Ivan-like hurricane storm surge

simulations for Tampa Bay, FL

with 3-D and 2-D models

R.H. Weisberg and L. Zheng

Storm Surge Workshop

St. Pete Beach

2/11/09

Inundation based on uniform sea level rise of:

5 ft 20 ft

Potential for Tampa Bay area inundation is

large due to land elevation and geometry

A house on the Waveland, MS coastline

water-gutted from below.

Rte 90 bridge across Bay St. Louis, MS. All

spans were knocked from their supports.

Destroyed homes in South Diamondhead, MS.

Hurricane Storm Surge Simulation

Requirements

1. A high resolution, physics-based circulation model with

flooding and drying capabilities.

2. A high resolution water depth (bathymetry) and land

elevation data set on which to overlay the model.

3. Accurate enough wind and pressure fields to drive the

model.

Here we use the Finite Volume Coastal Ocean

Model (FVCOM) of Chen et al. (2003).

Overall Model Domain and Grid

The Ivan track (red dots) and the tracks (black dots) used in

this study (with landfalls as Sarasota, Indian Rocks Beach,

Tarpon Springs, Bayport, and Cedar Keys.

Ivan Winds on approach and at Landfall

While Ivan reached category 5 in the Caribbean it was a 4

upon approach and a 3 at landfall.Category mph knots m/s

1 74-95 64-82 33-43

2 96-110 83-95 44-49

3 111-130 96-113 50-59

4 131-155 113-135 60-70

5 >155 >135 >70

Surge elevation relative to mean sea level (left) and land elevation

(middle), plus wind vectors on wind speed contours (right) 3 hours before

and at IRB landfall (hrs. 27 and 30, respectively).

Maximum IRB landfall surge relative to land at sub-domains emphasizing

St. Pete Be. (left), Old Tampa Bay (middle), and Hillsborough Bay (right).

Time series of surge height sampled at selected locations

where:

ζ is the sea level,

H=h+ζ is the total water depth,

is the surface wind stress,

is the bottom friction stress, and

is the sum of the local and Coriolis accelerations,

the advective accelerations, and the horizontal diffusion,

each calculated separately before summation.

Rρ

τ

ρ

τζgH-

0

b

0

s

bτ sτ

R

Vertically Integrated Momentum Balance

Stress Parameterizations

wwads VVρCτ

1-

w

1-

w

1-

w

-1

w

3

d

ms25.0V250.0650.49

ms25.0Vms11.0V0.0650.49

ms11.0V1.2

10C

bbwzb VVρCτ

,0.0025])H/zσ[ln(1

kmaxC

2

01kb

2

z

-3

wd 10)V067.075.0(C

H

HC b

f 1C minf

Wind Stress

3-D FVCOM Cd

2-D IPET Cd

Bottom Stress

3-D FVCOM Cz

2-D IPET Cf

Parameter Examples

Wind Stress

For 50 ms-1 wind Cd (3-D) is 2.11x10-3, whereas Cd (IPET 2-D) is 4.1x10-3

Bottom Stress

For the 3-D FVCOM used here, Cz is capped at 0.005

For IPET Katrina ADCIRC Cf (2-D) is 0.003 at 2 m depth, 0.007 at 1m

depth, and larger for shallower depths.

Hence,

The IPET 2-D ADCIRC compensates for larger bottom stress by using

larger wind stress.

Absolute (black) and percent (red) differences between 3-D and 2-D

surges at four positions from the mouth to the head of the bay.

Conclusions

1. Local:

Tampa Bay, FL is as vulnerable to hurricane storm surge

inundation as was coastal Mississippi for H. Katrina.

2. General:

A. Storm surge simulation is sensitive to model

construction: i.e., 3-D, versus 2-D. The explanation is

bottom stress. A 2-D model overestimates bottom stress,

and hence underestimates surge.

B. Calibration can mitigate this (e.g., the IPET, Katrina

analyses are excellent), but forecasts without calibration

data may be in significant error.

C. Agencies (NOAA, FEMA, USACE) employ 2-D models. Our

findings suggest the importance of 3-D.

D. Studies are necessary to improve surface and bottom

stress parameterizations.

Acknowledgments

This work began with support by ONR, grant #s

N00014-05-1-0483 and N00014-02-1-0972, and it

continues with support from NOAA, grant #

NA07NOS4730211. The second of these was for the

Southeast Atlantic Coastal Ocean Observing System

(SEACOOS), and the third is related to the Southeast

Atlantic Coastal Ocean Observing Regional

Association (SECOORA). Changsheng Chen

(UMassD) kindly shared the FVCOM code.

Coupled Wave Effects

To investigate the effects of waves on the combined surge

and waves we (with Yong Huang) coupled an unstructured

version of SWAN to the FVCOM and repeated the

experiments. The following figures show: 1) significant wave

height, radiation stress and wind stress at several times

during the simulation, 2) surge heights with and without the

wave coupling by radiation stress, and 3) surge height

differences with and without the wave coupling. To the surge

heights with wave coupling must also be added the wave

amplitude (between approximately 0.5 to 0.7 times the wave

height) to get the total vertical reach of water over the

evolution of the storm.

Surge height with/without wave radiation stress.

Surge height difference with/without wave radiation stress.

Significant wave height, radiation stress, and wind stress

Significant wave height, radiation stress, and wind stress