Sensitivity Studies of the New Coastal Surge and Inundation
Prediction System Andrew Condon1,2 and Jay Veeramony2
1American Society for Engineering Education, 2Naval Research Laboratory, Stennis Space Center, USA
[email protected], [email protected]
4. Baseline Simulation The baseline simulation captures the
coastal water levels accurately but does
not perform as well in capturing the
inundation due to the coarse domain
resolution.
Figure 2. Domains, station locations, and track of Hurricane Ike used in hindcast
simulations
2. The Modeling System Delft3D is an integrated modeling suite, which simulates 2D and
3D flow, sediment transport and morphology, waves, water quality
and ecology. This study utilizes two Delft3D modules:
• FLOW – hydrodynamic module
• WAVE – wave module
FLOW is dynamically coupled to WAVE by passing water level,
currents, winds, and bed level to WAVE and in return gets
radiation stresses for wave setup calculations.
The Delft Dashboard (Figure 1) is a GUI specific to Delft3D
FLOW/WAVE and allows for the rapid setup of storm surge and
inundation simulations, including:
• Selection of an area from the world map; specifying the size
and resolution to create a rectangular grid
• Interpolation of bathymetry from available datasets
• Specifying type of boundary conditions and forcing
• Nesting domains and defining observation points
• Defining wave parameters and FLOW / WAVE coupling
6. Conclusions As expected the model results are very sensitive to a number of
input parameters
• Proper grid resolution is necessary for quality inundation
results, can get by with coarse resolution to model coastal
water levels
• Bathymetry / Topography data matter – High resolution data
can have a large influence even on a coarse grid
• Land cover / bottom roughness data is needed to improve
inundation simulation results
References • Amorocho, J and JJ DeVries, "A new evaluation of the wind stress coefficient over water surfaces,"
J. Geophys. Res., vol. 85, no. C1, pp. 433 - 442, 1980.
• Donelan, MA, et al., "On the limiting aerodynamic roughness of the ocean in very strong winds,"
Geophys. Res. Let., vol. 31, no. L18306, 2004.
• Garratt, JR, "Review of drag coefficients over oceans and continents," Mon. Wea. Rev., vol. 105,
pp. 915 - 929, 1977.
• Holland, GJ, "An analytic model of the wind and pressure profiles in hurricanes," Monthly
Weather Review, vol. 108, pp. 1212 - 1218, 1980.
• Jarosz, E, Mitchell, DA, Wang, DW and WJ Teague, "Bottom-up determination of air-sea
momentum exchange under a major tropical cyclone," Science, vol. 315, pp. 1707 - 1709, 2007.
• Large. WG and S Pond, "Open ocean momentum flux measurements in moderate to strong winds,"
J. Phys. Ocean., vol. 11, pp. 324 - 336, 1981.
• Mattocks, C, and C Forbes, "A real-time, event-triggered storm surge forecasting system for the
state of North Carolina," Ocean Modelling, vol. 25, pp. 95 - 119, 2008.
3. Domains The sensitivity studies involved hindcast simulations of Hurricane
Ike (2008) in the multiple domains depicted in Figure 2.
5. Sensitivity Studies
Model setup Gulf of Mexico domain
Bathymetry / Topography data:
• SURA inundation testbed GoM dataset
• NOAA NGDC Coastal Relief Model
• SRTM topography data
• GEBCO 08
Riemann (weakly reflective) boundary conditions
Tidal constituents:
• OSU global model of ocean tides based on
TOPEX7.2 satellite altimeter data
Initial water level: 0.11 m – seasonal trends
Wave coupling every 60 minutes
Variable bottom roughness (Manning’s N coefficient)
Wind drag coefficient from Powell et al. (2012)
Reanalysis winds from OceanWeather Inc. (OWI)
• 0.02° spatial resolution
• 10 min temporal resolution
NH21A-1583
1. Overall Objective To develop a tool (GUI) for storm surge and inundation prediction
which results in a minimization in man-hours and required operator
knowledge for model set-up as well as result in more optimal surge
and inundation forecasts.
Figure 1. The Delft Dashboard Graphical User Interface
Gulf of Mexico Domain • 0.1° x 0.1°
• 179 x 129
• 1 day runtime ~ 30
min @ Δt = 10 min
Nearshore Domain • 0.02° x 0.02°
• 462 x 302
• 1 day runtime ~ 60
min @ Δt = 5 min
Coastal Domains • 0.004° x 0.004°
• Various sizes
• 1 day runtime ~ 30
min @ Δt = 1 min
LAWMA Freshwater
Canal Locks
Calcasieu
Pass
Sabine
Pass North
Galveston
Pleasure Pier
High Water
Marks
-12.34 -0.59 0.78 4.49 1.72 23.72
Table. 1. Percent error of peak water level at 5 NOAA stations and average for all HWMs
Figure 3. Hydrograph analysis at five coastal
locations and HWM analysis for baseline simulation
Air-Sea Drag Coefficient Considered 12 formulations (Figure 4)
Found best results with newer formulations:
• Formulation of Powell et al. (2012)
produced the best results
• Peak CD at hurricane force
• Decrease in CD with higher wind speeds
Figure 4. 12 Drag coefficient formulations used in
sensitivity studies as a function of wind speed
Grid Resolution / Nesting Water level results are good with coarse resolution.
Inundation results improve with increasing resolution
as in Figure 5.
• 0.5° domain
• 2 – way nesting
• 0.1° to 0.004°
• 3 – way nesting
• 0.1° to 0.02° to 0.004°
Figure 5. HWM analysis for various 0.1, 0.05 resolution
GoM domains and 2-way and 3-way nesting
Bathymetry / Topography High resolution bathymetry / topography is not available worldwide. GEBCO bathy/topo and SRTM
topo are widely available and provide global coverage, but results are not as good (Figure 6).
GEBCO - Baseline GEBCO + SRTM - Baseline
Figure 6.
Comparison of
Baseline
simulation with
GEBCO only
bathy / topo
dataset and
GEBCO +
SRTM bathy
topo dataset
Forecast wind model Considered 3 analytic wind models and developed a new model.
Example: Ws = 95 KT, Rmax = 50 n. mi., Pn= 1007 hPa, Pc = 954 hPa,
RO = 300 n. mi.
64 KT… 105NE 90SE 60SW 75NW
50 KT… 150NE 150SE 90SW 105NW
34 KT… 240NE 205SE 150SW 175NW
Figure 7. Bathymetry and topography differences between GEBCO and GEBCO + SRTM and
Baseline
Figure 8. Wind snapshot from OceanWeather reanalysis
winds, 3 analytic wind models and newly developed model
Figure 9. Example of construction of wind field for
new wind model
• Place forecast
information on grid
(red data points)
• Obtain Holland B for
black data points
• Obtain wind speed at
black points using
Holland equation
• Use multivariate
interpolation to
determine velocity as a
function of distance
and angle from center Figure 10. HWM comparison between various wind models
Proper wind forcing is needed
• Many modeling studies use
older drag formulations with a
simple cap; newer formulations
may offer better results
• Various analytic wind models
can offer very different results
• New model offers promising
results … more testing is
needed
• Moon, IJ, Ginis, I, Hara, T and B Thomas, "A physics-based parameterization of air-sea
momentum flux at high wind speeds and its impact of hurricane intensity predictions," Mon. Wea.
Rev., vol. 135, pp. 2869 - 2878, 2007.
• Powell, MD, Vickery, PJ and TA Reinhold, "Reduced drag coefficient for high wind speeds in
tropical cyclones," Nature, vol. 422, pp. 279 - 283, 2003.
• Powell, MD, "High wind drag coefficient and sea surface roughness in shallow water," Final
Report to the Joint Hurricane Testbed, NOAA HRD-AOML, 2008.
• Powell, MD, Holthuijsen, L, and J Pietrzak, "Spatial variation of surface drag coefficient in
tropical cyclones," unpublished, 2012.
• Wu, J, "Wind-stress coefficients over sea surface from breeze to hurricane," J. Geophys. Res., vol.
87, no. C12, pp. 9704 - 9706, 1982.
• Xie, L, Bao, S, Pietrafesa, LJ, Foley, K, and M Fuentes, "A real-time hurricane surface wind
forecasting model: Formulation and verification," Mon. Wea. Rev., vol. 134, pp. 1355 - 1370,
2006.
• Zachry, BC, Letchford, CW, Zuo, D, Schroeder, JL, and AB Kennedy, "Surface drag coefficient
behavior during Hurricane Ike," in 11th Americas Conf. on Wind Engineering, San Juan, PR, 2009.
• Zijlema, M, van Vledder, GP, and LH Holthuijsen, "Bottom friction and wind drag for wave
models," Coastal Eng., vol. 65, pp. 19-26, 2012.