
#15
Model calibration and validation • A key activity in modeling • The amount of time/data used is directly reflected in model quality • Software quality assurance • • • •
Making the right decisions ISO certification Daily build and basic tests, update procedures New feature tests in real environment
• Application/projects • Use of relevant data is important • Typically 50 % for model calibration/validation
© DHI
22 May, 2013
#16
Tide and surge modeling, Sakhalin • Metocean study of waves and currents
Tidal calibration • MIKE 21 FM hydrodynamic model • DTU10 tidal data on boundary and inside • Customized wind fields (OWI)
The (modern) Limfjord models • Study of impact on storm surge from morphological changes in Thyborøn Inlet
The Limfjord
© DHI
22 May, 2013
#19
Model calibration and model verification Six storms were selected to calibrate and verify the model Performance criteria was defined 1.8 1.6 1.4
1.2 1 0.8 0.6
2.2
Lemvig Havn
1.8 1.6 1.4
1.2 1 0.8 0.6
2.2
Løgstør Havn
2
1.8 1.6 1.4
1.2 1 0.8 0.6
2.2
Thisted Havn
1.8
2
1.6 1.4
1.2 1 0.8 0.6
1.6 1.4
1.2 1 0.8 0.6
0.4
0.4
0.4
0.4
0.2
0.2
0.2
0.2
0.2
0
0
0
0
Målt vandstand [m]
Målt vandstand [m]
Målt vandstand [m]
2.2
Modelleret vandstand [m]
2
1.6 1.4
1.2 1 0.8 0.6
Hvalpsund
1.8
2
1.6 1.4
1.2 1 0.8 0.6
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
Målt vandstand [m]
2.2
Modelleret vandstand [m]
Nykøbing Mors
1.8
0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
Målt vandstand [m] 2.2
2.2
Rønbjerg
2
Modelleret vandstand [m]
2.2
2
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
1.8 1.6 1.4
1.2 1 0.8 0.6
Ålborg Øst
2
Modelleret vandstand [m]
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
1.8 1.6 1.4
1.2 1 0.8 0.6
1.6 1.4
1.2 1 0.8 0.6
0.4
0.4
0.4
0.4
0.2
0.2
0.2
0.2
0.2
0
0
0
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
Målt vandstand [m]
© DHI
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
Målt vandstand [m]
22 May, 2013
Målt vandstand [m]
#20
Alle stationer
1.8
0.4
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
Skive Havn
1.8
0.4
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
Modelleret vandstand [m]
2
Modelleret vandstand [m]
2
Modelleret vandstand [m]
Thyborøn Havn Modelleret vandstand [m]
Modelleret vandstand [m]
2
Modelleret vandstand [m]
2.2
2.2
0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
Målt vandstand [m]
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
Målt vandstand [m]
Model calibration and model verification Example: Comparison of modelled and observed water level at Lemvig
© DHI
22 May, 2013
#21
Femern Belt Link • 20 km combined hi-speed rail and car bridge or tunnel between Denmark and Germany • Planned completion 2021 • Very large EIA partly due to concern for Baltic ecosystem
© DHI
22 May, 2013
#22
MIKE 3 Model set up for Femern Belt Regional model • 3-5 km, 1 m vertical • 30+ years
Local model • 500 m, 1 m vertical • 1 year
Regional MIKE 3 Oceanographic Model 30 years simulation
Example of Calibration of 3D model
The model does reproduce the major inflows of saline water that are crucial for the ecosystem of the Baltic Sea
MIKE 21 SW Global wave prediction
0.5° spatial resolution requires app. 2.7 GB RAM 0.25° spatial resolution requires app. 6 GB RAM
25
Global wave prediction January storm 2005 in the North Atlantic
26
Horns Rev Objective (waves) Establishment of 4 day wave forecast for planning of marine operations (maintenance etc.)
(m)
Offshore wind farm with 80 turbines (168 MW) 27
Horns Rev Atlantic model
Number of elements:
8334
Maximum edge length: 1.0 degree Minimum edge length: 0.5 degree
(m)
North Sea-Baltic Sea model Number of elements: 8985 Maximum edge length: 40 km Minimum edge length: 1.5 km
(m)
28
Horns Rev Measurements stations Station Harald W Fjaltring Horns Rev S
Longitude (Degree E) 4.219722 8 058221 7.836733
Latitude (Degree N) 56.338056 56.475075 55.483617
Depth, MSL (m) 65.0 17.5 10.0
Fjaltring Harald W Horns Rev S (m)
29
Horns Rev Fjaltring Significant wave height, Hm0
Wave period, T02
Horns Rev S Significant wave height, Hm0
30
Wave period, T02
Horns Rev Fjaltring - frequency spectrum
Horns Rev S - frequency spectrum
31
Kirkwall Marina, UK • MIKE 21 Spectral Waves • MIKE 21 Boussinesq Wave • Physical model
*) Kirkwall Marine is used in the MIKE 21 BW step-by-step-guide
Kirkwall Marina Results from MIKE 21 SW*)
• Comparison between SW and BW
Results from MIKE 21 BW
*) Using the DS (directionally decoupled parametric formulation) with diffraction and reflection
MIKE 21 Spectral Waves FM - Kirkwall Marina, UK
Kirkwall Marina • Comparison between BW and Lab inside harbour
Comparison with measurements
Coastal flooding - two perspectives 1.
Extreme events and their consequences - short term
• Wave impact on sea defense => Breaching => Hinterland inundation.
2.
Long term changes in the coast’s resillience to potential extreme events • Dune and dike stability. • Importance of foreshore. • Natural evolution versus anthropological activity (including climate adaptation).
Dune breaching
Dune erosion Dune breach at Danish Westcoast 1991
Cross shore sediment transport in the surf zone
© DHI
22 May, 2013
#36
Validation of dune erosion model against results from DELTA flume (DELTARES) Hs=1,5m Tp=4,9s
Measured dune erosion
Modelled dune erosion during 6 hr
The dune erosion model illustrates that these narrow dunes are best strengthened by adding sand behind the crest
Which is being done…
Comparison: dynamic simulation >< stationary analysis based on topographic info only:
Extreme 40 hour storm event with max water level in the sea of 2.2 m ( i.e. present day extreme event). Max flooded area to the left. Estimate based on topography to the right (2 m contour line).
Dune overtopping • MIKE 21 BW • Hydrodynamics as storm surge events • Parameterisation of overtopping • Combined with Dune erosion and sand transport
© DHI 2013
Ribe Polder flooding • Flood Risk analysis of Polder • Dynamic breach and flood modelling with MIKE 21 • Carried out by Danish Coastal Authority as part of Climate Change Adaptation study
Levee breaching, Ribe • Breaching dynamics from detailled model
Hinterland indundation MIKE21 FM, HD – Development of breach
Hinterland indundation MIKE21 FM, HD – Development of breach
Hinterland indundation MIKE21 FM, HD – Development of breach
Know more ? http://www.dhigroup.com/Aboutus/DHIAnnualReport2011/MakingKnowledgeAccessible/GuidanceOnClimateChangeAdaptation.aspx
Longterm - Hybridmodel • Under development ! • Combines Flow, wave and STP models for long term morphology • Quasi-steady flow and waves • Morphology described in profiles
Know more ? Kaergaard, K. and Fredsoe, J. (2013). A numerical shoreline model for shorelines with large curvature. Coastal Engineering, Vol. 74, p. 19-32. Kaergaard, K., Fredsoe, J. and Knudsen, S. B. (2012). Coastline undulations on the West Coast of Denmark: Offshore extent, relation to breaker bars and transported sediment volume. Coastal Engineering, Vol. 60, p. 109-122
© DHI
Hybrid model – spit development
© DHI
Model coupling • Loads on wind turbine foundation, MIKE 21 BW + CFD
Nielsen, A.W, Mortensen, S. B., Christensen, E.D (2008) “Numerical Modeling of Wave Run-up on a Wind Turbine Foundation”, Proc. OMAE 2007, Lisbon, Portugal. © DHI
22 May, 2013
#48
Model coupling • Levee overtopping using MIKE 21 BW + CFD
© DHI
22 May, 2013
#49
Model coupling - Hvide Sande harbour
Distance along shoreline (km)
• Study of sand bypass using MIKE 21 HD + SW + ST
Shoreline evolution (m/year)
1957-2008 profile data
~400 m northern breakwater established in 1963: • updrift accumulation up to ~3,5 km north of breakwater in total 3,600,000 m3 80.000 m3/yr • shoreline advance ~3 m/yr 500 m to the north downdrift erosion counteracted by nourishment of 300,000 m3/yr along 2-4 km to the south
Sedimentation in access channel Navigation depth of 4,5 m is maintained
maintenance dredging ~170,000 m3/yr hydraulically pumped to downdrift shore
Yellow colour: depths between 3 and 4 m
With the present harbour layout, the natural depth of bypass bar is ~2.5 m
Calibration results
Bar crest elevation across the access channel during calibration period
Measured pre-storm Measured post-storm Modelled post-storm
Accurate reproduction of bypass bar evolution: 350 m southward migration of the bar crest elevation along the bar from -2.5 to -3.5 m
The Limfjord model anno 2012 Key problems in Thyboron Harbour after 2004: • Larger waves in the entry area • Navigation • Overtopping • Wave disturbance in the harbour • Sedimentation causing reduced depth in the harbour entry area http://www.youtube.com/watch?v=wPRQDpXHj6w
North Sea
Thyborøn Thyborøn Channel Harbour
Limfjord
The Limfjord • Since 2004 increasing overtopping at outer breakwater in Thyborøn Harbour • Possible cause in morphological active entrance channel
• Approach • Model sediment dynamics in entrance using HD + SW + ST • Model wave activity using Boussinesque model (MIKE 21 BW)
Thyborøn Thyborøn Channel Harbour
Niemann, S.L., Sloth, P., Buhl, J., Deigaard, R. and Brøker, I. Thyborøn Harbour - Study of Wave Agitation and Sedimentation. ICCE No 32 (2010): Proceedings of the 32nd Conference on Coastal Engineering, Shanghai, China, 2010.
© DHI
22 May, 2013
North Sea
Limfjord #54
Historical benchmark for Danish Coastal Engineers Before 1862: The shoreline was located more than 1500m further offshore and there was no connection between the North Sea and the Limfjord 1862: Large breach occurred North Sea
Limfjord
1875-1930’ies: A large groyne system were constructed along the 22 km barrier Today: Coastline is mitigated by nourishment of about 750000 m3/year
Change in bathymetry have changed the wave conditions Reduced depth and increased bed slope
Erosion, steeper beach
2004
2008 Reduced depth
Impact on waves of 400.000 m3 excavation 2008
After excavation
Hs decreases to 50-70% Hs increases by 10-50%
Summary • Model have developed since 1960 and we have been able to make use of this • Model and data should be an integrated entity
[email protected]
www.dhigroup.com
© DHI
22 May, 2013
#59
