11URS Team

Evaluation of the StormEvaluation of the StormSurge Hazard inSurge Hazard in

Coastal MississippiCoastal Mississippi

A. Niedoroda, D.Resio, G. Toro; D. Divoky, H. Das, C. Reed

22URS Team

ACKNOWLEDGEMENTSACKNOWLEDGEMENTSProject Staff:

Alex Whitworth, Richard Sanborn, Saiduz Zaman, Bikash Saha, Candace Beauvais, Dale Lehman, Lynn Mayo, Darryl Hatheway,

Maria Honeycutt, Jonica Gibson, Karen Schuckman, Heather Zhao, William O’Brien, Mark Johnson, David Divoky, Cheryl

Johnson, Ki Pak, Vince Cardonne, Andrew Cox, Jeffrey Gangai, Kevin Slover, Elena Drei-Horgan, Kumiko Yamazaki, Darren

Rose, Robert Dean, Don Slinn, Robert Weaver, Norm Scheffner, Leon Borgman, Stephen Baig, Gabriel Toro, Peter Vickery, and

Todd Walton

FEMA Staff:Rob Lowe, Dr. Shabbar Saifee, Lynda Charles, Jason Hunter

Henrietta Willliams, John LaBrune, Emile Hurst, Doug Bellomo

Companies:Risk Engineering, Watershed Concepts, Dewberry, Ayres Assoc.,

Computational Hydraulics,

33URS Team

MOTIVATIONMOTIVATION

New Flood Maps Needed New Flood Maps Needed (especially after (especially after Hurricane Katrina)Hurricane Katrina)New Methodology New Methodology NeededNeeded

44URS Team

NEW METHODOLOGYNEW METHODOLOGYStorm CharacterizationStorm Characterization

Numerical ModelsNumerical Models

JPM JPM –– JPMJPM--OSOS

Coastal Flood ElevationsCoastal Flood Elevations

55URS Team

SUMMARY OF CONCLUSIONSSUMMARY OF CONCLUSIONS

New working methodology for FEMA Coastal New working methodology for FEMA Coastal Flood Studies has been developedFlood Studies has been developedEfficient JPM has been shown to be accurate Efficient JPM has been shown to be accurate and efficientand efficientWave setup included within the Wave setup included within the hydrodynamic simulations improves the hydrodynamic simulations improves the resultsresultsResults are sensitive to subtle processesResults are sensitive to subtle processesNew efficient methods have been New efficient methods have been corroborated over related flood study corroborated over related flood study projectsprojects

66URS Team

PROJECT ORGANIZATIONPROJECT ORGANIZATION

Characterize the local storm climateCharacterize the local storm climateDevelop method of forward projectionDevelop method of forward projectionCreate surge heights with numerical Create surge heights with numerical modelsmodelsAnalyze recurrence statisticsAnalyze recurrence statistics

77URS Team

CharacterizationCharacterization

Storm ParametersStorm ParametersPeriod of RecordPeriod of RecordSchematic ShorelineSchematic ShorelineOptimize the Sample Error/Population Optimize the Sample Error/Population Error Trade OffError Trade OffParameter StatisticsParameter Statistics

88URS Team

Period of RecordPeriod of Record

1851:1851: Limit of Limit of HurdatHurdat1900:1900: Former FEMA Data LimitFormer FEMA Data Limit1940:1940: Offshore Aircraft ObservationsOffshore Aircraft Observations1960s:1960s: Satellites and Data BuoysSatellites and Data Buoys1990s:1990s: Doppler RadarDoppler Radar2006:2006: Limit Of Study DataLimit Of Study Data

99URS Team

Period of RecordPeriod of Record

1851:1851: Limit of Limit of HurdatHurdat1900:1900: Former FEMA Data LimitFormer FEMA Data Limit1940:1940: Offshore Aircraft ObservationsOffshore Aircraft Observations1960s: 1960s: Satellites and Data BuoysSatellites and Data Buoys1990s:1990s: Doppler RadarDoppler Radar2006:2006: Limit Of Study dataLimit Of Study data

1010URS Team

Schematic ShorelineSchematic Shoreline

100:W

100:W

95:W

95:W

90:W

90:W

85:W

85:W

80:W

80:W

20:N 20:N

25:N 25:N

30:N 30:N

Generalized Coastline

CRP

1111URS Team

Storm ParametersStorm ParametersPressure deficitPressure deficitPressure radiusPressure radiusAzimuthAzimuthLandfallLandfallForward speedForward speedHolland BHolland BAstronomical tideAstronomical tideTrack variationsTrack variationsWind/Pressure Field Wind/Pressure Field DetailsDetails

Major ParametersMajor Parameters–– Pressure deficitPressure deficit–– Pressure radiusPressure radius–– AzimuthAzimuth–– LandfallLandfall–– Forward speedForward speed

Special ParameterSpecial Parameter–– Holland BHolland B

Random ParametersRandom Parameters–– Astronomical tide Astronomical tide –– Track variationsTrack variations–– Wind/pressure field Wind/pressure field

variationsvariations–– Lack of model framework Lack of model framework

skillskill

1212URS Team

Optimize Spatial Resolution Optimize Spatial Resolution vsvs Statistical PrecisionStatistical Precision

Figure 3-1. Analysis of hurricane frequency from Toro (Risk Engineering) from an analysis using an optimized spatial kernel.

Optimal kernel Optimal kernel –– (width 200 (width 200 km)km)Total of 30 Storms (1940Total of 30 Storms (1940--2006; 852006; 85oo--9595ooW)W)Divide: Divide: ΔΔP>48 P>48 mbmb(14);31>(14);31>ΔΔP>48 (16)P>48 (16)ChouinardChouinard Method (Method (λλ,,ΔΔP,P,θθ))–– Weights based on trackWeights based on track--

toto--CRP distanceCRP distance

Conditional Conditional RpRp | | ΔΔPPSimple fit for USimple fit for UHurricane frequency (from Toro)

1313URS Team

Direction & RateDirection & Rate

Annual Rate 5.45E-4 storms/km/yr

0.E+00

2.E-07

4.E-07

6.E-07

8.E-07

1.E-06

1.E-06

1.E-06

2.E-06

-180 -150 -120 -90 -60 -30 0 30 60 90 120 150 180

Heading (degrees)

Rat

e (s

torm

s/km

/deg

/yr)

CalculatedNormal Fit

Lesser 2.57E-4 storms/km/yr

0.E+00

5.E-07

1.E-06

2.E-06

2.E-06

3.E-06

3.E-06

4.E-06

-180 -150 -120 -90 -60 -30 0 30 60 90 120 150 180

Heading (degrees)R

ate

(sto

rms/

km/d

eg/y

r)

Calculated RateBeta Fit

Greater 2.88E-4 storms/km/yr

1414URS Team

Pressure Pressure DeficitDeficit

Fitted with aWeibull Distributions

Complementary Cumulative Distribution of ΔP(Lesser+Greater Storms) at Coastal Reference Site (30.2 N, 89.3 W)

1.E-02

1.E-01

1.E+00

30 50 70 90 110

Pressure Deficit (mb)

Exc

eede

nce

Prob

abili

ty

Lesser + Greaterr Storms

Lesser Storms

Greater Storms

Probability Density Function of ΔP(Lesser+Greater Storms) at Coastal Reference Site (30.2 N, 89.3)

0.00

0.01

0.02

0.03

0.04

30 50 70 90 110

Pressure Deficit (mb)

Prob

abili

ty D

ensit

y (m

b-1) Lesser + Greater Storms

Greater StormsLesser Storms

1515URS Team

Pressure RadiusPressure Radius(depends on (depends on ΔΔP)P)

0

10

20

30

40

50

60

40 50 60 70 80 90 100 110 120 130

DP (mb)R

p (n

mi)

S of 29.5N of 29.584%Meanmedian16%

Model: ln(Rp[nmi])=5.91-0.711*ln(DP); sigma=0.44

0

20

40

60

80

100

120

30 40 50 60 70 80 90 100 110 120

DP (mb)

Rp

(nm

i)

Rp(offshore) vs. DP(coast) data

Rp vs DP (whole Gulf) 84%

Rp vs DP (whole Gulf) median

Rp vs DP (whole Gulf) 16%

Rp(o) vs DP(coast) 84%

Rp(o) vs DP(coast) median

Rp(o) vs DP(coast) 16%

Model: ln(Rp[nmi])=4.28-0.33*ln(DP); sigma=0.44

M

GREATER STORMSLESSER STORMS

DATA FROM WHOLE GULFOF MEXICO (1940-2006)DATA FROM WHOLE GULF

OF MEXICO (1950-2006)

Conditional Distribution Rp|ΔP => lognormal functions

1616URS Team

Forward SpeedForward Speed

0

0.05

0.1

0.15

0.2

0 2 4 6 8 10 12 14 16

Vf (m/s)Pr

obab

ility

Den

sity

Median (m/s): 6.04Logarithmic Std. Dev.: 0.41

Mean (m/s): 6.57

Std. Dev. (m/s): 2.79

Mean (m/s): 6.57

Std. Dev. (m/s): 2.79

Median (m/s): 6.04Logarithmic Std. Dev.: 0.41

Mean (m/s): 6.57

Std. Dev. (m/s): 2.79

GREATER STORMS

0

0.05

0.1

0.15

0.2

0.25

0 2 4 6 8 10 12 14 16

Vf (m/s)

Prob

abili

ty D

ensi

ty

Median (m/s): 5Logarithmic Std. Dev.: 0.43

Mean (m/s): 5.48

Std. Dev. (m/s): 2.47

Mean (m/s): 5.48

Std. Dev. (m/s): 2.47

Median (m/s): 5Logarithmic Std. Dev.: 0.43

Mean (m/s): 5.48

Std. Dev. (m/s): 2.47

LESSER STORMS

Forward speed => lognormal function

1717URS Team

Storm TracksStorm TracksWhole Gulf Tracks Curved & Similar to Corps’ BRICA

1818URS Team

AlongAlong--Track ProfilesTrack Profiles

For storms where Rp > 10 n. mi.

Rp (landfall) = 1.3 * Rp (offshore)

B (landfall) = 1.0 B (offshore) = 1.27

Decrease in ΔP(mb) = Rp (offshore n.mi.) -6

max ΔP = 18 mb, min ΔP = 5 mb

1919URS Team

Joint Probability Method (JPM)Joint Probability Method (JPM)

∫∫ >=>x Xyr xdxPxfP ])([)(...][ )1max( ηηληη

Annual occurrencerate

Surge from thenumerical modelingFrom

characterizationof parameters

2020URS Team

Joint Probability Method (JPM)Joint Probability Method (JPM)

])([][1

)1max( ∑=

>≈>n

iiiyr xPP ηηληη

∫∫ >=>x Xyr xdxPxfP ])([)(...][ )1max( ηηληη

Commonly approximated by the discrete function:

Rate for each storm Surge elevation fromnumerical modeling

2121URS Team

JPMJPM--OS Development StrategyOS Development StrategySLOSH model sensitivity testsSLOSH model sensitivity tests–– Yields relative importance of parametersYields relative importance of parameters

Create full JPM (Gold Standard) AnalysisCreate full JPM (Gold Standard) Analysis–– The 2,967 synthetic storms output at 147 points The 2,967 synthetic storms output at 147 points

form a basis of comparisonform a basis of comparisonUse Bayesian Use Bayesian quadraturequadrature method to create 4 JMPmethod to create 4 JMP--OS candidatesOS candidates–– OS4 = 303 , OS5 = 193, OS6 = 158 and OS7 = 147OS4 = 303 , OS5 = 193, OS6 = 158 and OS7 = 147

Compare JPMCompare JPM--OS candidates to the Gold Standard OS candidates to the Gold Standard (GS) and select (GS) and select ““bestbest””–– OS6 yield least synthetic storms while preserving OS6 yield least synthetic storms while preserving

GSGS--resultsresults

2222URS Team

Parameter Sensitivity TestsParameter Sensitivity Tests

Mobile

Jackson

St Tammany

Harrison

Hancock

Orleans

Lake Borgne

LakePontchartrain

Gulfof

Mexico

BiloxiBay

Bay St. Louis

¹Figure 1a

Preliminary Joint Probabillity Model Output Locations

Map Created November 29, 2006

Storm Track Line

LegendCoastal Stations

Riverine Stations

Inland_Stations

!

Coastal StationsVarying Central Pressure

0

5

10

15

20

25

28 38 48 58 68 78 88 98 108

Central Pressure Deficit (millibars)

Sto

rm S

urge

(ft)

Inland StationsVarying Central Pressure

0

5

10

15

20

25

0 20 40 60 80 100 120

Central Pressure Deficit (millibars)

Stor

m S

urge

(ft) Results:

ΔP > Rp > θ > U• Track spacing = Rp• 3 tracks west & 1 east

2323URS Team

Full JPM (Full JPM (““Gold StandardGold Standard””))

Mobile

Jackson

St Tammany

Harrison

Hancock

Orleans

Lake Borgne

LakePontchartrain

Gulfof

Mexico

BiloxiBay

Bay St. Louis

¹Figure 1a

Preliminary Joint Probabillity Model Output Locations

Map Created November 29, 2006

Storm Track Line

LegendCoastal Stations

Riverine Stations

Inland_Stations

!

Central Pressure Deficit (mb)

Normalized Probability 0.01100 0.22200 0.53400 0.22200 0.01100

45.6 0.04958 6.94 13.43 24.38 44.28 85.7148.6 0.16607 6.63 12.84 23.32 42.34 81.9656.1 0.28435 5.99 11.59 21.05 38.23 74.0069.1 0.28435 5.16 10.00 18.15 32.96 63.8087.6 0.16607 4.36 8.44 15.33 27.84 53.88111.8 0.04957 3.67 7.10 12.89 23.41 45.31

Forward Speed (m/s) 2.99 6.04 12.23Normalized Probability 0.16667 0.66667 0.16667

Headings (Theta) * -73.0 -32.7 7.3 49.4Normalized Probability 0.13299 0.36701 0.36701 0.13299

*(direction to; degrees clockwise from North)

Values of Rp(nmi)

360 Synthetic storms each on multiple tracks spaced at

Rp yields:2,967 model runs

2424URS Team

The Bayesian The Bayesian QuadratureQuadrature MethodMethod

])([][1

)1max( ∑=

>≈>n

iiiyr xPP ηηληη

∫∫ >=>x Xyr xdxPxfP ])([)(...][ )1max( ηηληη

2525URS Team

The Bayesian The Bayesian QuadratureQuadrature MethodMethod

DP(mb)

10 20 30 40 -60 -20 20 60

6080

100

120

1020

3040

Rp(nmi)

Vf(m/s)

46

810

1214

60 80 100 120

-60

-20

2060

4 6 8 10 12 14

Heading(deg)

JPM-OS6

2626URS Team

Comparison Of JMPComparison Of JMP--OS6 & OS6 & ““GSGS””Figure 32a: Comparison of Surge Elevation in JPM-OS-v1 v. JPM-Set2

02468

10121416

-90 -89.8 -89.6 -89.4 -89.2 -89 -88.8 -88.6 -88.4 -88.2

Point Location (Longitude)

Feet

100yr-JPM-Set2

100yr-OS-v4

Figure 32b: Error between Surge Elevations in JPM-OS-v1 and JPM-Set2

-5

-3

-1

1

3

5

-90 -89.8 -89.6 -89.4 -89.2 -89 -88.8 -88.6 -88.4 -88.2

Point Location (Longitude)

Feet

Error

6

GP100 Year Surge Elevations

Surge Elevation Differences

2727URS Team

The JPMThe JPM--OS6 CaseOS6 Case

StormID (OWI notation) dp(mb;coast) Rp(nmi;offshore) Vf(m/s) theta(deg) Prob. Annual Rate (each

JOS6###% track)JOS6% 66.69 18.61 6.047 -38.91 1.33E-01 1.32E-03JOS6% 57.17 39.82 6.047 -13.49 1.20E-01 2.55E-03JOS6% 49.72 22.93 6.047 -38.92 1.33E-01 1.63E-03JOS6% 57.17 10.83 6.047 -13.49 1.20E-01 6.94E-04JOS6% 57.17 20.77 6.047 56.66 1.08E-01 1.19E-03JOS6% 92.95 14.7 5.943 -12.81 3.42E-02 2.68E-04JOS6% 78.59 30.8 6.014 -12.82 5.34E-02 8.77E-04JOS6% 78.59 16.56 4.349 47.33 4.20E-02 3.71E-04JOS6% 78.59 8.904 6.014 -12.82 5.34E-02 2.54E-04JOS6% 78.59 16.56 14.54 -12.86 3.49E-02 3.08E-04JOS6% 70.02 17.98 5.943 -12.82 3.42E-02 3.28E-04JOS6% 78.59 16.56 4.346 -71.04 4.20E-02 3.71E-04JOS6% 128.7 11.66 5.943 -12.81 1.06E-02 6.58E-05JOS6% 103.7 25.3 6.014 -12.82 1.65E-02 2.23E-04JOS6% 103.7 13.6 4.349 47.33 1.30E-02 9.44E-05JOS6% 103.7 7.313 6.014 -12.82 1.65E-02 6.44E-05JOS6% 103.7 13.6 14.54 -12.86 1.08E-02 7.83E-05JOS6% 94.47 14.53 5.943 -12.82 1.06E-02 8.20E-05JOS6% 103.7 13.6 4.346 -71.04 1.30E-02 9.43E-05

2. The annual rate for each storm is calculated as the storm probability displayed here, times the annual rate of greater storms (2.88E-4 storms/km/yr), times the storm spacing (Rp) in km. 3. The annual rates in column I are the lambda terms in the report text

Notes

1: the Reference storms (e.g., JOS6001) are not assigned any rate. Only JOS6001A, JOS6001B, etc. are used in the probability calculations.

156 Greater Storms 72 Lesser Storms

Total of 228 Synthetic storms to be modelled

2828URS Team

Detailed TracksDetailed Tracks

72 LESSER STORMS152 GREATER STORMS

2929URS Team

““A Whole Bunch Of ModelingA Whole Bunch Of Modeling””

Risk Engineer – Storms and Tracks

D. Slinn50K ADCIRC & SWAN

URS TallahasseeADCIRC & RI Analysis

OWIWinds, Pressure & Deepwater Waves

WAM

3030URS Team

JPM With The JPM With The ““Random TermRandom Term””])([])([)(...][

1)1max( ∑∫∫

=

>+≈>+=>n

iimix mXyr xPxdxPxfP ηεηληεηληη

24

23

22

21 εεεεε σσσσσ +++=

Astronomical tide σε1 = 0.647 feet

Holland B variations σε2 = 0.15 * surge elevation

Model imprecision σε3 = 0.77 feet

Real storm variability σε4 = 1.17 feet

3131URS Team

Cumulative Distribution FunctionCumulative Distribution Function

1.0

0.99

0.98

8.0 m 7.0 m 6.0 m

3232URS Team

CompletionCompletion

Reconcile/combine with Reconcile/combine with MsCIPMsCIPAdd WHAFIS WavesAdd WHAFIS WavesPrepare DRAFT mapsPrepare DRAFT mapsFinal Digital MapsFinal Digital Maps