Integrated Modeling of Storm Surges during Hurricanes Isabel, Charley, and Frances

UFE0012989.pdf

UFE0012989_pdf.txt

PAGE 4

Mymostprofoundthanksgotomyparents,NadezhdaandVladimir,fortheirloveandendlesssupport.Iwishtoexpressmysincereappreciationtomyadvisorandsupervisorycommitteechairman,Dr.Y.PeterSheng.Iwouldalsoliketothankthemembersofmysupervisorycommittee,Dr.RobertG.Dean,Dr.UlrichH.Kurzweg,Dr.RobertJ.Thieke,andDr.AndrewB.Kennedy.IwouldalsoliketothankYanfengZhang,VladimirParamygin,JeKing,KijinPark,TaeyunKim,JunLee,JustinDavis,DetongSun,DaveChristian,EnriqueGutierrez,andTatianaLomasko. iv

PAGE 5

page ACKNOWLEDGMENTS ............................. iv LISTOFTABLES ................................. ix LISTOFFIGURES ................................ xii ABSTRACT .................................... xxii CHAPTER 1INTRODUCTION .............................. 1 1.1LiteratureReview ............................ 2 1.1.1WaveEectonSurfaceStress ................. 3 1.1.2WaveEectinWave-CurrentInteractionattheBottom ... 5 1.1.3WaveEectthroughRadiationStress ............. 9 1.1.4Miscellaneous .......................... 12 1.2StormSurgeModelReview ...................... 12 1.2.1SLOSH .............................. 14 1.2.2TAOS .............................. 14 1.2.3SPH/WIFM ........................... 15 1.2.4HAZUS .............................. 15 1.2.5ADCIRC ............................. 16 1.2.6SURGE ............................. 17 1.2.7POMcoupledwithWAVEWATCH-IIwavemodel ...... 18 1.2.8CH3D .............................. 19 2THISSTUDY ................................. 20 2.1CH3D-SSMS:WhatMakesitaBetterModel? ............ 20 2.2GoalsandQuestionstobeAnswered ................. 21 2.3ComponentsofCH3D-SSMS ...................... 22 2.3.1Wind ............................... 22 2.3.2RegionalCirculationModel:ADCIRC ............. 29 2.3.3RegionalWaveModel:WAVEWATCH-III .......... 31 2.3.4LocalCirculationModel:CH3D ................ 32 2.3.4.1Governingequations ................. 33 2.3.4.2ImplementationofWetting-and-DryingAlgorithmintoCH3D ...................... 38 2.3.4.3SurfaceandBottomStresses ............. 44 v

PAGE 6

... 45 2.3.4.5RadiationStress ................... 51 2.3.5LocalWaveModel:SWAN ................... 53 3METHODOLOGY .............................. 58 3.1Introduction ............................... 58 3.2CouplingMechanism .......................... 58 4TESTSIMULATIONS ............................ 64 4.1ValidationofWetting-and-DryingSchemeImplementedinCH3D . 64 4.1.1Description ............................ 64 4.1.2Validation ............................ 65 4.1.2.1TestCase1:Wall ................... 65 4.1.2.2TestCase2:Wind .................. 67 4.1.2.3TestCase3:AnalyticalSolution .......... 68 4.2ValidationofAtmosphericPressureGradientTermsImplementedinCH3D ................................. 73 4.2.1Description ............................ 73 4.2.2Validation ............................ 74 4.3ValidationofNear-BottomWave-CurrentInteraction ........ 76 4.3.1Description ............................ 76 4.3.2Validation ............................ 76 4.3.2.1PureOscillatoryFlow ................ 77 4.3.2.2CurrentSuperimposedonanOscillatoryFlow ... 79 4.4ValidationofWaveSetupCalculatedbasedonSWAN-CH3Dcou-pling ................................... 81 4.4.1Description ............................ 81 4.4.2Validation ............................ 82 4.5ValidationofCrossandLongshoreCurrentsBasedonREF/DIF-CH3DCoupling ............................. 84 4.5.1DescriptionofCross-shoreandLongshoreCurrents ..... 84 4.5.2Validation ............................ 86 4.6ValidationofWaveHeightSimulatedbySWANUnderStormCon-ditions .................................. 91 5VALIDATIONOFTHESTORMSURGEMODELINGSYSTEM .... 94 5.1HurricaneIsabel(2003) ......................... 94 5.1.1DescriptionAccordingtoNHC ................. 94 5.1.2ComputationalDomain ..................... 96 5.1.3FieldData ............................ 97 5.1.4ForcingandBoundaryConditions ............... 100 5.1.5Results:SimulatedWave .................... 108 5.1.6Results:SimulatedWaterLevel ................ 114 5.1.7ErrorAnalysisofCalculatedWaterLevel ........... 122 vi

PAGE 7

................ 136 5.1.9Results:SimulatedCurrents .................. 146 5.2HurricaneCharley(2004) ........................ 150 5.2.1DescriptionAccordingtoNHC ................. 150 5.2.2ComputationalDomain ..................... 152 5.2.3Data ............................... 153 5.2.4Results:SimulatedWaterLevel ................ 156 5.2.5ErrorAnalysisofCalculatedWaterLevel ........... 167 5.2.6Results:SimulatedFloodLevel ................ 175 5.3HurricaneFrances(2004) ........................ 183 5.3.1DescriptionAccordingtoNHC ................. 183 5.3.2ComputationalDomain ..................... 185 5.3.3Data ............................... 186 5.3.4Results:SimulatedWaterLevel ................ 188 5.3.5ErrorAnalysisofCalculatedWaterLevel ........... 192 5.3.6Results:SimulatedFloodLevel ................ 199 6FUTUREENHANCEMENTSANDAPPLICATIONS .......... 201 6.1ModelingofMorphologicalImpactsofExtremeStorms ....... 201 6.2RipCurrentForecasting ........................ 201 7CONCLUSIONS ............................... 203 APPENDIX ASAFFIR-SIMPSONHURRICANESCALE ................. 208 BFORMULAETOCALCULATEERRORS ................. 210 CBESTTRACKSFORISABEL,CHARLEY,ANDFRANCES ...... 211 DWINDSPEEDANDDIRECTIONDURINGHURRICANEISABEL:WNAANDWINDGENVS.MEASURED ................. 216 EOUTERBANKS/CHESAPEAKEBAYCOMPUTATIONALGRIDEX-AMPLEPLOT ................................ 228 FHURRICANEISABEL:SIMULATEDRESULTSVS.MEASUREDDATA 230 F.1Simulatedvs.MeasuredWaterLevel ................. 230 F.2Simulatedvs.MeasuredSurge ..................... 238 GHURRICANECHARLEY:SIMULATEDVS.MEASUREDWATERLEVEL .................................... 242 HHURRICANEFRANCES:SIMULATEDRESULTSVS.MEASUREDDATA ..................................... 247 vii

PAGE 8

................. 247 H.2Simulatedvs.MeasuredSurge ..................... 250 ILOW-PASSFILTER ............................. 252 REFERENCES ................................... 254 BIOGRAPHICALSKETCH ............................ 261 viii

PAGE 9

Table page 1{1Asummaryofstormsurgemodels. ..................... 13 2{1Winddatasummary. ............................. 24 2{2Parametersusedtocreatethe\lookuptable". ............... 51 4{1WaveparametersusedtoimposeHurricaneFloyd(1999)boundarycon-ditions. .................................... 92 4{2ComparisonofcalculatedandmeasuredwaveheightduringHurricaneFloyd(1999). ................................. 92 4{3ComparisonofcalculatedandmeasuredwaveheightduringHurricaneFloyd(1999)withwavesetupbeingaccountedfor. ............ 93 4{4WaveparametersusedtoimposeHurricaneBonnie(1998)boundarycon-ditions. .................................... 93 4{5ComparisonofcalculatedandmeasuredwaveheightduringHurricaneBonnie(1998). ................................ 93 5{1MeasuredstormtidecrestsatseveralsitesinNorthCarolina,Virginia,andMaryland. ................................ 96 5{2Tide,windandwavestationsusedforvalidationofthemodelduringHurricaneIsabel. ............................... 99 5{3ADCIRCtidalconstituentsandtheirperiodsusedintheCH3DmodeltosimulateHurricaneIsabel. ......................... 101 5{4TidalconstituentparametersatDuckPier,NCcalculatedbasedonAD-CIRCtidalconstituentsandIOSprogram. ................. 102 5{5TidalconstituentparametersatBeaufort,NCcalculatedbasedonAD-CIRCtidalconstituentsandIOSprogram. ................. 103 5{6ErrorsofWNAandWINDGENwindspeedanddirectioncomparedwithmeasuredatwindstationsduringHurricaneIsabel. ............ 105 5{7Alistofsimulationswithvariouscombinationsofsixmodelfeatures(psymboldenotesthefeaturewasincludedduringthesimulation). ..... 122 ix

PAGE 10

......... 124 5{9MeasuredpeakwaterelevationsatsevenstationsduringHurricaneIs-abelusingWNAwindandvariouscombinationsofstormsurgemodelfeatures. .................................... 130 5{10Calculatedpeakstormsurge(withtidessubtracted)atsevenstationsduringHurricaneIsabelusingWNAwindandvariouscombinationsofstormsurgemodelfeatures. ......................... 131 5{11Alistofsimulationswithvariouscombinationsofsixmodelfeatures(psymboldenotesthefeaturewasincludedduringthesimulation). ..... 167 5{12ErrorsofwaterelevationattidestationsduringHurricaneCharley. ... 169 5{13MeasuredpeakwaterelevationsatfourstationsduringHurricaneCharleyusingWINDGENwindandvariouscombinationsofstormsurgemodelfeatures. .................................... 173 5{14Calculatedpeakstormsurge(withtidessubtracted)atfourstationsdur-ingHurricaneCharleyusingWINDGENwindandvariouscombinationsofstormsurgemodelfeatures. ........................ 173 5{15Comparisonbetweenreportedhighwatermarkvaluesandoodlevelscalculatedusingtwotechniques ....................... 181 5{16Alistofsimulationswithvariouscombinationsofsixmodelfeatures(psymboldenotesthefeaturewasincludedduringthesimulation). ..... 193 5{17ErrorsofwaterelevationattidestationsduringHurricaneFrances. ... 194 5{18MeasuredpeakwaterelevationsatthreestationsduringHurricaneFrancesusingWNAwindandvariouscombinationsofstormsurgemodelfeatures. 198 5{19Calculatedpeakstormsurge(withtidessubtracted)atthreestationsduringHurricaneFrancesusingWNAwindandvariouscombinationsofstormsurgemodelfeatures. ......................... 198 7{1Summaryofsimulatedhurricanes. ...................... 207 C{1BesttrackforHurricaneIsabel,6-19September2003. ........... 212 C{2BesttrackforHurricaneCharley,9-14August2004. ............ 214 C{3BesttrackforHurricaneFrances,31August-7September2004. ..... 215 F{1Alistofsimulationswithvariouscombinationsofsixmodelfeatures(psymboldenotesthefeaturewasincludedduringthesimulation). ..... 230 x

PAGE 11

..... 242 H{1Alistofsimulationswithvariouscombinationsofsixmodelfeatures(psymboldenotesthefeaturewasincludedduringthesimulation). ..... 247 xi

PAGE 12

Figure page 2{1TheADCIRCcomputationalgrid. ...................... 30 2{2TheWAVEWATCH-IIINorthAtlanticregionalcomputationalgrid. ... 32 3{1Adiagramofvariousphysicalprocesses.Thoseinredareaccountedforinthismethodology. ............................. 59 3{2Adiagramofthecouplingprocess. ..................... 63 4{1Thewalltestcase:computationallayout. .................. 66 4{2Thewalltestcase:calculatedwatersurfaceelevation. ........... 66 4{3Thewindtestcase:computationallayout. ................. 67 4{4Thewindtestcase:calculatedwatersurfaceelevation. .......... 68 4{5Tidalcase:comparisonwithanalyticsolutionatt=0. ........... 70 4{6Tidalcase:comparisonwithanalyticsolutionatt=/6. ......... 70 4{7Tidalcase:comparisonwithanalyticsolutionatt=/3. ......... 71 4{8Tidalcase:comparisonwithanalyticsolutionatt=/2. ......... 71 4{9Tidalcase:comparisonwithanalyticsolutionatt=2/3. ......... 72 4{10Tidalcase:comparisonwithanalyticsolutionatt=5/6. ......... 72 4{11Tidalcase:comparisonwithanalyticsolutionatt=. ........... 73 4{12Analyticalsolutionofwatersurfaceelevationduetoatmosphericpres-suregradientforasimpliedhurricane. ................... 75 4{13Dierenceinwaterelevationbetweentheanalyticalandnumericalsolu-tions. ...................................... 76 4{14Comparisonbetweenmeasured( JonssonandCarlsen , 1979 )[dashedlinewithsquares]andcalculated[solidline]velocityprolesforeightphaseangles. ..................................... 78 4{15Verticalproleofthecalculatedphaselagbetweenhorizontalvelocitiesandfreestreamvelocity. ........................... 78 xii

PAGE 13

JonssonandCarlsen ( 1979 )experiment[dashedlinewithsquares]. ............... 79 4{17Comparisonbetweenmeasured( BakkerandDorn , 1978 )[dashedlinewithsquares]andcalculated[solidline]velocityprolesforeightphaseangles. ..................................... 80 4{18Bottomstressduetowave-currentinteractioncalculatedusingthe1-DBBLmodelbasedonthenumericalsimulationofthe BakkerandDorn ( 1978 )laboratoryexperiment. ........................ 81 4{19LayoutofStiveandWindexperimentalsetup(from StiveandWind ( 1982 )). 82 4{20LayoutofMoryandHammexperimentalsetup(from MoryandHamm ( 1997 )). .................................... 84 4{21Comparisonbetweenmeasuredandcalculatedwavesetup( MoryandHamm ( 1997 )experiment). .......................... 85 4{22Calculatedfreesurfaceelevationandcurrentpatternalongwiththelo-cationswhereverticalvelocityprolesweremeasured(lettersAthroughN). ....................................... 87 4{23Simulated(reddashedline)vs.measured(greensolidline)longshoreve-locities:prolesA,B,C,andN. ....................... 88 4{24Simulated(reddashedline)vs.measured(greensolidline)cross-shorevelocities:prolesA,B,C,andN. ..................... 89 4{25Simulated(reddashedline)vs.measured(greensolidline)longshoreve-locities:prolesD,F,I,andH. ....................... 89 4{26Simulated(reddashedline)vs.measured(greensolidline)cross-shorevelocities:prolesD,F,I,andH. ...................... 90 4{27Simulated(reddashedline)vs.measured(greensolidline)longshoreve-locities:prolesEandG. .......................... 90 4{28Simulated(reddashedline)vs.measured(greensolidline)cross-shorevelocities:prolesEandG. ......................... 90 4{29TheFRFinstrumentsetupatDuck,NC .................. 91 5{1BesttrackofHurricaneIsabel(courtesyofNOAANHC). ......... 95 5{2TheOuterBanksandChesapeakeBaygriddomainforIsabelsimulation. 98 5{3LocationofthenineRiverInputMonitoringsites(courtesyofUSGS). . 100 xiii

PAGE 14

................................... 101 5{5WINDGENandWNAvs.measuredwindspeedanddirectionatCapeLookout,NCduringHurricaneIsabel. ................... 104 5{6WINDGENandWNAvs.measuredwindspeedanddirectionatDuckPier,NCduringHurricaneIsabel. ...................... 104 5{7SignicantwaveheightandpeakwaveperiodobtainedfromWAVEWATCH-IIIcomparedwithmeasuredwaveheightatNDBCstation41001. .... 107 5{8SignicantwaveheightandpeakwaveperiodobtainedfromWAVEWATCH-IIIcomparedwithmeasuredwaveheightatNDBCstation41002. .... 107 5{9LocationoftheVIMSinstrumentpackageatGloucesterPoint,VA. ... 108 5{10Simulatedsignicantwaveheightvs.measuredfromtheFRFWaveriderbuoyduringHurricaneIsabel. ........................ 109 5{11Simulatedpeakwaveperiodvs.measuredfromtheFRFWaveriderbuoyduringHurricaneIsabel. ........................... 110 5{12Simulatedwavedirectionvs.measuredfromtheFRFWaveriderbuoyduringHurricaneIsabel. ........................... 111 5{13Atestcase:wavesetupandcurrentsinducedbywavesapproachingtheshorefromsouth-westtonorth-east(toppanel),andfromnorth-westtosouth-east(bottompanel). .......................... 112 5{14Simulatedsignicantwaveheightandpeakwaveperiodvs.measuredfromtheFRFpierduringHurricaneIsabel. ................ 113 5{15Simulatedsignicantwaveheightandpeakwaveperiodvs.measuredatVIMSduringHurricaneIsabel. ....................... 114 5{16Comparisonofsimulatedvs.measuredwaterelevationatBeaufort,NC.Twosimulatedresultsareshown:oneusingWNAwindandtheotherusingWINDGENwind. ........................... 115 5{17Comparisonofsimulatedvs.measuredwaterelevationatDuck,NC.Twosimulatedresultsareshown:oneusingWNAwindandtheotherusingWINDGENwind. ............................... 115 5{18Comparisonofsimulatedvs.measuredwaterelevationatChesapeakeBayBridge,VA.Twosimulatedresultsareshown:oneusingWNAwindandtheotherusingWINDGENwind. ................... 116 xiv

PAGE 15

........................... 116 5{20Comparisonofsimulatedvs.measuredwaterelevationatMoneyPoint,VA.Twosimulatedresultsareshown:oneusingWNAwindandtheotherusingWINDGENwind. ........................... 117 5{21Comparisonofsimulatedvs.measuredwaterelevationatKiptopeke,VA.Twosimulatedresultsareshown:oneusingWNAwindandtheotherusingWINDGENwind. ........................... 117 5{22Comparisonofsimulatedvs.measuredwaterelevationatLewisetta,VA.Twosimulatedresultsareshown:oneusingWNAwindandtheotherusingWINDGENwind. ........................... 118 5{23MaximumwaterelevationrelativetoNAVD88(includestide,surgeandwavesetup)calculatedduringsimulationofHurricaneIsabelintheOuterBanks/ChesapeakeBayusingWNAwind. ................. 120 5{24MaximumwavesetupelevationrelativetoNAVD88calculatedduringsimulationofHurricaneIsabelinthesouthernpartofOuterBanksus-ingWNAwind. ................................ 121 5{25Simulatedstormsurge(waterlevelminustide)atthesevenstationsthrough-outtheOuterBanks/ChesapeakeBayusingWNAwind. ......... 121 5{26Separatelysimulatedtide,wavesetup,andsurge,andtheirlinearsuper-positionatDuck. ............................... 133 5{27Linearlycoupledwaterelevationvs.waterelevationcalculatedthroughdynamiccouplingatDuck,NC. ....................... 133 5{28Linearlycoupledwaterelevationvs.waterelevationcalculatedthroughdynamiccouplingatDuck,NC. ....................... 134 5{29Linearlycoupledwaterelevationvs.waterelevationcalculatedthroughdynamiccouplingneartheSouthRiver,NC.ThelocationisinitiallydryandgetsoodedduringIsabel.Afterthesurgerecedes,itbecomesdryagain. ..................................... 135 5{30Linearlycoupledwaterelevationvs.waterelevationcalculatedthroughdynamiccouplingononeoftheemergentislandsoftheOuterBanks,NC.ThelocationisinitiallydryandgetsoodedduringIsabel.Afterthesurgerecedes,itbecomesdryagain. .................. 135 5{31Linearlycoupledwaterelevationvs.waterelevationcalculatedthroughdynamiccouplingnearGloucester,VA.ThelocationisinitiallydryandgetsoodedduringIsabel.Afterthesurgerecedes,itbecomesdryagain. 136 xv

PAGE 16

........ 137 5{33MaximumsimulatedinundationintheeasternpartoftheOuterBanksduringHurricaneIsabelusingWNAwind(toppanel).Thebottompanelshowsthetimeduringwhichthemaximumoodoccurred. ........ 138 5{34MaximumsimulatedinundationintheChesapeakeBayduringHurri-caneIsabelusingWNAwind(toppanel).Thebottompanelshowsthetimeduringwhichthemaximumoodoccurred. .............. 139 5{35MaximumsimulatedinundationinthesouthernpartoftheOuterBanksduringHurricaneIsabelusingWINDGENwind(toppanel).Thebot-tompanelshowsthetimeduringwhichthemaximumoodoccurred. .. 140 5{36MaximumsimulatedinundationintheeasternpartoftheOuterBanksduringHurricaneIsabelusingWINDGENwind(toppanel).Thebot-tompanelshowsthetimeduringwhichthemaximumoodoccurred. .. 141 5{37MaximumsimulatedinundationintheChesapeakeBayduringHurri-caneIsabelusingWINDGENwind(toppanel).Thebottompanelshowsthetimeduringwhichthemaximumoodoccurred. ........... 142 5{38Pre-storm(top)andpost-storm(middle)airphotostakeninthesouth-ernOuterBanks. ............................... 144 5{39Pre-storm(top)andpost-storm(middle)airphotostakenintheeasternOuterBanks. ................................. 145 5{40LocationofKittyHawk,NCwherecurrentsweremeasured. ....... 146 5{41LocationofGloucesterPoint,VAwherecurrentsweremeasured. ..... 147 5{42Measured(left)andsimulated(right)\SouthtoNorth"currentatKittyHawk,NCduringHurricaneIsabel. ..................... 148 5{43Measured(left)andsimulated(right)\WesttoEast"currentatKittyHawk,NCduringHurricaneIsabel. ..................... 148 5{44Measured(left)andsimulated(right)\SouthtoNorth"currentatGlouces-terPoint,VAduringHurricaneIsabel. ................... 149 5{45Measured(left)andsimulated(right)\WesttoEast"currentatGlouces-terPoint,VAduringHurricaneIsabel. ................... 149 5{46BesttrackofHurricaneCharley(courtesyofNOAANHC). ........ 150 5{47TheCharlotteHarborgriddomain. ..................... 153 xvi

PAGE 17

...................... 154 5{49Measuredwinddirectionvs.WINDGENandWNAwinddataatFtMy-ers,FLduringHurricaneCharley. ...................... 155 5{50Measuredwindspeedvs.WINDGENandWNAwinddataatNaples,FLduringHurricaneCharley. ........................ 155 5{51Measuredwinddirectionvs.WINDGENandWNAwinddataatNaples,FLduringHurricaneCharley. ........................ 156 5{52Comparisonofsimulatedvs.measuredwaterelevationatBigCarlosPass.Twosimulatedresultsareshown:oneusingWNAwindandanotherus-ingWINDGENwind. ............................. 157 5{53Comparisonofsimulatedvs.measuredwaterelevationatEsteroBay,lo-cation1.Twosimulatedresultsareshown:oneusingWNAwindandanotherusingWINDGENwind. ....................... 157 5{54Comparisonofsimulatedvs.measuredwaterelevationatEsteroBay,lo-cation2.Twosimulatedresultsareshown:oneusingWNAwindandanotherusingWINDGENwind. ....................... 158 5{55Comparisonofsimulatedvs.measuredwaterelevationatFtMyers.Twosimulatedresultsareshown:oneusingWNAwindandanotherusingWINDGENwind. ............................... 158 5{56Simulatedvs.measuredwaterelevationatEsteroBay,location1.Dashedlinesspecifythethreetimeinstantswhenwindsnapshotsweretaken. .. 159 5{57WNAwindeldsnapshot1(Aug-1320:55,JulianDay=226.872)alongwithtotaldepthcontoursintheEsteroBayarea. ............. 160 5{58WINDGENwindeldsnapshot1(Aug-1320:55,JulianDay=226.872)alongwithtotaldepthcontoursintheEsteroBayarea. .......... 160 5{59WNAwindeldsnapshot2(Aug-1320:55,JulianDay=226.872)alongwithtotaldepthcontoursintheEsteroBayarea. ............. 161 5{60WINDGENwindeldsnapshot2(Aug-1320:55,JulianDay=226.872)alongwithtotaldepthcontoursintheEsteroBayarea. .......... 161 5{61WNAwindeldsnapshot3(Aug-1401:20,JulianDay=227.055)alongwithtotaldepthcontoursintheEsteroBayarea. ............. 162 5{62WINDGENwindeldsnapshot3(Aug-1401:20,JulianDay=227.055)alongwithtotaldepthcontoursintheEsteroBayarea. .......... 162 xvii

PAGE 18

....... 163 5{64MaximumwaterelevationrelativetoNAVD88(includestide,surgeandwavesetup)calculatedduringsimulationofHurricaneCharleyinChar-lotteHarborusingWINDGENwind. .................... 165 5{65Simulatedstormsurge(waterlevelminustide)atthefourstationsusingWINDGENwind. ............................... 166 5{66MaximumsimulatedinundationinCharlotteHarborusingWINDGENwind(toppanel).Thebottompanelshowsthetimeduringwhichthemaximumoodoccurred. .......................... 176 5{67MaximumsimulatedinundationinCharlotteHarborusingWNAwind(toppanel).Thebottompanelshowsthetimeduringwhichthemaxi-mumoodoccurred. ............................. 177 5{68Pre-storm(top)andpost-storm(middle)airphotostakennearCaptivaIsland.Aclose-upofourcalculatedoodmap(bottom)veriesthepres-enceofwaterovertheland. ......................... 179 5{69Pre-storm(top)andpost-storm(middle)airphotostakennearSanibelIsland. ..................................... 180 5{70NauticalchartofcoastalareasintheCharlotteHarborareaimpactedbyHurricaneCharley. ............................ 181 5{71ManpointsatahighwatermarkleftbystormsurgecausedbyHurri-caneCharleyonNorthCaptivaIsland. ................... 182 5{72BesttrackofHurricaneFrances(courtesyofNOAANHC). ........ 183 5{73TheTampaBaygriddomain. ........................ 187 5{74Comparisonofsimulatedvs.measuredwaterelevationatClearwater,FL.Twosimulatedresultsareshown:oneusingWNAwindandtheotherusingWINDGENwind. ........................... 188 5{75Comparisonofsimulatedvs.measuredwaterelevationatStPete,FL.Twosimulatedresultsareshown:oneusingWNAwindandtheotherusingWINDGENwind. ........................... 189 5{76Comparisonofsimulatedvs.measuredwaterelevationatPortMana-tee,FL.Twosimulatedresultsareshown:oneusingWNAwindandtheotherusingWINDGENwind. ........................ 189 xviii

PAGE 19

............................ 191 5{78Simulatedstormsurge(waterlevelminustide)atthethreestationsus-ingWNAwind. ................................ 192 5{79MaximumsimulatedinundationinTampaBayusingWNAwind(toppanel).Thebottompanelshowsthetimeduringwhichthemaximumoodoccurred. ................................ 200 D{1WINDGENandWNAvs.measuredwindspeedanddirectionatCapeLookout,NCduringHurricaneIsabel. ................... 217 D{2WINDGENandWNAvs.measuredwindspeedanddirectionatDuck,NCduringHurricaneIsabel. ......................... 218 D{3WINDGENandWNAvs.measuredwindspeedanddirectionatChesa-peakeLight,VAduringHurricaneIsabel. .................. 219 D{4WINDGENandWNAvs.measuredwindspeedanddirectionatChesa-peakeBayBridge,VAduringHurricaneIsabel. .............. 220 D{5WINDGENandWNAvs.measuredwindspeedanddirectionatKip-topeke,VAduringHurricaneIsabel. .................... 221 D{6WINDGENandWNAvs.measuredwindspeedanddirectionatMoneyPoint,VAduringHurricaneIsabel. ..................... 222 D{7WINDGENandWNAvs.measuredwindspeedanddirectionatGlouces-terPoint,VAduringHurricaneIsabel. ................... 223 D{8WINDGENandWNAvs.measuredwindspeedanddirectionatLe-wisetta,VAduringHurricaneIsabel. .................... 224 D{9WINDGENandWNAvs.measuredwindspeedanddirectionatHPLWS,VAduringHurricaneIsabel. ......................... 225 D{10WINDGENandWNAvs.measuredwindspeedanddirectionatChop-tankRiver,VAduringHurricaneIsabel. .................. 226 D{11WINDGENandWNAvs.measuredwindspeedanddirectionatNorthBay,VAduringHurricaneIsabel. ...................... 227 E{1ComputationalgridnearChesapeakeBaymouth. ............. 228 E{2ComputationalgridintheSouthOuterBanksarea. ............ 229 F{1Comparisonofsimulatedvs.measuredwaterelevationatBeaufort,NC. 231 xix

PAGE 20

.. 232 F{3Comparisonofsimulatedvs.measuredwaterelevationatChesapeakeBayBridge,VA. ............................... 233 F{4Comparisonofsimulatedvs.measuredwaterelevationatGloucesterPoint,VA. ...................................... 234 F{5Comparisonofsimulatedvs.measuredwaterelevationatMoneyPoint,VA. ...................................... 235 F{6Comparisonofsimulatedvs.measuredwaterelevationatKiptopeke,VA. 236 F{7Comparisonofsimulatedvs.measuredwaterelevationat,Lewisetta,VA. 237 F{8Comparisonofsimulatedvs.measuredstormsurgeelevationatBeau-fort,NC.CalculatedresultsarebasedonSimulation3usingWNAwind. 238 F{9Comparisonofsimulatedvs.measuredstormsurgeelevationatDuck,NC.CalculatedresultsarebasedonSimulation3usingWNAwind. ... 239 F{10Comparisonofsimulatedvs.measuredstormsurgeelevationatChesa-peakeBayBridge,VA.CalculatedresultsarebasedonSimulation3us-ingWNAwind. ................................ 239 F{11Comparisonofsimulatedvs.measuredstormsurgeelevationatGlouces-terPoint,VA.CalculatedresultsarebasedonSimulation3usingWNAwind. ...................................... 240 F{12Comparisonofsimulatedvs.measuredstormsurgeelevationatMoneyPoint,VA.CalculatedresultsarebasedonSimulation3usingWNAwind. 240 F{13Comparisonofsimulatedvs.measuredstormsurgeelevationatKiptopeke,VA.CalculatedresultsarebasedonSimulation3usingWNAwind. ... 241 F{14Comparisonofsimulatedvs.measuredstormsurgeelevationatLewisetta,VA.CalculatedresultsarebasedonSimulation3usingWNAwind. ... 241 G{1Comparisonofsimulatedvs.measuredwaterelevationatBigCarlosPass,FL. ....................................... 243 G{2Comparisonofsimulatedvs.measuredwaterelevationatEsteroBay#1,FL. .................................... 244 G{3Comparisonofsimulatedvs.measuredwaterelevationatEsteroBay#2,FL. .................................... 245 G{4Comparisonofsimulatedvs.measuredwaterelevationatFtMyers,FL. 246 xx

PAGE 21

. 248 H{2Comparisonofsimulated(usingWNAwind)vs.measuredwatereleva-tionatStPete,FL.Calculatedresultsarebasedonvesimulations. .. 249 H{3Comparisonofsimulated(usingWNAwind)vs.measuredwatereleva-tionatPortManatee,FL.Calculatedresultsarebasedonvesimulations. 249 H{4Comparisonofsimulatedvs.measuredstormsurgeelevationatClear-water,FL.CalculatedresultsarebasedonSimulation3usingWNAwind. 250 H{5Comparisonofsimulatedvs.measuredstormsurgeelevationatStPete,FL.CalculatedresultsarebasedonSimulation3usingWNAwind. ... 251 H{6Comparisonofsimulatedvs.measuredstormsurgeelevationatPortMan-atee,FL.CalculatedresultsarebasedonSimulation3usingWNAwind. 251 xxi

PAGE 22

Astormsurgemodelingsystem,CH3D-SSMS,thatcouplesregionalandlocalscalecirculationandwavemodelswasdeveloped.ThemodelcalculatesstormsurgeelevationduringhurricaneeventsusingeithersimpleanalyticwindeldorwindeldsproducedbysophisticatedwindmodelssuchasNCEPWNAandWINDGEN.TheCH3Dmodelisdynamicallycoupledwithawavemodel,SWAN,accountingforwavesetup,waveenhancedsurfacestress,andwaveenhancedbottomfriction.Themodelalsofeaturesarobustoodinganddryingschemethatallowssimulatingofstorminducedinundation.TheCH3Dmodelisalsocoupledwitharegionalscalecirculationmodel,ADCIRC,thatprovidesstormsurgeelevationconditionsalongopenboundaries.ThemodelwasvalidatedbysimulatingHurricanesIsabel,Charley,andFrances. Theeectsofvariousinteractionsamongstormsurge,tide,windandwaveonsurgewereinvestigated.ForIsabelandFrances,WNAwindwasmoreaccuratethanWINDGENwindandproducedmoreaccuratestormsurge.ForCharley,WINDGENwasmoreaccuratethanWNAandproducedmoreaccuratesurge.WhenIsabel,Charley,andFrancesweresimulatedusingtideandwindonly,the xxii

PAGE 23

Thedynamicallycoupledwaterelevationwascomparedwithlinearlysuperim-posedresultsofindependentlysimulatedtide,wavesetup,andsurge.Theeectistwofold:overopenwater,dynamiccouplingproducesslightlymoreaccuratestormsurge,andoverland,theinundationcalculatedthroughdynamiccouplingoccursearlierandismoresignicant. Theeectofexcludingthewetting-and-dryingfeatureduringstormsurgesimulationswasalsoexaminedandfoundsignicant.DuringCharley,whenthefeaturewasdisabled,thecalculatedwaterelevationatitspeakwassignicantlyoverestimated. xxiii

PAGE 24

HurricanesarethemostdevastatinganddamaginghazardsimpactingtheUnitedStates.Today,hurricanedamagecostsbillionsofdollars.Accordingtothe NationalOceanicandAtmosphericAdministration(NOAA) ( 2005 ),duringthelastcentury,23hurricaneshadeachcauseddamageinexcessof$1billiondollars.DamagefromhurricaneAndrew(1992)alonewasestimatedatmorethan$25billiondollarsinSouthFloridaandLouisiana.Industrydatashowthat65%ofinsuredlossesfromnaturalhazardsintheU.S.overthepast50yearsareduetohurricanes.From1990through1999hurricanescaused140deathsand$50billioninpropertydamageintheU.S.Coastalstormsaccountfor71%ofrecentU.S.disasterlossesannuallywitheacheventcostingroughly$500million.In2004,forthersttimeinhistory,fourmajorhurricanes,Charley,Frances,IvanandJeanne,madelandfallinFlorida.The2004hurricaneseasonwillgodownasthemostcostlyseasononrecordintheU.S.( NOAA , 2005 ),with$42billionestimateddamage,deathstotaling59,anddeathsoutsideoftheU.S.atover3,000.Inaddition,Floridalostmanylivesandpartofthe2,170milesofshorelines.AbridgeandsectionsofI-10weredestroyedandtransportationinterruptedformanydays.Withpopulationanddevelopmentcontinuingtoincreasealongcoastalareas,agreaternumberofpeopleandpropertyarevulnerabletohurricanethreat.Hurricanescannotbecontrolledbutthevulnerabilitycanbereducedthroughaccurateforecasting. Themajordamagecausedbyhurricanesisassociatedwithstormsurgesandcoastalooding.Accordingto NOAA ( 1999 ),astormsurgeisalargedomeofwater,80to160kmwide,thatsweepsacrossthecoastlinenearwhereahurricane 1

PAGE 25

makeslandfall.Itcanbemorethan4.5mdeepatitspeak.Thesurgeofhighwatertoppedbywavescanbedevastating.Alongthecoast,stormsurgeisthegreatestthreattolifeandproperty. Notonlycanhurricanesdamagehousesandbuildingsinhighlypopulatedcoastalresidentialandcommercialareasbutalso,withinjustafewhours,theycancausedrasticchangesinthecoastlineasanoutcomeofmorphologicalresponse.Thismayresultinanecologicalimbalanceofestuarinesystems,especiallythosethatareseparatedfromtheoceanbybarrierislands,whichareacommonfeatureofFlorida'scoastline.Inordertoreducecoastalhazardsassociatedwithhurri-canes,itisnecessarytohaveanaccuratepredictionmodelofstormsurgeandcoastalooding,whichisessentialfordevelopingcosteectivestormmitigationandpreparation. Accuratestormsurgesimulationsarealsoessentialforproducingaccurateoodinsuranceratemaps(FIRMs)forcoastalcounties.Floridacoastalcountiesalonecontributemorethan40%ofthetotalinsurancepremiumscollectedbytheNationalFloodInsuranceProgram(NFIP)administeredbytheFederalEmergencyManagementAgency(FEMA). ShengandAlymov ( 2002 )showedthattheFEMAmethodologyonoodinsuranceratesinPinellascounty,Florida( FEMA , 1988 ),whichisbasedonthe1-DWHAFISmodel,overestimatespossibledamagethatmaybecausedbythe100-yearstormevent.Theuseofamorerobuststormsurgemodelwilllikelyresultinsignicantsavingsininsurancepremiums. BodeandHardy ( 1997 )pointedoutthelackof

PAGE 26

robuststormsurgemodelsfortropicalstorms.Duringthelasttenyears,morehurricaneandstormsurgedatahavebeencollectedinFloridaandelsewhere,providingagoodopportunitytodevelopandvalidatenewstormsurgemodels.Manyexistingstormsurgemodelscontainrathersimplephysicseventhoughphysicallymeasurableattributes,suchaswaterlevel,actuallyincludethecombinedeectsofphysicalprocessessuchaswavesandtides.Theytakeintoaccountonlyafewhurricaneparameterssuchaspressuredecit,sizeofthestorm,itstranslationspeed,anddirection. Janssen ( 1991 )theseasurfacestressdependsnotonlyonwindspeedbutroughnessduetowavesaswell.Thetotalstressneartheseasurfaceisthesumoftheturbulentpartandwave-inducedpart, wheretistheturbulentstresswhichaccordingtothemixing-lengthhypothesiscanbeparameterizedasfollows: @z2(1{2) where=0:4isvonKarmanconstant;U(z)isthewindspeedatheightz;aistheairdensity. Janssen ( 1991 )alsointroducedtheeectiveroughnesslength,ze,asopposedtotheroughnesslength,z0,whenwavesareabsent.Theeectiveroughnesslengthisafunctionofthewave-inducedstress. Inderivationofhiswave-inducedstress, Janssen ( 1991 )usedthefollowingwindprole: whereuisthefrictionvelocity.

PAGE 27

Ifequation 1{3 isdierentiated,squaredandcomparedwithequation 1{2 atz=z0,thefollowingrelationshipbetweenz0andzeisobtained: AssumingthatCharnok-likeexpressionz0=u2/gisvalid,thevalueofistunedinsuchawaythatze=u2/gforoldwindsea.Theoldwindseatermmeansthatwavesarenolongerdevelopingundercurrentwindconditionandthewave-inducedstressforsuchseadiminishes,yieldingze!z0.Foryoungwaves(travelingmuchslowerthatthewind)almosttheentiresurfacestressisduetowaves;therefore,w=approachesone. ZhangandLi ( 1996 )appliedthetheoryofJanssenintheircouplingofathird-generationwavemodelandatwo-dimensionalstormsurgemodel.ComparingtheirresultswithmeasureddataoftwostormeventsthattookplaceinthenorthernSouthChinaSea,theyfoundthattheintroductionofawavedependentdraggivesasignicantimprovementovertheuseofthe SmithandBanke ( 1975 )stressrelationwhichunderestimatedthesurgesby10%. whereCD=(0:066jU10j+0:63)103. Mastenbroeketal. ( 1993 )alsostudiedtheeectofawave-dependentdragcoecientonthegenerationofstormsurgesintheNorthSea.Toestimatetheeectsofwavesontheboundarylayer,thetheoryofJanssenwasused.Theresultswerecomparedtomeasureddataforthreestormperiods.Thecalculationswiththewave-dependentdraggaveasignicantimprovement.InturnthecalculationswiththeSmithandBankestressrelationunderestimatedthesurgesby20%. Donelanetal. ( 1993 )investigatedtheaerodynamicroughnessoftheseasurface,z0,usingdatafromLakeOntario,fromtheNorthSeaneartheDutch

PAGE 28

coast,andfromanexposedsiteintheAtlanticOceanothecoastofNovaScotia.Theyfoundthatnormalizedroughnessdependsstronglyonwaveage(Cp/u)whereCpisthephasespeedofthewavesatthespectralpeak.Theirequationforthewaveenhanceddragcoecientis TheauthorsnormalizedroughnessbytheRMSwaveheightandusingthefrictionvelocity,u,ofthewindstressandconcludedthatinbothcasesyoungwaveswererougherthanmaturewaves. Xieetal. ( 2001 )investigatedtheinuenceofsurfacewavesonoceancurrentsinthecoastalwatersbyusingacoupledwave-currentmodelingsystem.Theytookintoaccountthefactthatthewave-inducedwindstressisnotonlyafunctionofwindspeedbutthewave-modieddragcoecientaswell,whichinturnisafunctionofthespectralpeakfrequencyofwaves.Foryoungwavesthespectralpeakfrequencyislarge,andaccordingly,thewave-inducedsurfacestressislarge.However,forfullydevelopedwindwavesthespectralpeakfrequencyissmall,andaccordingly,theeectofwavesonsurfacestressisrelativelysmall.Theauthorsnotethatthewavespectralpeakfrequencyincreasesasthewaterdepthdecreases.Intheirstudythemagnitudeofthepeakspectralfrequencyincreasesfromabout0.6rads1intherelativelydeeposhorewatertoapproximately0.9-1.0rads1intheshallowcoastalwaters.Asaconsequence,underaconstantwindtheeectofwavesonwindstressislargerintheshallowerwaterthaninthedeeperwater. GraberandMadsen ( 1988 ),theshapeofthewavespectruminnite-depth

PAGE 29

watersissignicantlyinuencedbythebottomfriction.Duringastormeventwhenwavesarelargetheareaofsuchaninuenceextendsfaroshore. GrantandMadsen ( 1979 )developedananalyticaltheorytodescribethecombinedmotionofwavesandcurrentsinthevicinityofaroughbottomandtheassociatedboundaryshearstressbyconsideringacombinedwave-currentfrictionfactor.Themagnitudeofthemaximumboundaryshearstressduetocombinedwaveandcurrentis 2fcwj~ubj2(1{7) wherethecombinedfrictionfactorfcwisafunctionofj~uaj=j~ubj;j~uajisthemagnitudeofthesteadycurrentvelocityvectorataheightaabovethebottom;j~ubjisthemaximumnear-bottomorbitalvelocityfromlinearwavetheory; Schoellhamer ( 1993 )pointedoutweaknessesofthe GrantandMadsen ( 1979 )methodologywhichincludetheintroductionofactitiousreferencevelocityatanunknownlevel,theassumptionofthelogarithmiclayerbeingconstantwhichisnotcorrectwhenwavesarepresent. TangandGrimshaw ( 1996 )adaptedthe GrantandMadsen ( 1979 )bottomboundarylayertheorytostudytheeectofincreasedbottomfrictionduetowind-wavecurrentinteractionusinga2-Dshallowwaternumericalmodel.Theyshowedthatthe GrantandMadsen ( 1979 )theorymaybreakdowninveryshal-lowwaterwherethewaveamplitudesbecomelarge.Toavoidtheproblemtheauthorsintroducedanempiricalwave-breakingcriterionintothebottomfrictionformulation: ifaW

PAGE 30

Basedontheirnumerical(hencenotveried)results, TangandGrimshaw ( 1996 )concludedthatalthoughthewind-waveenhancementofthebottomstressissignicantonlyinthenearshorezoneofshallowwater,thereisadramaticreductionintheseasurfaceelevationandthecurrentsinthisregion. SignellandList ( 1997 )studiedtheeectofwave-enhancedbottomfrictiononstorm-drivencirculationinMassachusettsBaybasedonasimpliedformofthe GrantandMadsen ( 1979 )theorydescribedby Signelletal. ( 1990 ).Theyfoundthatthedragcoecientincreasesdramaticallybyafactorof2-6.Themostsignicantdragcoecientenhancementtookplaceintheshallowregionsneartheshoreline.Inresponsetotheincreasedbottomdrag,however,bottomcurrentswerereducedby30%-70%.Sincethebottomstressisproportionaltobottomdragandtothesquareofthebottomvelocity,themeanbottomshearstressincreasedonlyby10%-60%insteadofafactorof2-6. Wangetal. ( 2000 )analyzedseveralimportantmechanismsforstorm-inducedentrainmentofestuarinesedimentsusingeldmeasurements.Theirstudyshowedthatthebottomshearstress,computedusingawave-currentinteractionmodelbased,again,onthe GrantandMadsen ( 1979 )theory,increasedsignicantlyduringepisodicwindevents.Thecurrentsandwavestendedtoenhanceeachothersothattheshearstressesduringthepeaksofstorms,computedfromthewave-currentinteractionmodel,wereapproximatelythreetimeslargerthanusingthetraditionalquadraticlaw.Alargere-suspensioneventwascausedbyafrontalpassagewhenstrongwind-drivencurrentsaugmentedthetidalcurrents.Itwasalsopointedoutinthisstudythatthetimingofstormwaveswithrespecttotidalphasewasacriticalfactor. LiuandDalrymple ( 1978 )proposedasimpleempiricalmodelwhichwasalsousedby SunandSheng ( 2002 )intheirstudyof3-Dwave-inducedcirculation.The

PAGE 31

bottomshearstressisdenedas whereUwbisthemaximumnear-bedwaveorbitalvelocityestimatedfromlinearwavetheory;~ubisthenear-bedwave-averagedvelocity;Cdisthefrictioncoecientwhichcanbecalculatedaccordingtothelawofthewall Sheng ( 1986 ) whereisvonKarmanconstant,z0=ks=30withksbeingtheNikuradseequivalentsandroughness,andzbistheverticaldistanceofthelowergridpointabovethebottom. Thissimplemodelassumesthatwavesandcurrentsareco-linearwhichisratherunrealistic.Asaresultoftheassumption,thebottomstresswillbeoverestimatedwhenwaveandcurrentdirectionsdeviatefromeachother. Xieetal. ( 2001 )alsoemphasizethatsurfacewavesproducetwooppositeeectsoncirculation:energyinputthroughsurfacestressandenergydissipationthroughbottomstress.Theneteectofwave-inducedsurfaceandbottomstressescanbequitedierentunderdierentwinddirections.Thiseectcaneitherenhanceordampthesurfacecurrent.Theauthorsshowedthattheeectofwave-inducedbottomstressismoresignicantforalongshorewindsthanforcross-shorewinds.Theirresultsindicatedthattheeectofwavesoncurrentsismainlypresentinshallowcoastalwatersandattenuatesrapidlyoshoreaswaterdepthincreases. Xieetal. ( 2001 )alsonotethattheeectsofwave-inducedsurfaceandbottomstressesalsodependonthewindspeed.Intropicalcyclonesituationsthewave-inducedsurfaceshearstressisgenerallymoreimportantthanthatdueto

PAGE 32

wave-inducedbottomstress,andhencetheeectofwindwavesusuallyincreasesthemagnitudeofstormsurge. JohnsonandKofoed-Hansen ( 2000 )studiedtheinuenceofbottomfrictiononseasurfaceroughness.Theirinvestigationsshowthatthebottomdissipationkeepswavesyoung,whichresultsinincreasedwindfriction. Wavesetupgenerallyoccursinthesurfzone.Thebreakingwavesproduceex-cessmomentumuxintheshorewarddirectionwhichisusuallytermed\radiationstress."Asthebrokenwavescontinuetopropagatetowardtheshore,theexcessmomentumuxorradiationstressdiminishes.Inthesteadystate,theshorewarddecreaseinradiationstressisbalancedbyashorewardincreaseinthewaterlevel.Thisraisesthewatersurfaceelevationwithinthesurfzonetohigherthanthestillwaterlevel(SWL)producingsetup.ItalsopushesthewaterleveloutsideofthesurfzonetolowerthantheSWLproducingsetdown. Itisnecessarytoaccountforwavesetupwithinthesurfzoneduringthecalculationofstormsurgeelevation.Thereasonforthatisthatthesetupmaybeverysignicantespeciallyduringastormevent.Accordingto DeanandDalrymple ( 1991 )thewavesetupneartheshoreisabout19%ofthebreakingwaveheight.Thesignicantwaveheightduringthe100-yearstormestimatedby FederalEmergencyManagementAgency(FEMA) ( 1988 )intheGulfofMexicois8.86mwiththeperiodof11.5sec.Thebreakingwaveheightduringthestormaccordingto DeanandDalrymple ( 1991 ,p.116)mayreach8.5massumingaplanebeach,

PAGE 33

and0odeepwaterincidentangle.Therefore,thewavesetupmayreach1.62mattheshore,whichisasignicantvalueasfarasoodingcausedbythewavesetupisconcerned. Itwouldbeincorrecttoaddthecalculatedwavesetupontopofthecalculatedstormsurgeelevationlinearlysincethereisanon-linearinteractionbetweenthetwo.Moreover,itispracticallyimpossibletodistinctlyseparatethewavesetupfromeverythingelsebasedonthesurfaceelevationdatacollectedintheeld.Therefore,wavesetupshouldbeintroducedintheequationsintermsofradiationstresses,sothatthecalculatedsurfaceelevationwouldincludeitinternallyinanon-linearfashion. Longuet-HigginsandStewart ( 1964 )deriveddepth-integratedradiationstressquantitiesthatareusedinmanynumericalmodelsincludingstormsurgemodelsaccountingforwave-inducedsetup. Mastenbroeketal. ( 1993 )and ZhangandLi ( 1996 )incorporatedradiationstresstermsbasedonthe Longuet-HigginsandStewart ( 1964 )formulationintotwo-dimensionaloceancirculationequationsandinvestigatedtheimportanceofradiationstressincalculationofstormsurge. Mastenbroeketal. ( 1993 )reportedthatonlyinoneofthethreecasestheystudiedtheradiationstressincreasedthesurgesome5%.Intheothertwocasestheeectoftheradiationstresswasinsignicant.Similarly, ZhangandLi ( 1996 )concludedthattheinclusionoftheradiationstressimprovestheaccuracyofthecomputedresultsslightlyby2%. Ithastobepointedoutthatthestudyby Mastenbroeketal. ( 1993 )tookplaceintheNorthSeawithstationsalongtheDutchandBritishcoasts.Theauthorsdonotspecifyhowdeepthelocationsoftheirstationsare.Itishardtoestimatetheimportanceoftheradiationstressifitisnotbeingestimatedinrelativelyshallowwaterwherewavesetupisformedundertheinuenceofthebreakingwaves.Also,thesurgemodeltheyuseddoesnotaccountforooding

PAGE 34

anddrying.Thismakesthemodelinadequateinshallowwaterregions.Thesameconclusionisvalidforthemodelusedby ZhangandLi ( 1996 ).TheirgridislocatedinthenorthernSouthChinaSeaandthedepthsoftheirstationlocationsarenotspeciedeither. ShengandAlymov ( 2002 )implementedthe Longuet-HigginsandStewart ( 1964 )radiationstressesintheCH3D( Sheng , 1987 )modelandsimulated2-Dwavesetupeldsduringthe100-yearstormeventfortwostudyareasinPinellasCounty,Florida.Thesetupvaluevariedfromapproximately0.5mto1.0m.Therewassomeeectonwavesetupwhenwaveswereapproachingat40oanglewhich,accordingto Deanetal. ( 1995 ),isthemostlikelyangleofapproachofthe100-yearstormeventinPinellasCounty.Also,thestudyshowedthatthegridresolutionhassomeeectoncalculatedwavesetupespeciallyintheareaswherebathymetryhassteepergradients.Coarsegridsarenotcapableofresolvingthesebathymetricgradientsandasaresultthecalculatedwavesetupistypicallylowerthanthatcalculatedusinganergrid. SunandSheng ( 2002 )coupledCH3DwiththeREF/DIFwavemodel( KirbyandDalrymple ( 1994 ))andshowedsignicanteectsofwavesonwaterlevelandcoastalcurrents.Theycomparedsimulatedwavesetuptomeasuredlaboratorydataandfoundthatthecalculatedwavesetupisusuallyoverestimatedandproposedthattheoriginalwaveforcingshouldbereducedbymultiplyingwaveforcingbyacoecientof0.8.ThisisratheranadhocapproachandthehighradiationstressesmighthavecomefromoverestimatedwavescalculatedbyREF/DIF. Recently,someeortshavebeenmadeinordertoderiveverticallyvaryingradiationstress. Mellor ( 2003 )exploitedthree-dimensionalequationsofmotiondecomposingvelocitiesintothreecomponents:meancurrent,wave,andturbulence.

PAGE 35

Hisradiationstresstermsareverticallydependantand,ifdepthintegrated,appearinamoreconventionalformasin Longuet-HigginsandStewart ( 1964 ). ( 2002 )foundthatthemaximumsurgeheightsdependonthetidalphasewhenthehurricanelandfalloccurs:maximumsurgeheightoccurswhenhurricanelandfalloccursatseveralhoursafterthepeaktide. ShengandAlymov ( 2002 )simulatedthestormsurgeinPinellasCountyusingthePEMmodel( DavisandSheng , 2002 )andfoundthatthemaximumsurgeheightsandoodingpatternsforthecountyaredramaticallydierentwhenthehighresolutionALSMdataareusedasopposedtotheUSGS0.25-degreedata. ShengandAlymov ( 2002 )usedthe2-DversionofCH3DandREF/DIFtosimulatethewavesetup,andusedtheSWAN( Holthuijsenetal. , 2000 )tosimulatethewind-inducedsurgeinPinellasCounty.However,thecalculationsofthesurge,setup,andwave-inducedsurgewereperformedseparatelyandaddedlinearlyforsimplicity.

PAGE 36

Table1{1: Asummaryofstormsurgemodels. AnalyAssimiBoundary WaveWaveWave Rain/ River Tide ValiOperaLocal/ 2-D 3-D ting tical lated Fitted enhanced induced Setup EvapoDisdation tional Regional Model and Wind Wind Grid Surface Bottom ration charge Grid Drying Stress Friction Nesting p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p1 p p p p p p p p p p p2 p p p p p p p p p p p p3 p ( 2004 )-HurricaneGeorges2 ( 2000 )-TyphoonWinnie3Thisstudy-HurricanesIsabel,Charley,andFrances

PAGE 37

Jelesnianskietal. , 1992 ).ThemodelisrunbytheNationalHurricaneCentertoestimatestormsurgeheightsandwindsresultingfromhistorical,hypothetical,orpredictedhurricanesbytakingintoaccountpressuredecit,size,forwardspeed,track,andwinds.Themodeldoesnottakeintoaccounttide,precipitation/evaporation,riverow,wind-drivenwaves. SLOSHisusedbyNOAANWSandtheU.S.ArmyCorpsofEngineerstocreateoodmapsrepresentingtheMaximumoftheMaximum(MOM)stormsurgecompositeofhypotheticalstorms. Comments:SLOSHisanoutdatedmodelwhichneedstoberevisedandimproved.Ittendstoproducelargeuncertaintyinthepredictedoodedareabecauseofitsrelativelycoarseresolution(0.5-7km)andinabilitytotconvolutedshorelines. Watson , 1995 ; WatsonandJohnson , 1999 ).TAOSissimilartoSLOSHandiscapableofcalculatinganestimateofstormsurge,waveheight,maximumwinds,inlandooding,debrisandstructuraldamage.Themodelhasbeenrunonover25signicanthistoricalstormsfromaroundtheworld.Comparingover500peaksurgeobservationswithTAOSmodelestimatesforthesamelocationandtime,themodelgeneratesresultswithin0.3m80%ofthetime,andlessthan0.6m90%ofthetime. Comments:Asfarasthestormsurgepartisconcerned,theTAOSmodeldoesnotdiermuchfromtheSLOSHmodel,sincethestormsurgephysicsarerepresentedinasimilarwayinbothmodelsand,therefore,itdoesnottakeinto

PAGE 38

accountimportantphysicalprocessesassociatedwithstormsurge,e.g.,wave-inducedeects. Schwerdtetal. , 1979 ; Cialone , 1991 )arecomponentsofCoastalEngineeringResearchCenter'sCoastalModelingSystemusedbytheU.S.ArmyCorpsofEngineers. TheSPHisatwo-dimensional,parametricmodeldevelopedinastretchedCartesiancoordinatesystemforrepresentingwindandatmosphericpressureeldsgeneratedbyhurricanes.ItisbasedontheStandardProjectHurricanecriteriadevelopedbyNOAA,andthemodel'sprimaryoutputsareresultingwindvelocityandatmosphericpressureeldswhichcanbeusedinstormsurgemodeling.TheSPHmodelcanberunindependently,oritcanbeinvokedfromwithinmodelWIFM. TheWIFMisatwo-dimensional,time-dependent,long-wavemodelforsolvingtheverticallyintegratedNavier-StokesequationsinastretchedCartesiancoordinatesystem.Themodelsimulatesshallow-water,long-wavehydrodynamicssuchastidalcirculationand,makinguseofwindeldsproducedbySPH,stormsurges.WIFMcontainsmanyfeaturessuchasmovingboundariestosimulatewettinganddryingoflow-lyingareasandsubgridowboundariestosimulatesmallbarrierislands,jetties,dunes,orotherstructuralfeatures.Modeloutputincludesverticallyintegratedwatervelocitiesandwatersurfaceelevations. Comments:WIFMisasimpleandoutdatedmodelwhichdoesnotaccountforwaveeect.

PAGE 39

Ithasthecapabilitytoestimateearthquakelosses,andoodandwindmodelsarebeingdeveloped. TheHurricaneLossEstimationModelwhichisapartoftheHAZUSmodelincorporatesseasurfacetemperatureintheboundarylayeranalysis,andcalculateswindspeedasafunctionofcentralpressure,translationspeed,andsurfaceroughness.Themodeladdresseswindpressure,windbornedebris,surge,waves,atmosphericpressurechange,duration/fatigue,andrain. TheFloodLossEstimationModeliscapableofassessingriverineandcoastalooding.Itestimatespotentialdamagestoallclassesofbuildings,essentialfacilities,transportationandutilitylifelines,andagriculturalareas.Themodelestimatesdebris,shelterandcasualties.Directlossesareestimatedbasedonphysicaldamagetostructure,contents,andbuildinginteriors.Theeectsofoodwarningandvelocityaretakenintoaccount. Theoodmodelusesgeographicinformationsystemsoftwaretomapanddisplayoodhazarddata,andtheresultsofdamageandlossestimatesforbuildingandinfrastructure.Italsoenablesuserstoestimatetheeectsofoodingonpopulations. Luet-tichetal. , 1992 )solvestheequationsofmotionforauidonarotatingearth.TheseequationsarebasedonhydrostaticpressureandBoussinesqapproximationsandhavebeendiscretizedinspaceusingtheniteelementmethodandintimeusingthenitedierencemethod. ADCIRCcanberuneitherasatwo-dimensionaldepthintegrated(2DDI)modelorasathree-dimensional(3D)model.Ineithercase,elevationisobtainedfromthesolutionofthedepth-integratedcontinuityequationinGeneralizedWave-ContinuityEquationform.Velocityisobtainedfromthesolutionofeitherthe

PAGE 40

2DDIor3Dmomentumequations.Allnonlineartermshavebeenretainedintheseequations.ADCIRCcanberunusingeitheraCartesianorasphericalcoordinatesystem. ADCIRCboundaryconditionsinclude:speciedelevation(harmonictidalconstituentsortimeseries),speciednormalow(harmonictidalconstituentsortimeseries),zeronormalowslipornoslipconditionsforvelocity,externalbarrieroverowoutofthedomain,internalbarrieroverowbetweensectionsofthedomain,surfacestress(windand/orwaveradiationstress),atmosphericpressure,outwardradiationofwaves(Sommereldcondition).ADCIRCcanbeforcedwith:elevationboundaryconditions,normalowboundaryconditions,surfacestressboundaryconditions,tidalpotential,andearthload/selfattractiontide. Comments:ADCIRCisawidelyusedmodel.Sincethemodelisbasedonniteelementnumerics,ithastheabilitytoexploitverylargecomputationaldomainswithsparseresolutionindeepwaterareasandsmallgridspacinginshallowwaterareasornearcomplexboundaries. Weaver ( 2004 )implementedaone-waycouplingofa2-DversionofADCIRCwithawavemodel,WAM-3G,toaccountforradiationstress;nootherwaveeectwasconsidered.HeperformedahindcastofthestormsurgeduringHurricaneGeorges(1998)intheNorthGulfofMexicoandconcludedthattheadditionofwaveforcingimprovedtheoverallpredictivecapabilitiesandreducedtheRMSerrorofthecalculatedstormsurgeby20%to50%. BlumbergandMellor ( 1987 ).SURGEsimulatesandpredictsstormsurge,ooding,overwash,waterrecession,andassociatedhorizontalcurrents.ThemodelmakesuseofNOAA/NOS

PAGE 41

bathymetrydataandhightresolutionUSGS/NOAALIDARsurveydata.Hurri-caneAndrew(1992)andHurricaneCarla(1967)wereusedformodelvericationinLouisianaandLavacaBay,TX,respectively. Comments:TheprosoftheSURGEmodelincludeitsthree-dimensionality,theabilitytousedneresolutioncomputationalgrids,thecapabilitytosimulatewetting-and-dryingofthecoastalarea.Themajordeciencyistheabsenceofwaveeects,e.g.,radiationstress,wave-enhancedsurfacestress,andwave-inducedbottomfriction. Moon , 2000 , 2005 ).Analyticwindmodel( Holland , 1980 )isusedtocalculatedhurricanewindeld.ThemodelwasappliedtonumericalexperimentsintheYellowandEastChinaSeasduringTyphoonWinnie(1997). Theoceancirculationmodelcalculatescurrentsandsurfaceelevation(newwa-terdepth)whichisfedbackintothewavemodeltocomputethewavedependentdragcoecienttobeusedinthewavemodelthenexttimestep.Thisprocessisrepeated.Eachmodelhasitsowntimestepduetothereasonthattimescalesofchangeofwaveparametersandtidalcurrentsarequitedierent.Thewavemodelhasa360sectimestepandtheoceanmodelhasa1800sectimestep.Therefore,afterevery5timestepsofrunningthewavemodeltheoceanmodelisrunandthecouplingtakesplace. Comments:ThePOM/WAVEWATCH-IIcouplingwasperhapstherstat-tempttoexploitatwo-waycouplingbetweenanoceancirculationmodelandawavemodel.Thecouplingtakesintoaccounttheeectsofunsteadyandinhomo-geneouscurrents,unsteadydepth,tides,wind,surfaceheatux,riverdischarge.

PAGE 42

Someofthedecienciesincludetheinapplicabilityofthemodelinshallowwa-terregions;bottomfrictiondependsonlyoncurrents,i.e.noeectofwavesisconsidered;wetting-and-dryingisnotconsidered;wavesetupisnottakenintoaccount. Sheng ( 1986 , 1990 ).Themodelcanbeusedtosimulatetheestuarine,coastal,andriverinecirculationdrivenbywind,tide,anddensitygradients.Themodelusesaboundaryttedcurvilineargridinthehorizontaldirectionstoresolvethecomplexshorelineandgeometry,andaterrain-following-gridintheverticaldirection.ThemodelusesaSmagorinskitypehorizontalturbulentdiusioncoecient,andarobustturbulenceclosuremodel( ShengandVillaret , 1989 )fortheverticalturbulentmixing. CH3Dhasbeenappliedtosimulatethe2-Dand3-DcirculationinnumerouswaterbodiesinFlorida(e.g.,TampaBay,SarasotaBay,IndianRiverLagoon,FloridaBay,BiscayneBay,St.JohnsRiver,andLakeOkeechobee)andU.S.(e.g.,ChesapeakeBay).ManyoftheCH3Dapplications,aswellastheCH3Dformulationanddevelopment,aredescribedon In2002,CH3Dwasmodiedtoincludewetting-and-dryingcapability( Shengetal. , 2002 )andcoupledwithawavemodelSWANtosimulatetheoodelevationinPinellasCountyduringthe100-yearstorm( ShengandAlymov , 2002 ).Thewetting-and-dryingversionofCH3DwillbethefoundationoftheCH3D-SSMSforthisstudy.

PAGE 43

ThischapterprovidesadetaileddescriptionoftheCH3D-SSMSintegratedstormsurgemodelingsystemincludingeachofthefourmodelsitisbasedon:tworegionalmodels,ADCIRC(circulation)andWAVEWATCH-III(wave);andtwolocalmodels,CH3D(circulation)andSWAN(wave). Shengetal. ( 2004 ).Themodelingsystemincludessurge-wave-tide-windcouplinginthecoastal-estuarine-overlandregion,aswellascouplingbetweenlocalandregionalscales. ThetableshownatthebeginningofSection 1.2 demonstratesthattheCH3D-SSMSmodelingsystemhasmorefeaturesthananyotherexistingstormsurgemodel.SuchanimportantelementaswaveswhichisincludedinCH3Dthroughcouplingwithawavemodel,SWAN,isunjustlyneglectedbymostoftheothermodels.Dynamiccouplingwithtideisalsogenerallyignoredassumingthatpredictedtidecanbelinearlyaddedontopofthecalculatedstormsurge.Local/Regionalcouplingissomethingthatisbeingexploitedbymanylately.Theusefulnessofthisfeatureistobeabletopredictandforecaststormsurgelocallyusingnegrids,whichincludehighresolutionshorelines,bathymetryandtopography.Theboundaryconditionsforthelocalmodelareprovidedbymeansofnestedcouplingwiththeregionalmodel.DetailsofthemethodologyusedinthisstudycanbefoundinChapter 3 . 20

PAGE 44

G1)Produceanadvancedstormsurgemodelwithrobustphysicsbyincorpo-ratingthenonlinearinteractionbetweensurge,tide,wave,andwindandallowingtheuseofaverynespatialresolution.Themodelwillbecapableofperforminginshallowwaterregionsandsimulatingwetting-and-drying. G2)Producenelyresolvedboundary-ttedcurvilineargridsfortheOuterBanks/ChesapeakeBay,TampaBay,andCharlotteHarborareasbyutilizinghigh-resolutionbathymetryandtopographydata. G3)ValidatethemodelingsystembysimulatingHurricanesIsabel(2003),Frances(2004),andCharley(2004)andcomparingthecalculatedresultswithmeasureddata. G4)ProduceoodmapsbasedonsimulationsofIsabel,Frances,andCharley. G5)Performasensitivityanalysisoftheeectofnonlinearinteractionsamongstormsurge,tide,wind,andwave,aswellastheeectofwetting-and-dryingandtheeectofthedynamiccouplingversusalinearsuperpositionofseparatelysimulatedtide,wavesetup,andsurge. Q1)Howsignicantistheeectofthenonlinearinteractionbetweentheturbulentandwave-inducedstresses? Q2)Howsignicantistheeectofthenonlinearinteractionbetweenbottomstressesduetocurrentsandwaves? Q3)Howdoeswetting-and-dryingaectstormsurgesimulations? Q4)Isdynamiccouplingbetterthanlinearsuperposition?

PAGE 45

2.3.1Wind Severaltypesofwinddataareusedinthisstudy.Thersttypeisassociatedwiththeactualwindmeasuredfrombuoysintheopenoceanorwindtowersonland.TheNationalDataBuoyCenter(NDBC)isanagencywithintheNationalWeatherService(NWS)oftheNationalOceanicandAtmosphericAdministration(NOAA),whichoperatesandmaintainsanetworkofdatacollectingbuoysandcoastalstationsalongtheU.S.coastline( AnothernetworkofstationsdeployedintheGulfofMexicoistheCoastalOceanMonitoringandPredictionSystem(COMPS).COMPS(

PAGE 46

takemeteorologicalmeasurementssuchaswind,airtemperature,humidity,baro-metricpressure,precipitation,radiation,visibility;andmarinemeasurementssuchaswaterlevel,watertemperature,salinity,currentvelocity,andwaveparameters. Theabovetwotypesofwinddatafromeldmeasurementsareusefulforvalidatingwindmodels,butcannotbeusedalonetogeneratethewindeldneededforstormsurgemodeling. Anothertypeofwinddataiswindsnapshotdatawhichareproducedbyvariouswindmodels,rangingfromsimpletohighlysophisticated.Thesedatacoverlargeareasandaremoresuitableforstormsurgemodeling.Thereareafewdierentwindsnapshotdatasets.ThesummaryofthistypeofwinddataispresentedinTable 2{1 .

PAGE 47

Table2{1: Winddatasummary. Source Type Vert. Cycles Cycle Mean Analysis/ AssimiWind Data (BGD/ ResoluLevel Length/ Sea Forecast/ lated Over Set HUR/ tion Snapshot Level Measured Land CMB) Frequency Pressure NCEP BGD 12km 10m 00,06, 84hrs yes FCAST no yes 12,18 6hrs NDAS BGD 12km 10m 00,06, 6hrs yes ANL yes yes 12,18 6hrs GFDL NCEP HUR varies 35m 00,06, 126hrs yes FCAST no yes 12,18 6hrs GDAS HUR varies 35m 00,06, 6hrs yes ANL yes yes 12,18 6hrs HRD NOAA HUR 6km 10m varies no MEAS no no WINDGEN Ocean CMB 22km 10m 00,06, ?hrs yes ANL+FCAST yes yes weather 12,18 1hr WNA NCEP CMB 28km 10m 00,06, 120hrs no ANL+FCAST yes no 12,18 3hrs PBL Ocean HUR any 20m yes no yes weather Analytical Holland ( 1980 ) HUR any sfc yes no yes WRF NCEP BGD 4km 10m 00,06, 36/84hrs yes ANL+FCAST yes yes 12,18 3hrs MM5 NCEP BGD 12km 10m 00,06, 48hrs yes ANL+FCAST yes yes 12,18 3hrs

PAGE 48

WNAandWINDGENwindswereextensivelyutilizedinthisstudy.Twootherwindmodelsthatweretested:acomplexPlanetaryBoundaryLayer(PBL)modelandamuchsimpleranalyticwindmodel( Holland , 1980 )basedonthehypothesisofanexponentialdecayofatmosphericpressurefromthecenterofastorm. ThePBLmodelisbasedon Chow ( 1971 )vortexmodel.ThemodeliscapableofcalculatingverticallyaveragedthroughthedepthofthePBLvelocitiesduringastormevent. Themodel'sgoverningequationofhorizontalmotionwrittenincoordinatesxedtotheearthis( Cardoneetal. , 1992 ) dt+f~K~V~Vg=1 where~V=~Vave~Vcisthehorizontalwindrelativetothecenterofthecyclone; Theinteractionbetweentheboundarylayerandthefreeatmosphereisexpressedintermsofthegeostrophicwindeld(verticalvelocityatthetopoftheboundarylayer)andthesurfacestress(frictionaldissipationofthekineticenergy

PAGE 49

intheboundarylayer).Furtherparameterizationofthemodelincludesverticaluxesofmomentum,heatandmoisture. Thisparameterizationisbasedonmatchingofmeanprolesofwind,tempera-ture,andmoisturebysurfaceandouterlayersimilaritytheories. Thegeneralformoftheparametricrelationsmaybewrittenas u=(ln[z0/h]+Am)v u=Bmsign(f)(V0)/=(ln[z0/h]0+Cm)(qq0)/q=(ln[z0/h]+Dm)(2{2) whereuandvaretheverticallyaveragedhorizontalvelocitycomponents(inthedirectionofthesurfaceshearandperpendiculartoit,respectively);z0istheroughnessparameter;isvonKarman'sconstant;Vandqarethemeanlayervirtualpotentialtemperatureandspecichumidity,respectively(thesubscript0denotesthevalueatz0);isapotentialtemperaturescaleexpressedintermsoftheheatux,H;qisaspecichumidityscaleinvolvingthemoistureux;andAm,Bm,Cm,andDmareuniversalfunctionsofdimensionlesssimilarityparameters. Arya ( 1977 )presentedthefollowingexpressionsforthesimilarityfunctionsinwhichthedepthofthePBL,h,isspeciedasanindependentvariable.Ifh L2(unstable)then ue0:2h/LCm=ln(h/L)+3:7(2{3)

PAGE 50

andifh L+2(stable)then whereL=u3VCP Fornear-neutralconditions,2
PAGE 51

wherePaistherelativeatmosphericpressureand4P0=P0P1isthecentralpressuredropofthestorm. Thecyclostrophicwindvelocity,Uc,is reR/r(2{7) Thegeostrophicwindvelocity,Ug,is r2eR/r(2{8) Thegradientwindvelocity,UG,is where 2Vs andtheresolvedpart,Vs,ofthetranslationalvelocityofthestorm,Vsis whereistheanglefromthedirectionofbearingofthestorm,,toanypointinsidethestorm.Thesurfacewindvelocity,Us,inthexandydirectionsisthenwrittenas

PAGE 52

whereisaninwardrotationangleof18oandKistheratioofsurfacewindvelocitytogradientwindvelocity. ADCIRCcanberuneitherasatwo-dimensionaldepthintegrated(2DDI)modelorasathree-dimensional(3D)model.Ineithercase,elevationisobtainedfromthesolutionofthedepth-integratedcontinuityequationinGeneralizedWave-ContinuityEquationform.Velocityisobtainedfromthesolutionofeitherthe2DDIor3Dmomentumequations.Allnonlineartermshavebeenretainedintheseequations.ADCIRCcanberunusingeitheraCartesianorasphericalcoordinatesystem. ADCIRCboundaryconditionsinclude:speciedelevation(harmonictidalconstituentsortimeseries),speciednormalow(harmonictidalconstituentsortimeseries),zeronormalowslipornoslipconditionsforvelocity,externalbarrieroverowoutofthedomain,internalbarrieroverowbetweensectionsofthedomain,surfacestress(windand/orwaveradiationstress),atmosphericpressure,outwardradiationofwaves(Sommereldcondition).ADCIRCcanbeforcedwith:elevationboundaryconditions,normalowboundaryconditions,surfacestressboundaryconditions,tidalpotential,andearthload/selfattractiontide. TheadvantageofusingADCIRCisthatitscomputationalgrid(showninFigure 2{1 )coversthewesternpartoftheNorthAtlanticincludingtheGulfofMexicoandtheCaribbeanandconsistsofonly58369elementsand31435nodeswithvaryinggridspacingwhichiscoarseoshore(upto100km)butnernearthecoast(5-6kmintheChesapeakeBayandTampaBayareas).Thismakesthe

PAGE 53

modelcomputationallyecientwithoutmuchofaslowdownwhencoupledwiththelocalcirculationmodel,CH3D.Bothmodelscanusethesametimestep. SomeofthedecienciesthatADCIRChasincludeitsinabilitytoexploitboundaryttedgrids.Nearland,shorelineapproximationcanbereachedbyincreasingtheresolutionofthecomputationalgrid.TheADCIRCcomputationalgridthatwasprovidedtoushadrathercoarseresolution(5-6km)alongthecoastlinewhichwouldresultinlosingsomeaccuracyofthecalculatedstormsurgeinthenearshoreregions.Also,theversionofADCIRCwhichwasusedinthisstudydoesnotcalculatewetting-and-drying.Overall,ADCIRCisarobustcirculationmodelwhichcanbeusefulinestimatingtheresponseoftheoceantoamovinghurricaneonalargeregionalscaleandprovidingboundaryconditionstolocalcirculationmodels. Figure2{1: TheADCIRCcomputationalgrid.

PAGE 54

Tolman ( 1997 )and Tolman ( 1999 ))isathirdgenerationNOAA/NCEPoperationalwavemodel.Itsolvesthespectralactiondensitybalanceequationforwavenumber-directionspectra.Theimplicitassumptionoftheseequationsisthatthemedium(depthandcurrent)aswellasthewaveeldvaryontimeandspacescalesthataremuchlargerthanthecorrespondingscalesofasinglewave.Furthermore,thephysicsincludedinthemodeldonotcoverconditionswherethewavesareseverelydepth-limited.Thisimpliesthatthemodelcangenerallybyappliedonspatialscales(gridincrements)largerthan1to10km,andoutsidethesurfzone. Thegoverningequationsincluderefractionandstrainingofthewaveeldduetotemporalandspatialvariationsofthemeanwaterdepthandthemeancurrent(tides,surgesetc.),andwavegrowthanddecayduetotheactionsofwind,nonlinearresonantinteractions,dissipation(`whitecapping')andbottomfriction.Wavepropagationisconsideredtobelinear.Relevantnonlineareectssuchasresonantinteractionsarethereforeincludedinthesourceterms(physics). TheWAVEWATCH-IIINorthAtlanticcomputationalgrid(showninFigure 2{2 )coversthewesternNorthAtlanticincludingtheGulfofMexicoandCaribbean.Thesizeofthegridis275203withspatialresolutionof0.25degrees(28km). WAVEWATCH-IIIproductsare6-hourhindcastand120-hourforecastwith6-hourintervals.RegionalanalysiswavedataforHurricanesIsabel(2003),Charley(2004),andFrances(2004)wereprovidedbycourtesyofNOAA/NCEP. AswillbeexplainedinChapter 3 ,wedonotactuallyrunWAVEWATCH-IIIwhichprovidesboundaryconditionsforthelocalwavemodel,SWAN.WAVEWATCH-IIIwaveforcingwasobtainedbythecourtesyofNOAA/NCEP.Thisisoneoftheadvantagesofusingthemodel:nocomputationalburdenisinvolvedatallexceptforthepreprocessingphase.Anotherreasonforusing

PAGE 55

WAVEWATCH-IIIisthat,aswillbeshowninSection 5.1.4 ,themodelproducesgoodresultscomparedwithmeasureddataduringhurricaneevents. Figure2{2: TheWAVEWATCH-IIINorthAtlanticregionalcomputationalgrid. Sheng , 1987 , 1990 )isa3-Dcurvilinear-gridhydrodynamicmodel.Themodelsolvesthecontinuityequationandtwomomentumequationsinanon-orthogonalboundaryttedcoordinatesystem.TheequationsarederivedfromtheNavier-Stokesequationsusingfoursimplifyingapproximations.First,itisassumedthatwaterisincompressible,whichresultsinasimpliedcontinuityequation.Second,basedonthefactthatcharacteristicverticallengthscaleis

PAGE 56

muchsmallerthanthehorizontallengthscale,theverticalvelocityissmallandverticalaccelerationmaybeneglected.Thus,theverticalmomentumequationcanbereducedtothehydrostaticpressurerelation.Third,withtheBoussinesqapproximation,anaveragedensitycanbeusedintheequationsexceptinthebuoyancyterm.Finally,theeddy-viscosityconcept,whichassumesthattheturbulentReynoldsstressesaretheproductofthemeanvelocitygradientsandeddyviscosities. @x+@v @y+@w @z=0;(2{14)@u @t+@uu @x+@uv @y+@uw @z+1 @x1 @x2+@2u @y2)+@ @z(AV@u @z); @t+@vu @x+@vv @y+@vw @z+1 @y1 @x2+@2v @y2)+@ @z(AV@v @z); whereu(x;y;z;t),v(x;y;z;t),andw(x;y;z;t)arethevelocityvectorcomponents[LT1]inx-,y-,andz-coordinatedirections,respectively;tistime[T];(x;y;t)isthefreesurfaceelevation[L];gistheaccelerationofgravity[LT2];AHandAVarethehorizontalandverticalturbulenteddycoecients,respectively[L2T1];Sxx,Sxy,Syyareradiationstresses,PaisatmosphericpressureandfistheCorioliscomponent[T1]. Thenecessaryconditionsforthesolutionarethedenitionofthecomputa-tionaldomain,theinitialconditionsonthedomain,andtheboundaryconditions.

PAGE 57

Waterelevationisrstsolvedbyusingapre-conditionedconjugategradi-entmethodforthePoissonequationofwaterelevation.Contravariantvelocitycomponentsarethenobtainedbysolvingthemomentumequations. Athreedimensionaladvection-diusionequationforsalinityissolvedcoin-cidentallywiththeequationsofmotionandcontinuity,whichallowsforvariabledensityandbaroclinicforcing.InCartesiancoordinates,theconservationofsaltandtemperaturecanbewrittenas: @t+@uS @x+@vS @y+@wS @z=@ @x(DH@S @x)+@ @y(DH@S @y)+@ @z(Dv@S @z);(2{17) @t+@uT @x+@vT @y+@wT @z=@ @x(KH@T @x)+@ @y(KH@T @y)+@ @z(Kv@T @z);(2{18) whereSissalinityandTistemperature,DH,KH,DvandKvareturbulenteddydiusivitycoecientsforsalinityandtemperatureinhorizontalandverticaldirection,respectively. Afterdeningdimensionlessvariablesas (x;y;z)=(x;y;zXr/Zr)/Xr(u;v;w)=(u;v;wXr/Zr)/Urt=tf=g/(fUrXr)Sij=Sij/(wU2r)Pa=Pa/(wfUrXr)=(wg#)/(wfUrXr)where#ispressureheadAH=AH/AHrAV=AV/AVr(2{19) equations 2{14 through 2{16 becomedimensionless: @x+@v @y+@w @z=0(2{20)

PAGE 58

@t+R0@uu @x+@uv @y+@uw @z+R0@Sxx @x@# @x+v+EHAH@2u @x2+@2u @y2+EV@ @zAV@u @z @t+R0@vu @x+@vv @y+@vw @z+R0@Syx @y@# @yu+EHAH@2v @x2+@2v @y2+EV@ @zAV@v @z where Inacurvilinearnon-orthogonalboundaryttedgridsystem,thenon-dimensionalformofthecontinuityandxandymomentumequationscanbewrittenas: @t+ @(p @(p @=0;(2{24) 1 @t=(g11@ @+g12@ @)(g11@# @+g12@# @)+(g12 (2{25) @(yp @(yp @(xp @(xp @(yp @(yp @(xp @(xp @g+Ev @(Av@u @+EHAH(HorizontalDiffusionofu)R0 @+g12@ @)d+(g11@H @+g12@H @)(Z0d+)]

PAGE 59

1 @t=(g21@ @+g22@ @)(g21@# @+g22@# @)(g11 (2{26) @(yp @(yp @(xp @(xp @(yp @(yp @(xp @(xp @g+Ev @(Av@v @+EHAH(HorizontalDiffusionofv)R0 @+g22@ @)d+(g21@H @+g22@H @)(Z0d+)] where h+=z Hstretchinginthevertical (2{27) @@x @+@y @@y @Horizontalmeasuresoflengths (2{28) @x @@y @@x @@y @2Jacobianofhorizontaltransformation (2{29) Thesalinitytransportequationcanbewrittenas @t=Ev @Dv@S @R0@H!S @R0 @(p @(p @(p @+p @)]; +Eh @(p @+p @)] ModesplittingtechniqueisappliedinCH3Dtosolvethefullthree-dimensionalequations.First,\externalmode"isusedtosolvetheverticallyintegratedequa-tionsofmotionandcontinuityoverthewholedomain.Thenthethree-dimensional

PAGE 60

equationsofmotion,continuityandtransportforagivencellissolvedinthe\in-ternalmode."InthecurrentversionofCH3D,sweepingmethodisusedintheexternalmode.Thei-sweepcouplesthecontinuityandu-momentumequationsandsolveforintermediateelevationandnewstepvelocityu;Thenj-sweepcombinesthecontinuityandv-momentumequationsandsolvefornewstepelevationandvelocityv. Theboundaryconditionsthatmustbespeciedarethetidalandwindforcing,riverinow,andsalinityprolesalongopenboundaries.Thetide,riverinowcanbespeciedaseitherconstantortimevarying.Theinitialconditionswhichmustbespeciedarethethreedimensionaloweld,salinityeldaswellasthewatersurfaceelevation.Theparameterizationofturbulenceinthemodelhasthreeoptions,aconstanteddycoecient,aRichardson-numberdependenteddycoecient,andasimpliedapplicationofthesecond-orderturbulenceclosuremodel. Slightlymodiedversionofarobustwetting-and-dryingschemeby CasulliandCheng ( 1992 )isincorporatedintotheCH3Dmodelasdescribedin Davis ( 1996 )and Shengetal. ( 2002 ). Duetotheuseofthenon-orthogonalboundaryttedequationsofmotionandcontinuity,themodelcanhandlefairlycomplexgeometrieswithoutexcessivenumberofgridcells.Inaddition,thecodeusesasigma-stretchingintheverticaldirectionwhichallowsforvariationinthebottombathymetry. Themodelsimulatesthestormsurgeandtidesubjectedtoprescribedhur-ricanewindandoshoretideforcing.Themodelhasbeentestedwithanalyticalsolutionaswellasstormsurgedataduringrealstorms( Peeneetal. , 1993 ).Cou-pledwithsomewavemodels(REF/DIFandSWAN),theCH3Dmodelhasalsobeenusedtoestimatewavesetup(e.g., SunandSheng ( 2002 ); ShengandAlymov ( 2002 )).

PAGE 61

Toallowecientsimulations,theCH3Dmodelhasbeenmodiedtoallowparalleloperationonasharedmemorycomputer( DavisandSheng , 2000 )and( DavisandSheng , 2002 ). CasulliandCheng ( 1992 )hasbeenimplementedinCH3D( Shengetal. , 2002 ).First,thealgorithmisim-plementedintotheverticallyaveragedequationsofCH3D.Duringeachtimestep,theverticallyaveragedequationsofCH3Darerstsolved,andanewshorelineiscalculated.Thisnewshorelineisthenimplementedinthecalculationofthethree-dimensionalbaroclinicoweld. Inthecurvilinearcoordinatesystem,thetwo-dimensionalverticallyaveraged,non-dimensionalequationscanbewrittenas @t+ @(p u)+@ @(p v)]=0(2{31) u @t+g11@ @g12 v @t+g22@ @+g11 whereFandFaretheremainingnonlinear,horizontaldiusion,windstress,bottomfriction,radiationstress,atmosphericpressuregradientandsurfaceslopeintheoppositedirectionsterms,and whereH=h+istotalwaterdepth.

PAGE 62

Thenitedierenceformofthesimpliedequations 2{31 , 2{32 ,and 2{33 canbewrittenas (2{36) +(11) (2{37) (2{38) where1and2arethedegreesoftheimplicitnessofthesurfaceslopeandCoriolisterms.Substitutingtheverticallyaveragedvelocitynitedierenceequationsintothecontinuityequationyields: nnw;i;jn+1i1;j+1+nn;i;jn+1i;j+1+nne;i;jn+1i+1;j+nw;i;jn+1i1;j+nc;i;jn+1i;j+ne;i;jn+1i+1;j+1+nsw;i;jn+1i1;j1+ns;i;jn+1i;j1+nse;i;jn+1i+1;j1=(RHS)ni;j where

PAGE 63

nnw;i;j=i;j+1 ns;i;j=nsw;i;jnse;i;jv;i;jnw;i;j=+i;j (RHS)ni;j=ni;jp (p +t(11) (p where

PAGE 64

(2{42) 1+t2g21v;i;j CasulliandCheng ( 1992 ),thenormalizedformoftheequationcanbewrittenas: which,byletting isequivalentto

PAGE 65

where Theconjugategradientalgorithm( CasulliandCheng , 1992 )solvesthesystemasdescribedin Davis ( 1996 ): (1)Guesse(0)i;j (3)Fork=0,1,2,...anduntil(r(k);r(k))<,calculate where

PAGE 66

(p(k);Mp(k))r(k+1)i;j=r(k)i;j(k)(Mp(k))i;jp(k+1)i;j=r(k+1)i;j+(k)p(k)i;j(2{49) where (r(k);r(k))(2{50) Intheequation,Mpissolvedasfollows (Mp(k))i;j=p(k)i;j+anw;i;jp(k)i1;j+1+an;i;jp(k)i;j+1+ane;i;jp(k)i+1;j+1+aw;i;jp(k)i1;j+ae;i;jp(k)i+1;j+asw;i;jp(k)i1;j1+as;i;jp(k)i;j1+ase;i;jp(k)i+1;j1(2{51) Oncethefreesurfacehasbeencomputedthroughthecomputationaldomain,beforeproceedingtothenexttimestepthenewtotaldepthatuandvhorizontallocationshavetobeupdated. AresultingzerovalueofthetotaldepthinthecellcenterHn+1i;j=hi;j+ni;jmeansthecellisdryanditmaybeoodedwhenthetotalwaterdepthbecomespositive.SincetheCH3Dmodelusesthetotaldepthasadenominatorinsomeofitsequations'terms,thezerototaldepthvalueisnotalwaysagoodwayofdistinguishingbetweenwetanddrycells.Also,verysmalltotaldepthmightbringininstability,e.g.,verystrongwindblowingoverextremelyshallowwaterwillresultinahighlyunpredictablebehaviorofhorizontalvelocitieswhichmaygrowunrealisticallyhighandcausethemodeltoblowup.Tosolvetheproblem,a\critical"totaldepthvalueasopposedtothezerototaldepthvaluewasusedtodistinguishbetweenwetanddrycells.Ifthetotaldepthofacellissmallerthanthecriticalvaluethenthecellisdry,otherwiseitiswet.Thevalueof30

PAGE 67

cmperformedwellunderhurricanestrongwindconditions.Underconditionslessextremeasmallervalueofthecriticaltotaldepthmaybemorepractical. whereWs=p Garratt ( 1977 )formulation: Theboundaryconditionatthebottomisexpressedintermsofbottomstressgivenbythequadraticlaw: whereubandvbarebottomvelocitiesandCdisthedragcoecientwhichisdenedusingtheformulationof Sheng ( 1983 ): where=0:4isthevonKarmanconstant. Theformulationstatesthatthecoecientisafunctionofthesizeofthebottomroughness,z0,andtheheightatwhichubismeasured,z1iswithintheconstantuxlayerabovethebottom.ThesizeofthebottomroughnesscanbeexpressedintermsoftheNikuradseequivalentsandgrainsize,ks,usingtherelationz0=ks=30.

PAGE 68

Inthetwo-dimensionalmode,thebottomboundaryconditionsaregivenusingaChezyformulation: @z=gUp @z=gVp whereUandVaredepthaveragedvelocities,andCzistheChezyfrictioncoecientdenedas: 6 whereRisthehydraulicradiuswhichcanbeapproximatedbythetotaldepthgivenincentimeters,andnisManning'sn. Donelanetal. ( 1993 )forsurfaceroughness,z0,anddragcoecient,Cdareusedtocalculatewindstressatthefreesurface.Botharefunctionsofwaveage.Whenwavesareyoungtheroughnessincreasesmakingthewindstresshigherasopposedtowhenwavesarenottakenintoconsideration. whereWsisthewindspeedata10maltitude. Followingtherelationbetweenz0andCd,z0=zexp(=p whereCpiswavephasespeedandWs=Cprepresentstheinversewaveage.

PAGE 69

WaveenhancementofbottomstressisimplementedinCH3Dusingtwomethodologies.Therstmethodologyexploitsthe GrantandMadsen ( 1979 )the-orydescribedinasimpliedformby Signelletal. ( 1990 ).Thesecondmethodologymakesuseofaone-dimensionalwave-currentbottomboundarylayermodel( ShengandVillaret , 1989 )asdescribedin ShengandVillaret ( 1989 )and Sun ( 2001 ). The GrantandMadsen ( 1979 )formulationisgivenbythetypicalquadraticlawwithonedistinction:Cdeisthewaveenhanceddragcoecient. ThemainideausedintheformulationinordertondCdeisthat whereCisthebottomstressduetocurrentandWisthemaximumstressduetowaveswhichcanbedenedas 2fWU2W(2{64) whereUWisthenear-bottomwaveorbitalvelocityandfWisthewavefrictionfactorwhichdependsonthebottomroughness,ksandthenear-bottomexcursionamplitudeAB=UW=!andwhosevaluesareobtainedusingtheempiricalexpressionsfrom GrantandMadsen ( 1982 ):

PAGE 70

WithuBdetermined,aniterationprocedureisusedtodetermineCdeatzr.WithaninitialguessofCde,thesteadyshearstresscomponentuBis Thecombinedwave-currentshearvelocityuCWisdenedby Fromequation 2{63 ,uCWisdeterminedas TheapparentbottomroughnesskBC,whichindicatestheturbulencelevelduetothecombinationofthewaveboundarylayerandthephysicalbottomroughness,isexpressedas wheretheexponentisgivenby Theapparentroughnessisthenusedtodeterminethevelocityproleintheconstantstressregionabovethewaveboundarylayerusingthelaw-of-the-wallrelation kBC/30(2{71) Thenalexpressionforthewaveenhanceddragcoecientis wherezrisareferenceheightchosentolieabovethewaveboundarylayerandkBCistheapparentbottomroughnesswhichaccountsfortheturbulenceinducedby

PAGE 71

boththewaveboundarylayerandphysicalbottomroughness.Accordingto Signelletal. ( 1990 ),thereferenceheightwasspeciedas20cmandks=0.1cmwasselectedtocorrespondtoadragcoecientof1.5103atonemeterabovethebedintheabsenceofwaves. OncetheeectivedragcoecientCdeiscalculated,itisusedinCH3Dtocomputebottomstressasdenedbyequation 2{62 . AswaspointedoutinSection 1.1.2 ,the GrantandMadsen ( 1979 )method-ologyhassomedeciencieswhichincludeactitiousreferencevelocityatanunknownlevelandtheassumptionofthelogarithmiclayerbeingconstantwhichisnotcorrectwhenwavesarepresent. Sun ( 2001 )useda1-Dwave-currentbottomboundarylayermodelbasedon ShengandVillaret ( 1989 )tocalculatebottomshearstressthroughnonlinearinteractionbetweenwavesandcurrents.ThismodelwasadoptedandimplementedinCH3Dinthefollowingparagraphs. Thegoverningequationsforthecombinedwave-currentbottomboundarylayermodelaretheverticalone-dimensionalequationsofmotion: @t=1 @x+@ @zAv@u @z(2{73) @t=1 @y+@ @zAv@v @z(2{74) Boundaryconditionsatthebottomare: @z=Cdu1q @z=Cdv1q whereu1,v1arevelocitycomponentsatthelowestgridpoint,z1,andCdiscomputedby:

PAGE 72

wherez0isthebottomroughness(itwassetto0.1cm)andisthevonKarmanconstant.Thesmallestgridspacingnearthebottomis0.03cm. Boundaryconditionsabovethebottomboundarylayerwhichwassetto30cmare: @z=0(2{78) @z=0(2{79) Todriveanoscillatorymotionduetowaves,apressuregradientfromthelinearwavetheoryisapplied: @xw=1 2gkHcosh(kz) cosh(kh)sin'cos(t)(2{80) @yw=1 2gkHcosh(kz) cosh(kh)cos'cos(t)(2{81) wheregisgravitationalacceleration,kiswavenumber,Hiswaveheight,'iswavedirection,andisangularwavefrequency. Todriveacurrent,aconstantpressuregradientisappliedintheydirection: @yc=const(2{82) TheeddyviscosityAvisdeterminedusingaTKEclosuremodeldevelopedby ShengandVillaret ( 1989 ).Themodelsolvesanequationfortheturbulentkineticenergy,q2: @z2hv0w0i@v @z+0:3@ @zq@q2 Thesecond-ordercorrelationtermsofuctuatingvelocitiesaresolvedusingthefollowingequilibriumcondition: 4(2{84) where

PAGE 73

=q =81q Themacro-scaleisdeterminedbythefollowingintegralconstraints: N(2{87) whereC1isbetween0.1and0.25,Histhetotaldepth,Hpisthedepthofpycnocline,C2rangingbetween0.1and0.25isthefractionalcut-olimitationofturbulentmacro-scalebasedonq2,thespreadoftheturbulencedeterminedfromtheturbulentkineticenergyprole,andNistheBrunt-Vaisalafrequencydenedas: @ @z1=2(2{88) ShengandVillaret ( 1989 )and Sun ( 2001 )validatedthemodelforanoscil-latoryboundarylayerbycomparingthecalculatedvelocityproleswithvelocityprolesobtainedinthe JonssonandCarlsen ( 1979 )experimentontheoscillatoryboundarylayerunderroughturbulentowconditions.Twotestsimulations(refertoSection 4.3.2 )wereperformedtovalidatethemodel:forapureoscillatoryow( JonssonandCarlsen , 1979 )anduniformcurrentsuperimposedonanoscillatoryow( BakkerandDorn , 1978 ).Themodelresultsofbothtestsagreedwellwiththeexperimentaldata. Aseriesofmodelruns(atotalof145200runs)usingvariouscombinationsofwaterdepth,waveheight,waveperiod,wavedirectionandcurrentmagnitude

PAGE 74

(showninTable 2{2 )wasperformedinordertodevelopa\look-uptable"ofbottomstressduetowave-currentinteraction.WhenthetableisusedinCH3D,thebottomstressvalueineachgridcellisdeterminedbasedonthevalueobtainedbyinterpolatingthetablevaluesinave-dimensionalspace(i.e.,waterdepth,waveheight,waveperiod,wavedirection,currentmagnitude).Thecurrentisspeciedatthelowestgridpoint,z1,whereCH3Dcalculatesitsbottomcurrents.Therefore,thewaterdepth(i.e.,watercolumnwithinwhichthe1-Dmodelisapplied)isdenedas1 2(h+) sinhk(h+)(2{89) wherehislocalwaterdepth,iswatersurfaceelevation,KMisthenumberofverticallayersinCH3D,andH(z=)iswaveheightatthesurface. Table2{2:Parametersusedtocreatethe\lookuptable". Parameter Values WaterDepth 0.5mto5.0mwith0.5mincrements WaveHeight 0.0mto2.0mwith0.2mincrements WavePeriod 2sto16swith1sincrements WaveDirection 0degto315degwith45degincrements Current 0.0m/sto1.0m/swith0.1m/sincrements 2{15 and 2{16 ).Twoformulationsareimplementedinthemodel:verticallyuniform( Longuet-HigginsandStewart , 1964 )andverticallyvarying( Mellor , 2003 ).

PAGE 75

Thederivationoftheverticallyuniformradiationstressandtherelationbetweenradiationstressgradientsandwavesetupfollowsinthemannerof DeanandDalrymple ( 1991 ). Ifawaveispropagatingatsomeangletothexaxis(representingonshoredirection),thentheradiationstressinthisdirectionwillbe: 2(2{90) Similarly,theradiationstressinthetransverse(longshore)directionwillbe: 2(2{91) wherenistheratioofgroupvelocitytowavecelerity. Thereisanadditionaltermwhichrepresentstheuxinthexdirectionoftheycomponentofmomentum Inasimple1-Dcase,therelationbetweentheradiationstressandwavesetupcanbeexpressedasfollows: dx(2{93) whereisthemeanwatersurfaceslopeorwavesetup. In2-Dcase,asystemoftwodierentialequationshastobesolved: dx(2{94) dy(2{95) Mellor ( 2003 )derivedequationsforthree-dimensionaloceancirculationmodelsthathandlesurfacewaves.Inhisderivationradiationstressesvaryverticallyand

PAGE 76

areexpressedasfollows: whereHisthetotaldepth;=(z)/H;kisthewavenumberwhosexandycomponentsarekandk,respectively;Eisthetotalwaveenergy;istheKroneckerdelta;and sinhKHFCS=coshKH(1+) sinhKHFSC=sinhKH(1+) coshKHFCC=coshKH(1+) coshKH(2{97) IfMellor'sequationsareintegratedvertically,hisradiationstressesbecomeidenticaltothoseof Longuet-HigginsandStewart ( 1964 ).SinceCH3Dcalculatesthewatersurfaceelevationinthe\externalmode"whereverticallyintegratedequationsareused,thecalculatedwavesetupwillbethesamenomatterwhichradiationstressformulationisused.Thedierencebetweenthetwoformulationswilltakeplaceinthevelocityeld. Booijetal. , 1999 )isathird-generationwavemodelwhichcomputesrandom,short-crestedwind-generatedwavesincoastalregionsandinlandwaters.Itaccountsforwavepropagationintimeandspace,shoaling,refractionduetocurrentanddepth,frequencyshiftingduetocurrentsandnon-stationarydepth,wavegenerationbywind,bottomfriction,depth-inducedbreaking,andtransmissionthroughandreectionfromobstacles.ThemodelwasdevelopedatDelftUniversityofTechnology,theNetherlands.SWANisusedincoastalapplicationsbymanyinstitutionsintheUnitedStatesandEurope.

PAGE 77

SWANiscanbeappliedtoaboundary-ttedcurvilineargridwhichisirreg-ular,quadrangular,andnotnecessarilyorthogonal.Itcalculatesmanyimportantwaveandwaverelatedparameterssuchassignicantwaveheight,swellwaveheight,meanwavedirection,peakwavedirection,directionofenergytransport,meanabsolutewaveperiod,meanrelativewaveperiod,currentvelocity,energydissipationduetobottomfriction,wavebreakingandwhitecapping,fractionofbreakingwavesduetodepth-inducedbreaking,transportofenergy,waveinducedforce,theRMS-valueofthemaximaoftheorbitalvelocitynearthebottom,theRMS-valueoftheorbitalvelocitynearthebottom,averagewavelength,averagewavesteepness,andsomeothers. InSWANthewavesaredescribedwiththetwo-dimensionalwaveactiondensityspectrumN(;)equaltotheenergydensitydividedbytherelativefrequency:N(;)=E(;)=. TheevolutionofthewavespectrumisdescribedbythespectralactionbalanceequationwhichforCartesiancoordinatesis: @tN+@ @xcxN+@ @ycyN+@ @cN+@ @cN=S (2{98) Thersttermintheleft-handsideofthisequationrepresentsthelocalrateofchangeofactiondensityintime,thesecondandthirdtermrepresentpropagationofactioningeographicalspace(withpropagationvelocitiescxandcyinx-andy-space,respectively).Thefourthtermrepresentsshiftingoftherelativefrequencyduetovariationsindepthsandcurrents(withpropagationvelocitycin-space).Thefthtermrepresentsdepth-inducedandcurrent-inducedrefraction(withpropagationvelocitycintheta-space).Theexpressionsforthesepropagationspeedsaretakenfromlinearwavetheory.ThetermS(=S(;))attherighthandsideoftheactionbalanceequationisthesourcetermintermsofenergy

PAGE 78

densityrepresentingtheeectsofgeneration,dissipationandnonlinearwave-waveinteractions. TransferofwindenergytothewavesisdescribedinSWANwitharesonancemechanismandafeed-backmechanism.Thecorrespondingsourcetermforthesemechanismsiscommonlydescribedasthesumoflinearandexponentialgrowth: inwhichAandBdependonwavefrequencyanddirection,andwindspeedanddirection. Thedissipationtermofwaveenergyisrepresentedbythesummationofthreedierentcontributions:whitecappingSds;w(;),bottomfrictionSds;b(;)anddepth-inducedbreakingSds;br(;). Whitecappingisprimarilycontrolledbythesteepnessofthewaves.InSWANthewhitecappingformulationsarebasedonapulse-basedmodel: whereisasteepnessdependentcoecient,kiswavenumberand~and~kdenoteameanfrequencyandameanwavenumber,respectively. Depth-induceddissipationmaybecausedbybottomfrictionwhichcangenerallyberepresentedas: inwhichCbottomisabottomfrictioncoecient.JONSWAPsuggestedtouseanempiricallyobtainedconstant.Itperformswellinmanydierentconditionsaslongasasuitablevalueischosen(typicallydierentforswellandwindsea). Thetotaldissipation(i.e.,integratedoverthespectrum)duetodepth-inducedwavebreakinginshallowwatercanbewellmodeledwiththedissipationofabore

PAGE 79

appliedtothebreakingwavesinarandomeld.TheexpressionusedinSWANis: whereEtotisthetotalwaveenergyandDtot(whichisnegative)istherateofdissipationofthetotalenergyduetowavebreaking. Indeepwater,quadrupletwave-waveinteractionsdominatetheevolutionofthespectrum.Theytransferwaveenergyfromthespectralpeaktolowerfrequencies(thusmovingthepeakfrequencytolowervalues)andtohigherfrequencies(wheretheenergyisdissipatedbywhitecapping).Inveryshallowwater,triadwave-waveinteractionstransferenergyfromlowerfrequenciestohigherfrequenciesoftenresultinginhigherharmonics(low-frequencyenergygenerationbytriadwave-waveinteractionsisnotconsideredhere).InSWANthecomputationsarecarriedoutwiththeDiscreteInteractionApproximationandtheLumpedTriadApproximation. SWANhasbeensuccessfullytestedandappliedtostormconditionsinsimulationof1995HurricaneLuisby Wornometal. ( 2001 ).ThemodeliscurrentlyusedbytheNavalResearchLaboratorywhichcreatedthreesub-regionalscalewaveforecastingsystemsfortheNationalWeatherServicesCoastalStormsProgram( Rogers , 2005 ).Oneoftheirsystems,northeastFloridasystem,wasvalidatedfortheperiodofSeptember2003whenHurricaneIsabeltookplaceandfortheperiodofAugust2004whenHurricaneCharleywentacrosstheFloridapeninsula.TheresultsofthesevalidationsshowthatonecanhavecondenceinapplyingtheSWANmodeltoshallow-waterregionsduringseverestormevenssuchashurricanes. ShengandAlymov ( 2002 )usedanextremelylargeandnegridintheirSWANsimulations.Thesizeofthegridwas1200800gridcellswiththegridspacingof5m.Althoughtorunasimulationonsuchanegridrequiresalotof

PAGE 80

computermemoryandcomputationaltime,itisworthwhiledoingit.Forstormsurgemodelingsuchneresolutionmaynotbenecessary.Increasingthegridspacingto100-400mwillallowtostillhavenelyresolveddomainbutalsoplacetheopenboundarymanytensofkilometersoshore. SomeofthedecienciesoftheSWANmodelincludetheassumptionofthewavespectrumbeingGaussianwhichmaynotbetrueinthebreakerzone( Ochi , 1998 ),andtheincapabilityofreplicatingextremedissipationimpartedbymuddybottom( KaihatuandSheremet , 2004 ).

PAGE 81

3{1 .Thoseinredareaccountedforinthemethodologypresentedherein. Oneofthemajoradvantagesofthismethodologyisthatunlikemanyotherstormsurgemodels,thewaterelevationiscalculateddynamicallyandnotinpartswhichincludestormsurge,tideandwavesetupinthesurfzone.Itisalldoneasawholeduetothereasonthatallthesepartsthatwaterelevationconsistsofinteractbetweeneachotherinanon-linearfashionanditwouldbephysicallyincorrecttoneglectthisinteraction. 58

PAGE 82

Figure3{1: Adiagramofvariousphysicalprocesses.Thoseinredareaccountedforinthismethodology.

PAGE 83

event;andacirculationmodelwhichcalculatessurfacewaterelevationandcurrentsbasedonthemeteorologicalconditionsprovidedbytheatmosphericmodel. Aswasstatedearlier,waveeectscanbecomeasignicantfactorduringstormevents.Wavesbreakinginthesurfzonetransfertheirmomentumtothewatercolumnwhichresultsinwaveset-uporincreaseinwaterelevationfromthebreakerpointtothepointwherethewavescompletelydissipatebyrunningupthebeach.Wavesetupismainlyafunctionofthebreakingwaveheight.Stormyseasgeneratelargewaveswhichbreakfurtheroshoreasopposedtowavesduringregularconditions,thusextendingthesurfzone.Asaresult,alargerareaisaectedbytherisingwaterduetowaveset-upwhichcanbesignicantandhastobeaccountedfor.Therefore,addingawavemodelasanothercomponentofstormsurgecalculationaddsrobustnesstotheentirestormsurgemodelingsystem. Horizontalscaleandgridspacingarealsoimportantaspectsofstormsurgemodeling.Sincehurricanesaectlargeareas,computationaldomainsmustbelargeaswell.Havingsmallgridspacingresultsinmoreaccurateresultsbutitalsomeansmorecomputationalcellsand,asaresult,morecomputationalresourcesandtime. Modelsoftwodierentscalesareusedinstormsurgemodelingingeneralandthisstudyinparticular:(1)Regionalmodelswhichsimulatelargeregions(e.g.,theNorthAtlantic),havearelativelycoarsespatialresolutionandprovideboundaryconditionsforlocalmodels;(2)Localmodelswhichsimulatesmallerareasofparticularinterest(e.g.,TampaBayorChesapeakeBay)andhavenespatialresolution. Themethodologypresentedhereinconsistsoffourmodels.Tworegionalmodels,ADCIRCandWAVEWATCH-III,andtwolocalmodelsCH3DandSWAN.ADCIRCandCH3Darecirculationmodels,andWAVEWATCH-IIIandSWANarewavemodels.

PAGE 84

Asfortheatmosphericmodel,resultsfromtwosophisticatedatmosphericmodels,NCEPWNAandWINDGEN,wereused.VericationofthesemodelresultsforHurricanesIsabel(2003),Charley(2004),andFrances(2004)canbefoundinChapter 5 . Forsimulationofstormsurgeduringaparticularhurricane,thebasiccouplingprocessisasfollows: First,windandatmosphericpressuresnapshotsareinitialized,i.e.,thewindandatmosphericpressuredatafromoneofthewindmodelsareprocessedtomakethemavailableforthecirculationmodels,ADCIRCandCH3D. Second,waveboundaryconditionsforthelocalwavemodel,SWAN,areinitializedbyprocessingthewavedatacalculatedbytheregionalwavemodel,WAVEWATCH-III,byNOAA. Third,tidalconstituentsalongtheopenboundariesofthelocalcirculationmodel,CH3D,areinitialized.TheseconstituentsarebasedontheADCIRCtidaldatabase( CATS/tides/tides.htm Attheconclusionoftheinitializationphase,thesimulationphasestarts.Therearethreemodelsthatareinvolvedintheactualsimulation:ADCIRC,CH3D,andSWAN.Therearetwotypesofcouplingbetweenthesemodels.Therstoneisone-waycouplingmeaningthattheresultsofmodelAarefedtomodelBwhoseresultsarenotfedbacktomodelA.ThistypeofcouplingoccursbetweenADCIRCandCH3D. Theaccuracyofresultscalculatedbythelocalmodel,CH3D,dependsonaccuraterepresentationofphysicalprocessesinsidethecomputationaldomainandCH3Dboundaryconditions.TheopenboundaryconditionscanbeprovidedeitherbyavailableelddataorthroughcouplingwitharegionalscalemodelsuchasADCIRC.TheADCIRCdomaincoverstheEasternNorthAtlantic,theGulfof

PAGE 85

MexicoandtheCaribbeanSea.Therefore,ADCIRCiscapableofprovidingwaterelevationalongopenboundariesofalocalmodel(CH3D)domain(e.g.,NorthCarolinaorFlorida)duringhurricaneeventswhentheactualhurricaneislocatedthousandsofkilometersaway.Bothmodelscanrunconcurrentlyusingthesametimestep.ADCIRCresultsarenotaectedbyresultsofCH3D.Thesameone-waycouplingisusedtocoupleWAVEWATCH-IIIandSWAN. Thesecondtypeofcouplingistwo-waycouplingwhichisalsoknownasdynamiccoupling,wheretheresultscalculatedbymodelAarefedtomodelBwhoseresultsarefedbacktomodelA.Thistypeofcouplingisusedtocouplethelocalcirculationmodel,CH3D,andthelocalwavemodel,SWAN. CH3Dsolvesforwaterelevationandcurrents.Italsoaccountsforsituationswhenlandcellsbecomeoodedandviceversa.SWANcomputeswaveconditionswithinthesamecurvilineargridasusedinCH3D.Sincewaveconditionschangerelativelyslowly,thewavemodelsimulationwasconductedevery30minutestoeasethecomputationalburden.Thismeansthatafterthirty60-secondCH3Dtimesteps,thetwomodelsmutuallyexchangeinformation.CH3Dloadsinwaveinformation(waveheight,waveperiod,andwavedirection)toaccountforwavesetup.SWAN,inreturn,updatesbathymetrythatchangesintimeduetotide,stormsurge,wavesetup,andinundationofpreviouslandareas.ThecurrenteldusedinSWANsimulationgetsupdatedaswell.Also,theupdatedwindeldispassedontothewavemodelviaCH3D. AdiagramoftheentirecouplingprocessfrominitializationtoconcurrentsimulationandcouplingofADCIRC,CH3D,andSWANisshowninFigure 3{2 .

PAGE 86

Figure3{2: Adiagramofthecouplingprocess.

PAGE 87

4.1.1Description HubbertandMcInnes ( 1999 )showedthattheirstormsurgeheightsatthecoastproducedby'xedcoastline'versionofthemodelwereoverestimatedby17%comparedwiththeinundationversionoftheirmodel.Thisoverestimationcomesasaresultofthewaterbeingpiledupnearthecoastbytheactionofhighwindandthe'xedcoastline'stormsurgemodelnotallowingthewatertopropagateinland. 64

PAGE 88

TheCH3Dmodeliscapableofaccountingforwetting-and-drying.Themethodisbasedonalightlymodiedversionofarobustwetting-and-dryingschemedevelopedby CasulliandCheng ( 1992 ).ThetechnicaldescriptioncanbefoundinSection 2.3.4.2 .Theideais,everytimestepthemodelcalculatesfreesurfaceelevationwhichisthenusedtocalculatetotaldepth.Ifthetotaldepthinacomputationalcellexceedssome\critical"value(i.e.,30cm),thenthecellisconsideredtobewet,otherwisedry. 4{1 below.Thereisaverticalwallinmiddleofthecomputationaldomainwhichextendsfromthebottomtohalfoftheverticalcolumn.Initially,thegridcellstotheleftofthewallarelledwithwaterandtheonestotherightaredry.Therearefouroutputlocationsinthemiddleofeachverticalcolumn.TheresultsofcomputedwatersurfaceelevationineachoutputlocationareshowninFigure 4{2 . Theresultsareinagreementwithwhatonewouldexpect.Thewatersurfaceelevationincolumns1and2dropsquicklyandtendstoreachhalfofthetotalwaterdepthastimeprogresses.Ontheotherhand,watersurfaceelevationincolumns3and4rapidlyincreasesaswaterllsupthepartofthedomaintotherightofthewall.Italsotendstoreachhalfofthetotaldepth.Eventually,the

PAGE 89

Figure4{1: Thewalltestcase:computationallayout. watersurfaceelevationbecomesequaltohalfofthetotaldepththroughouttheentiredomain. Figure4{2: Thewalltestcase:calculatedwatersurfaceelevation.

PAGE 90

makethewaterpileupagainstthebeachandmovethewatermassinlandcausingooding. ThecomputationallayoutofthetestisshowninFigure 4{3 .Thebedismildlysloped(1:100,000)andthewind(0.1dyne/cm2)blowsonshore.Therearefouroutputlocations.Location1isinitiallydryandbecomeswetduringthesimulation.Locations2through4arewetatalltimes.TheresultsofwatersurfaceelevationcalculatedineachlocationareshowninFigure 4{4 . Figure4{3: Thewindtestcase:computationallayout. Theresultsareconsistentwithwhatonewouldexpect.InLocation1,whichisinitiallydry,surfaceelevationdoesnotchangeduringsomeperiodoftimeduetothereasonthatthewaterwhichispilingupagainstthebeachhasnotyetreachedthelocation.ToclarifythebehaviorofthebluelineinFigure 4{4 ,ithastobenotedthatitwasplottedthewaythatwhenacellisdryithasnovalueforwaterelevation.Lateron,whenLocation1getsinundatedbythewaterthatisbeingpushedonshorebytheblowingwind,thewaterelevationstartstogrowfromtheinitialelevationwhichisgroundlevel. Location2isinitiallyunderwaterandwatersurfaceelevationincreasesintimeasexpected.Location3isalittleoshoreandthewatersurfaceelevationincreases

PAGE 91

Figure4{4: Thewindtestcase:calculatedwatersurfaceelevation. morerapidlythanthatofLocation2asaresultofLocation3beingclosertotheopenboundarywherethewindisblowingfromand,therefore,ittakeslesstimeforthewindtoreachLocation3andstarttopileupwaterthere. Location4isfartheroshoreneartheopenboundary.Watersurfaceelevationstartstogrowinthebeginningbutthenitstartstodeclinedueconservationofmass. CarrierandGreenspan ( 1958 )forpropagationofwavesonalinearlyslopingbeachwasapplied.Theone-dimensionalnonlinearshallowwaterequationcanbewrittenas: @t[(+h)u]=0 (4{1) (4{2)

PAGE 92

Theobtainedsolutiontotheequationintheformofapotentialisasfollows: (4{3) whereA0isanarbitraryamplitudeparameterandJ0isazeroBesselfunctionoftherstkind.Thispotentialrepresentsastandingwavesolutionresultingfromaperfectreectionofaunitfrequencywave. Thefollowingcomputationallayoutwassetuptosolvefor(x;t)andu(x;t)forgivenlocation,x,andtime,t.A1615orthogonalgrid62kmlongand10kmwidewithbottomslopingat1:2500.Thedepth,h,variesfrom2mabovethemeansealevelto22.8mbelowthemeansealevelatx=57km.Thegridspacinginthelongshorey-directionisxedandequalto2kmwhereasthegridspacinginthecross-shorex-directionvaries.Fortherst10.5km(goingoshore)thegridspacingisxedat100m,thenfrom10.5kmto15km,thegridspacinglinearlygrowsfrom100mto1kmadding100meverygridcell.From15kmto62km,thegridspacingisxedat1km.Thetidalforcingappliedinthemodelatx=57kmis Tt) (4{4) wheretheamplitude,A,is11.24cmandtheperiod,T,is3600s. Duringthissimulationthenon-lineartermsintheCH3Dmodelwereturnedonsincetheanalyticsolutionwasderivedusingthoseterms.TheresultsofcomparisonofcalculatedsurfaceelevationwithanalyticsolutionatdierenttimesareshowninFigures 4{5 through 4{11 . 4.2.1Description

PAGE 93

Figure4{5: Tidalcase:comparisonwithanalyticsolutionatt=0. Figure4{6: Tidalcase:comparisonwithanalyticsolutionatt=/6.

PAGE 94

Figure4{7: Tidalcase:comparisonwithanalyticsolutionatt=/3. Figure4{8: Tidalcase:comparisonwithanalyticsolutionatt=/2.

PAGE 95

Figure4{9: Tidalcase:comparisonwithanalyticsolutionatt=2/3. Figure4{10: Tidalcase:comparisonwithanalyticsolutionatt=5/6.

PAGE 96

Figure4{11: Tidalcase:comparisonwithanalyticsolutionatt=. atmosphericpressuregradientgivesrisetowatersurfaceelevationintheocean.Thesignicanceoftherisedependsonthemagnitudeofthegradient.InCartesiancoordinatesystem,themomentumequationincludingairpressuretermcanbewrittenas: @t+advection=@Pa @x+diffusion(4{5) @t+advection=@Pa @y+diffusion(4{6) SincetheairpressuretermcanbewrittenaswaterheadPa=wg#,equations 4{5 and 4{6 canberewrittenas: @t+advection=g@# @xg@ @x+diffusion(4{7) @t+advection=g@# @yg@ @y+diffusion(4{8)

PAGE 97

Incurvilinearcoordinatesystemafternon-dimensionalization,theequationstransformintothefollowingform: @t=H(g11@# @+g12@# @)H(g11@ @+g12@ @)+otherterms(4{9) @t=H(g21@# @+g22@# @)H(g21@ @+g22@ @)+otherterms(4{10) whereothertermsincludenonlinearterms,Coriolisterm,diusionterms,surfaceandbottomfrictionterms.Theairpressuretermsinequations 4{9 and 4{10 aretreatedfullyexplicitlywhenthe2Dequationsaresolved. Holland ( 1980 )andsimpliedasin Wilson ( 1957 )(seeSection 2.3.1 ).Thelocalpressureinastormcanbewrittenas: whereP0isthepressureinthecenterofthestorm,P1isthefreestreampressure,risthedistancefromthecenterofthestorm,andRistheradiustomaximumwind. Forsteadystatecondition,thewatersurfaceelevationduetothepressuregradientofthestormcanbewrittenas: whereCisspeciedbyboundaryconditionsorsimplyzeroincaseofsteadystateconditions. Thecomputationaldomainforthetestwasa100km100kmrectangulargrid.Thefollowingparameterswereusedforanalyticalstorm: atmosphericpressureatcenter,P0=960mb;

PAGE 98

freestreampressure,P1=1013mb; radiustomaximumwind,R=30km. TheanalyticalsolutionofwaterlevelisshowninFigure 4{12 andthedier-encebetweentheanalyticalsolutionandnumericalsolutionisshowninFigure 4{13 .Thedierenceisontheorderof0.0001cm,whichvalidatestheaccuracyoftheimplementedintheCH3Dmodelatmosphericpressuregradientterms. Figure4{12: Analyticalsolutionofwatersurfaceelevationduetoatmosphericpressuregradientforasimpliedhurricane. 4.3.1Description 2.3.4.4 ,a1-Dwave-currentbottomboundarylayer(BBL)modelisexploitedinCH3Dtocalculatebottomshearstressthroughnonlinearinteractionbetweenwavesandcurrents.TheeddyviscosityinthemodelisdeterminedusingaTKEclosuremodeldevelopedby Shengand

PAGE 99

Figure4{13: Dierenceinwaterelevationbetweentheanalyticalandnumericalsolutions. Villaret ( 1989 ).TheBBLmodel'sgoverningequationsandintegralconstrainscanalsobefoundinthatsection. ( 1979 )carriedoutalaboratoryexperimentinaU-shapedwatertunnelduringwhichtheymeasuredoscillatoryvelocitiesthroughoutthewatercolumnand,basedonthemeasurements,determinedbottomstresses.Thefollowingparameterswereusedduringtheexperiment: WaterDepth=10m; WaveHeight=5.3m;

PAGE 100

WavePeriod=8.39s; BottomRoughness=0.077cm. Inthenumericalsimulations,theverticalcomputationalgridwassetupinawaythatthegridspacingnearthebottomwasontheorderof0.0001cm;itgrewasthedistancefromthebottomincreased.Thecomputationaltimestepwas0.01s. AscanbeseeninFigure 4{14 ,thecalculatedvelocityprolesareingoodagreementwithmeasuredvelocities.TheRMSerrorsareshownforeachindividualproleandforallprolesaltogether.TherelativeRMSerrorswerecalculatedbasedonthemaximumrangeofmeasuredvelocities(440cm/s).Ithastobenotedthatoneofthesourcesofthecalculatederrormaybeattributedtotheasymmetryofthemeasuredfreestreamvelocityabovetheboundarylayer.Themeasurementsshowedthatthefreestreamvelocityrangedfrom-220cm/sto201cm/swhereastheoscillatoryowimposedthroughboundaryconditionsinthemodelwasimpliedsymmetricallyandrangedfrom-210cm/sto210cm/s. Figure4{14: Comparisonbetweenmeasured( JonssonandCarlsen , 1979 )[dashedlinewithsquares]andcalculated[solidline]velocityprolesforeightphaseangles.

PAGE 101

Figure 4{15 showsthephaselagsofhorizontalvelocitiesatvariouslevels.Thehorizontalvelocitynearthebottomshowsaphaseleadof24o.Theabovemodelresultsareconsistentwiththeresultsof Sheng ( 1982 )and ShengandVillaret ( 1989 ),whichusedaReynoldsstressmodelandaTKEmodel. Figure4{15: Verticalproleofthecalculatedphaselagbetweenhorizontalveloci-tiesandfreestreamvelocity. Figure 4{16 showshowthecalculatedbottomstresscomparesagainstthebot-tomstressdeterminedbasedonthevelocitymeasurementsduringtheexperiment.TherelativeRMSerrorwascalculatedbasedonthemaximumrangeofmeasuredbottomstress(880dyne/cm2).Fromthiscomparison,itcanbeconcludedthatthecalculatedandmeasuredbottomstressesagreewell,whichvalidatestheuseofthe1-DBBLmodelforcomputingvelocitiesandbottomstressesintheoscillatorybottomboundarylayer. BakkerandDorn ( 1978 )laboratoryexperimentwasperformed.Thewaterdepthduringtheexperimentwas0.3m.Theoscillatoryow

PAGE 102

Figure4{16: Comparisonbetweencalculated[solidline]bottomstressandbottomstressdeterminedbasedonmeasurementsduringthe JonssonandCarlsen ( 1979 )experiment[dashedlinewithsquares]. withaperiodT=2swasimposedthoughoscillatorymotionboundaryconditionaccordingto: Tt1+U2sin22 Tt2+U3sin32 Tt3(4{13) wherevelocityamplitudesofthethreeharmonicsweredenedas: andtheircorrespondingphaseanglesweredenedas: Thebottomroughnesswas0.07cmandthecurrentvelocityspeciedatthetopofa6.2cmthickbottomboundarylayerwas22.7cm/s. AcomparisonbetweenmeasuredandcalculatedvelocityprolesshowninFigure 4{17 demonstratesthatthenumericalresultsareingoodagreementwithmeasuredvelocitieswhichvalidatestheuseofthemodelforcomputingvelocityproleswithinturbulentbottomboundarylayerforcombinedwave-currentows.

PAGE 103

TheRMSerrorsareshownforeachindividualproleandforallprolesaltogether.TherelativeRMSerrorswerecalculatedbasedonthemaximumrangeofmeasuredvelocities(55cm/s). Figure4{17: Comparisonbetweenmeasured( BakkerandDorn , 1978 )[dashedlinewithsquares]andcalculated[solidline]velocityprolesforeightphaseangles. BakkerandDorn ( 1978 )didnotdeterminebottomshearstressbasedontheirmeasurements,thusFigure 4{18 showsonlythecalculatedbottomstressoveronewavecyclealongwithitsaveragevalue. 4.4.1Description 1.1.3 ,breakingwavesproduceexcessmomentumuxintheshorewarddirection,radiationstress.Theshorewarddecreaseinradiationstressisbalancedbyashorewardincreaseinthewaterlevel,wavesetup.Wavesetupvariesincross-shoreandlongshoredirectionsasaresultofcomplexbathymetry.Thewidthofthesurfzonedependsonthebeachslopeandincident

PAGE 104

Figure4{18: Bottomstressduetowave-currentinteractioncalculatedusingthe1-DBBLmodelbasedonthenumericalsimulationofthe BakkerandDorn ( 1978 )laboratoryexperiment. waveheight.Forthesamebeachslope,amodestwavebreaksclosertotheshorewhilealargerwavebreaksfurtheroshore. AswaspointedoutinSection 2.3.4.5 ,twoformulationsareimplementedinCH3D:verticallyuniform( Longuet-HigginsandStewart , 1964 )andverticallyvarying( Mellor , 2003 ).WhencoupledwithSWAN,CH3Daccountsforwavesetup,whichisanimportantfactorinstormsurgemodelingandhastobeconsidered.

PAGE 105

Therstexperimentwasby StiveandWind ( 1982 ).Intheexperimenttheauthorsstudiedvariationsofradiationstressandmeanwaterlevelforthetwo-dimensionalshoalingandbreakingofprogressive,periodicwavesonaplane,gentlysloppinglaboratorybeach.Theexperimentwasconductedinawaveumewhichis55mlong,1mwideand1mhigh(seeFigure 4{19 ).Aplaneconcretebeachwitha1:40slopewasinstalled.Theslopeconsistsofthreeparts(i)aslopingzonewhichstartsinawaterdepthof0.85mwherewavesaregenerated,(ii)azonewithaconstantdepthof0.70mtoenabletheinstallationofinstruments,and(iii)anotherslopingbeachzone. Figure4{19: LayoutofStiveandWindexperimentalsetup(from StiveandWind ( 1982 )). Theexperimentwassimulatedusingaslightlydierentcomputationaldomain.Thereasonforthemodicationwasthatthewidthofthewaveumeusedintheexperimentwasonly1manditslengthwasapproximately45m.IntheSWANmodelthereareaectedareaswitherrorsalonglateralboundariesspreadingtotheshoreatanangleofapproximately30o.Therefore,thelateralboundariesshouldbesucientlyfarawayfromtheareaofinteresttoavoidpropagationoftheerrorintothisarea.Thus,thedomaininthetransversedirectionwasexpandedfrom1mto20m.Sincetheproblemisessentiallyone-dimensionalthewidthofthedomainshouldnotmatteraslongastheareaofinterestisnotaectedbythelateralboundaries. Duringtheexperimentthefollowingwaveparameterswereusedbythewavegenerator:waveheightHrms=0.159m,whichwasconvertedtosignicantwaveheightsincetheSWANmodeldoesnotoperatewithmonochromaticwaves,

PAGE 106

Thesecondexperimentwasby MoryandHamm ( 1997 ).Inthisexperimenttheauthorsstudiedtheimpactofadetachedbreakwateroncoastalmorphologyina3-Dwavebasin.Thebasin(seeFigure 4{20 ),30m30mbysize,consistsofthreeparts(i)azone4.4mwidewithaconstantdepthof0.33mwhichistheclosesttothewavemaker,(ii)anunderwaterplanebeachwiththeslopeof1:50,and(iii)anemergedplanebeachwiththeslopeof1:20.Adetachedbreakwater6.66mlongand0.87mwidewasbuiltperpendiculartooneofthelateralwalls. Figure4{20: LayoutofMoryandHammexperimentalsetup(from MoryandHamm ( 1997 )). Oneoftheirtests,whichwasextensivelystudiedexperimentallyandnumeri-cally,wasforaJONSWAPdistributionincidentwavegeneratedbythewavemakerwithHsig=11.5cmandTpeak=1.69s.JONSWAPdistributionisoneoftheoptionsusedinSWAN.TheothertwodistributionsthatthemodeliscapableofsimulatingareGaussianandPierson-Moskowitz.Detailedmeasurementsofwaveheights,

PAGE 107

setupandcurrentsweremadeduringtheexperiment.Wavesetupwasmeasuredat6locationsalongatransectperpendiculartotheshoreline.Theaccuracyofthemeasuredwavesetupvalueswas0.02cm.ComparisonofmeasuredandsimulatedwavesetupisshowninFigure 4{21 . Figure4{21: Comparisonbetweenmeasuredandcalculatedwavesetup( MoryandHamm ( 1997 )experiment). TheresultscompareverywellwiththeaverageRMSerrorbeing0.04cm,whichisontheorderofaccuracyofmeasuredwavesetup,andtheaveragerelativeerrorbetweenallstationsbeing6.25%. 4.5.1DescriptionofCross-shoreandLongshoreCurrents

PAGE 108

SunandSheng ( 2002 )studiedtheimportanceofwave-inducedcurrentsandotherwaverelatedeectsonthegeneralnearshorecirculation.Inordertoaccountfortheseeects,acoupleofwave-relatedfeatureswereincorporatedinCH3D. First,anewsocalledsurfacerollertermwasimplemented. Svendsen ( 1984 )showedthatthesurfacerollerofbreakingwavesplaysanimportantpartinmass,momentumandenergybalanceinthesurfzoneandistheprimarydrivingmechanismfortheundertow.Therollerrepresentsanincreaseinradiationstresswhichaccordingto Svendsen ( 1984 )canbewrittenasfollows: LgH2b(4{14) where,S+xxistheincreaseoftheradiationstressabovethewavetrough,Listhewavelength,hiswaterdepth,andHbisthebreakingwaveheight. Thesecondimplementedtermwasanadditionaltermtotheverticaleddyviscosityinordertoaccountforwaveeects.Following Battjes ( 1975 )and VriendandStive ( 1987 )thewave-enhancedverticaleddyviscosityhasthefollowingform: where,AzcistheeddyviscosityrelatedtothemeancurrentsascomputedbytheequilibriumclosuremodelimplementedintheCH3Dmodel.Dbisthewaveenergydissipationresultedfromwavebreakingandbottomfriction,histhewaterdepthandMisaconstant. Thewaveenergydissipation,Db,wascalculatedaccordingto BattjesandJanssen ( 1978 )asfollows: 4gQbH2m whereQbisthefractionofbreakingwaves,Hmisthemaximumwaveheightthatcanexistatthisdepth,andTisthewaveperiod.Thefractionofbreakingwaveswascalculatedfromthefollowingimplicitrelation:

PAGE 109

1Qb whereEtotisthetotalwaveenergy. MoryandHamm ( 1997 )experimentaldatawereusedforvalidation.Theexperimentstudiedtheimpactofadetachedbreakwateroncoastalmorphology.AdetaileddescriptionoftheexperimentalsetupcanbefoundinSection 4.4.2 .SincediractionisnotmodeledinSWAN,thewaveeldcomputedbySWANmaynotbeaccurateintheimmediatevicinityofobstaclesandwillcertainlynotbeaccurateinharborsorbehindbreakwaters.Therefore,insteadoftheSWANmodeltheREF/DIF,anearshorewavetransformationmodeldevelopedby KirbyandDalrymple ( 1994 )wasused. Oneoftheexperimentswasconductedusingamonochromaticincidentwavegeneratedbythewavemakerwithawaveheightof0.075mandawaveperiodof1.69s.ThemostprominentphenomenonobservedduringtheexperimentwasastrongeddybehindthedetachedbreakwaterwhichcanalsobeseeninFigure 4{22 thatshowscalculatedfreesurfaceelevationalongwithcalculatedcurrentpattern.ThelocationswhereverticalvelocityprolesweremeasuredduringtheexperimentaredepictedbylettersAthroughN.Tosimulatetheexperiment,theREF/DIFmodelwasusedtocalculatewaveparametersandprovideradiationstressestodrivethehydrodynamicmodel. Aftershoalingandapproachingthebreakwater,thewavesdiractedbehindthebreakwaterandeventuallybrokeonthebeach.Maximumcurrentvelocityfortheeddywasmorethan0.3m/swhichagreeswellwithobservations.Thevelocityoflongshorecurrentontheopenbeachwasontheorderof0.1m/sorless,alsoingoodagreementwiththeobservation.Theeddywasalmostuniformoverdepthcomparedwiththedistinctive3-Dstructureofthecurrentsontheopenbeach

PAGE 110

Figure4{22: Calculatedfreesurfaceelevationandcurrentpatternalongwiththelocationswhereverticalvelocityprolesweremeasured(lettersAthroughN).

PAGE 111

duetothepresenceoftheundertow.Overall,thecurrentmodelresultsinthisstudyshowverygoodagreementwithmeasurementsqualitativelyandfairlygoodagreementquantitatively. Manyrunsusingvariouscombinationsofdierentparameterssuchaswavebreakingparameter,,bottomroughness,z0,andverticaleddyviscosityequationconstant,M,wereperformed.Twovaluesforwereused,0.55and0.82.Therangeofz0wasfrom0.0003cmto0.4cm,whereasconstantMvariedform0.001to0.025.ForeachruntherelativeRMSerroroftheverticalvelocityproles(total10proles)wascalculated.TheRMSerrorrangedfrom0.190to0.396.ThelowestrelativeRMSerror(=0.190)wasachievedwhenthefollowingparametervalueswereused:=0.55;z0=0.005cm;M=0.003. Smallervalueofasopposedtomoreconventionalwavebreakingparametervalues(0.8)performedbetterduetothereasonthattheREF/DIFmodeltendedtooverestimatethewaveheightwhichledtooverestimationofcomputedradiationstresses. Comparisonofmeasuredandsimulated(withthelowestRMSerror)long-andcross-shorevelocityprolesisshowninFigures 4{23 through 4{28 . 4{29 showsadiagramoftheinstrumentsetup.Wavesaremeasuredatthreelocations:#630(approximately3900moshore),#625(atthetipofthepier,570moshore),and#641(240moshore). Twohurricaneeventsweretested:HurricaneFloyd(September1999)andHurricaneBonnie(August1998).Thecomputationalgridwassetuptheway

PAGE 112

Figure4{23: Simulated(reddashedline)vs.measured(greensolidline)longshorevelocities:prolesA,B,C,andN. Figure4{24: Simulated(reddashedline)vs.measured(greensolidline)cross-shorevelocities:prolesA,B,C,andN.

PAGE 113

Figure4{25: Simulated(reddashedline)vs.measured(greensolidline)longshorevelocities:prolesD,F,I,andH. Figure4{26: Simulated(reddashedline)vs.measured(greensolidline)cross-shorevelocities:prolesD,F,I,andH.

PAGE 114

Figure4{27: Simulated(reddashedline)vs.measured(greensolidline)longshorevelocities:prolesEandG. Figure4{28: Simulated(reddashedline)vs.measured(greensolidline)cross-shorevelocities:prolesEandG. TheFRFinstrumentsetupatDuck,NC(

PAGE 115

thattheoshoreopenboundarylayparalleltotheshoreatthelocationofwavegage#630.Thiswasdoneforthepurposeofexploitingwaveparameterssuchaswaveheight,periodanddirectionmeasuredatthislocationastheopenboundaryconditionsusedintheSWANmodel.Thus,thegridwentapproximately3900moshoreinthecross-shoredirectionand3400minthealongshoredirection.BottomsurveysatFRFaredoneeverymonth.Inordertousebathymetryasclosetorealityaspossible,bathymetricdatacollectedduringthemonthofSeptember1999wereutilizedtocreatethegrid.Table 4{1 showsHurricaneFloydwave/windinformationusedastheinputforSWANatthetimewhentheoshorestationrecordedthelargestwaveheightduringthehurricaneFloydevent(=4.78m). Table4{1: WaveparametersusedtoimposeHurricaneFloyd(1999)boundaryconditions. Parameter Value JulianDay 257.401 Hsig(#630) 4.78m Tpeak WaveDirection 98o(TrueNorth) WindSpeed 5.3m/s WindDirection 62o(TrueNorth) Tide 0.58m Thetidevaluewasaddedontopofthebathymetryassumingitwasconstantthroughoutthedomain. Table 4{2 showscomparisonbetweencalculatedandmeasuredsignicantwaveheightattwolocationsalongthepier. Table4{2: ComparisonofcalculatedandmeasuredwaveheightduringHurricaneFloyd(1999). Station# MeasuredHsig,m HsigcalculatedbySWAN,m RelativeError,% 641 1.95 1.82 6.6 625 3.16 3.09 2.2 Sincewavesetupduringstormeventscanbeasignicantfactoraectingthemeanwaterlevel,anothersimulationwasmade.First,theCH3Dmodelwasused

PAGE 116

tocalculatewavesetupasafunctionofcrossandlong-shoredirections.ThewavesetupvaluewasthenaddedontopoftheSWLandthewaveeldwasrecalculatedusingtheSWANmodel.ComparisonbetweencalculatedandmeasuredsignicantwaveheightatthetwolocationsalongthepiertakingintoaccountwavesetupisshowninTable 4{3 . Table4{3: ComparisonofcalculatedandmeasuredwaveheightduringHurricaneFloyd(1999)withwavesetupbeingaccountedfor. Station# MeasuredHsig,m HsigcalculatedbySWAN,m RelativeError,% 641 1.95 1.86 4.6 625 3.16 3.10 1.9 Thecomparisonshowsanexcellentagreementbetweenmeasuredandsimu-latedwaveheight.Theinclusionofwavesetuphelpedreducetherelativeerror. Analogously,Tables 4{4 and 4{5 showHurricaneBonniewave/windinforma-tionandcomparisonoftheresults,respectively. Table4{4: WaveparametersusedtoimposeHurricaneBonnie(1998)boundaryconditions. Parameter Value JulianDay 238.635 Hsig(#630) 3.95m Tpeak WaveDirection 98o(TrueNorth) WindSpeed 20.7m/s WindDirection 111o(TrueNorth) Tide 0.28m Table4{5: ComparisonofcalculatedandmeasuredwaveheightduringHurricaneBonnie(1998). Station# MeasuredHsig,m HsigcalculatedbySWAN,m RelativeError,% 641 1.82 1.72 5.5 625 2.96 2.78 6.1 FRFbathymetricsurveyscollectedduringthemonthofAugustof1998wereusedtocreateacomputationalgrid,whichcoveredthesamedomainandhad

PAGE 117

thesamedimensionsastheoneusedincaseofHurricaneFloyd.Wavesetupwasaccountedforinthesamefashionasbefore,whichhelpedslightlyreducetherelativeerror.

PAGE 118

Inthischapterthevalidationofthemodelingsystemisdiscussed.Themaincriterionformodelvalidationishowwellthesimulatedresultscomparewithmeasureddata.Threehurricanesareconsidered:(1)Isabel(2003)intheOuterBanks,NCandChesapeakeBay,VAarea,(2)Charley(2004)inCharlotteHarbor,FL,and(3)Frances(2004)inTampaBay.Athorougherroranalysisofwind,waveandwaterelevationwasperformed.Eachhurricanewassimulatedseveraltimesusingvariouscombinationsofsixmodelfeatures,e.g.,tide,wind,wavesetup,waveenhancedsurfacestress,waveenhancedbottomfriction,andwetting-and-drying.Thegoalofthesesimulationsistoidentifythoseprocessesthataredominantforeachhurricane. 5.1.1DescriptionAccordingtoNHC A fordetaileddescriptionofthescale).ItmadelandfallnearDrumInletontheOuterBanksofNorthCarolinaasaCategory2hurricanearound17:00UTConSeptember18.Ocialreportsstatethat51peoplediedasaresultofthestorm(17directly),withanocialdamageestimateof$3.37billion.ThetrackchartofIsabelisgiveninFigure 5{1 .IsabelbroughthurricaneconditionstoportionsofeasternNorthCarolinaandsoutheasternVirginia.AccordingtoNOAA'sNationalHurricaneCenter,thehighestobservedsustainedwindover 95

PAGE 119

Figure5{1: BesttrackofHurricaneIsabel(courtesyofNOAANHC). landwas35m/swithgustsupto44m/sataninstrumentedtowernearCapeHatteras,NCat16:22UTConSeptember18.AnothertowerinElizabethCity,NCreported33m/ssustainedwindwithagustto43m/sat18:53UTCthatday.TheNationalOceanServicestationatCapeHatterasreported35m/ssustainedwindwithagustto43m/sbeforecontactwaslost.TheCoastalMarineAutomatedStations(C-MAN)atChesapeakeLight,VAandDuck,NCreportedsimilarwinds.GloucesterPoint,VAreported31m/ssustainedwindswithagustto41m/sat22:00UTConSeptember18,whiletheNorfolkNavalAirStationreported26m/ssustainedwindswithagustto37m/sat21:00UTCthatday.Thewindrecordfromthemostseriouslyaectedareasisincomplete,asseveralobservingstationswereeitherdestroyedorlostpowerasIsabelpassed. ThestormsurgeofHurricaneIsabelwas0.3to1.0mhigherthanitwasforecast,especiallyinthenorthernChesapeakeBayandPotomacBasins.Table 5{1

PAGE 120

providesmeasuredstormsurgecrestsatseveralsitesinNorthCarolina,Virginia,andMaryland. Table5{1:MeasuredstormtidecrestsatseveralsitesinNorthCarolina,Virginia,andMaryland. StormTide(m,NAVD88) 0.68 Beaufort,NC 1.23 CapeHatteras,NC 2.04 OregonInletMarina,NC 1.48 Duck,NC 1.83 MoneyPoint,VA 2.06 ChesapeakeBayBridge,VA 1.87 Sewell'sPoint,VA 1.99 Scotland,VA 1.75 King'sMill,VA 1.61 GloucesterPoint,VA 1.46 ColonialBeach,VA 1.66 Kiptopeke,VA 1.54 Wachapreague,VA 1.86 Richmond,VA 2.44 Washington,D.C. 1.98 Baltimore,MD 2.24 Annapolis,MD 1.97 TolchesterBeach,MD 2.16 Cambridge,MD 1.57 OceanCityInlet,MD 0.80 Philadelphia,PA 1.83 ReedyPoint,DE 1.75 Lewes,DE 1.30 CapeMay,NJ 1.18 Burlington,NJ 2.00

PAGE 121

alltheaectedareaswascreated.Thegrid(Figure 5{2 )containstwoopenbound-aries:ThesouthernopenboundarystartsatWilmington,NCandgoes300kmtotheeastwherethecontinentalshelfends,whiletheeasternopenboundaryextends578kmtothenorth.Bothopenboundariesarefarawayfromthecoastlineoftheareasaectedbythehurricane.ThedistancefromtheSouthOuterBankstothesouthernopenboundaryrangesfrom40to80kmwhereasthedistancefromtheEastOuterBankstotheeasternopenboundaryisbetween40and60km.Theareaofthecomputationaldomainis134,385km2withatotalof548,240computationalgridcellsandanaveragegridspacingof500m.192,608(35%)ofthosecomputationalcellsarewatercells.ThegridcoverstheentireChesapeakeBayandallofitsriverbasinsincludinglandcellsforcalculationofwettinganddrying.TheUSGSNationalElevationDataset( E ).Thegridextendsinlandfarenoughtotheheightsoftensofmeterssothatitwouldbeimpossibleforthewatertoreachtheinlandgridboundariesduringhurricaneevents. 5{2 .

PAGE 122

Figure5{2: TheOuterBanksandChesapeakeBaygriddomainforIsabelsimula-tion.

PAGE 123

Table5{2: Tide,windandwavestationsusedforvalidationofthemodelduringHurricaneIsabel. Lat Lon Easting Northing WaterElev.Data WindData WaveData CurrentData (oN) (oW) (UTM18,m) (UTM18,m) Source Source Source Source Vert.Datum AnemometerHeight(m) Depth(m) Depth(m) CapeLookout 34.620 76.520 360653 3831954 C-MAN 9.8 Beaufort 34.720 76.670 347084 3843262 NOAA NAVD88 KittyHawk 36.101 75.710 436065 3995427 FRF 8.5 DuckPier 36.180 75.750 432557 4004174 NOAA C-MAN FRF FRF NAVD88 20.4 8.0 8.0 Duck630 36.168 75.701 436954 4002809 FRF 17.0 ChesapeakeBay 36.960 76.110 400923 4091430 NOAA NOAA Bridge NAVD88 ? ChesapeakeLight 36.910 75.710 436752 4085123 C-MAN 43.3 GloucesterPoint 37.244 76.500 366926 4123035 VIMS VIMS VIMS VIMS NAVD88 ? 8.5 8.5 Kiptopeke 37.170 75.990 412109 4114190 NOAA NOAA NAVD88 ? Lewisetta 38.000 76.470 370934 4206834 NOAA NOAA NAVD88 ? MoneyPoint 36.780 76.300 383996 4071255 NOAA NOAA NAVD88 ? HPLWS 38.590 76.133 401326 4271888 CBOS ? ChoptankRiver 38.634 76.159 399124 4276800 CBOS ? NorthBay 39.375 76.113 404427 4382844 CBOS ?

PAGE 124

RiverdischargedatawereobtainedfromUSGSChesapeakeBayRiverInputMonitoringProgram( 5{3 showsthelocationofthenineriverinputmonitoringsites.RiverdischargevaluesatthesestationsduringthemonthofSeptember,2003areshowninFigure 5{4 .Ascanbeseeninthegure,afterHurricaneIsabelmadelandfall(aroundJulianDay261),therewasasignicantincrease(upto40times)inriverdischarge. Figure5{3: LocationofthenineRiverInputMonitoringsites(courtesyofUSGS). Luettichetal. , 1992 )fortheWesternNorthAtlantic,Caribbean

PAGE 125

Figure5{4: RiverdischargeintoChesapeakeBaydataduringthemonthofSep-tember,2003. andGulfofMexico(allwaterswestofthe60oWMeridianandeastoftheNorthAmericancontinent).M2,S2,N2,K2,O1,K1,Q1,M4,M6andSTEADYtidalconstituentsareincluded. SevenconstituentsareusedintheCH3DmodelduringsimulationofHurricaneIsabel.TheconstituentsandtheircorrespondingperiodsarelistedinTable 5{3 . Table5{3: ADCIRCtidalconstituentsandtheirperiodsusedintheCH3DmodeltosimulateHurricaneIsabel. Constituent Period,hours K1 23.93446966 O1 25.81934167 Q1 26.86835668 M2 12.42060122 N2 12.65834826 S2 12.00000000 K2 11.96723479 Itshouldbenotedthatalthoughthetidaldatabasewaspartiallyvalidated(exceptnonlinearlygeneratedconstituentsM4,M6,andSTEADY)by Mukaietal. ( 2001 ),therearestillerrorsassociatedwiththeconstituents.Inordertoimprove

PAGE 126

tidalsimulationbyreducingerrorsassociatedwiththeADCIRCconstituentsintheOuterBanks/ChesapeakeBayarea,ananalysisoftidalconstituentsbasedonacomparisonbetweensimulatedandmeasuredwaterelevationwasperformed.TheanalysiswasdoneusingtheIOSprogram( Foreman , 1977 )forcalculatingtidalconstituentsbasedonmeasuredorsimulatedtimeseriesofwaterelevation.Atwo-monthperiod,Sep-15,2003throughNov-15,2003,waschosenwithtwotimeseriesofmeasuredwaterelevationsavailableduringthatperiodoftimeattwotidalstations:DuckPier,NCandBeaufort,NC.TheDuckPieranalysiscorrespondstotheeasternopenboundaryandtheBeaufortanalysiscorrespondstothesouthernopenboundary.Tables 5{4 and 5{5 belowshowhowconstituentparameters(amplitudeandphase)comparewitheachother. Table5{4: TidalconstituentparametersatDuckPier,NCcalculatedbasedonADCIRCtidalconstituentsandIOSprogram. Simulated(ADCIRC) Measured(IOS) Dierence(ADCIRC-IOS) Constituent Amplitude 9.1 191.2 8.4 176.4 0.7 14.8 O1 6.5 201.3 6.1 207.5 0.4 -6.2 Q1 1.1 195.8 1.2 240.3 -0.1 -44.5 M2 47.0 24.4 47.0 29.1 0.0 -4.7 N2 10.1 9.2 12.1 20.6 -2.0 -11.2 S2 10.3 41.9 10.4 41.7 -0.1 0.2 K2 1.9 29.0 N/A N/A N/A N/A TheseresultsshowthattherearesomediscrepanciesbetweenconstituentsbasedonmeasuredwaterelevationandADCIRCconstituents.ThesediscrepanciesaremorepronouncedattheBeaufortlocation.Thatisconsistentwithourprelimi-narysimulationsofHurricaneIsabelwhensimulatedwaterelevationwasnoticeablyoutofphaseatBeaufort. InordertoimproveHurricaneIsabelsimulationresults,theADCIRCtidalconstituentsusedintheCH3Dmodelalongtheopenboundarieswereadjustedaccordingtotheamplitudeandphasedierencesshownabove.TheDuckPier

PAGE 127

Table5{5: TidalconstituentparametersatBeaufort,NCcalculatedbasedonAD-CIRCtidalconstituentsandIOSprogram. Simulated(ADCIRC) Measured(IOS) Dierence(ADCIRC-IOS) Constituent Amplitude 9.4 193.4 6.7 210.7 2.7 -17.3 O1 7.0 205.5 6.6 221.5 0.4 -16.0 Q1 1.2 197.2 1.0 216.4 0.2 -19.2 M2 43.9 17.2 45.3 37.2 -1.4 -20.0 N2 10.3 2.0 10.8 28.5 -0.5 -26.5 S2 9.3 32.0 8.9 48.3 0.4 -16.3 K2 1.7 17.3 N/A N/A N/A N/A adjustmentwasappliedtotheentireeasternopenboundaryandtheBeaufortadjustmentwasappliedtothesouthernopenboundary.Asaresult,thesimulatedtidewasinmuchbetteragreementwithmeasuredtide,whichwasveryimportant. Windisamajorforcedrivingastormsurge.Sowhenitcomestousingwindinastormsurgemodel,itisveryimportanttovalidatethewindbecauseitsaccuracywillbeasignicantfactorintheoverallaccuracyofmodel'soutput.Twoanalysiswindeldswereusedtodrivethemodel:WINDGENandWNA(refertoTable 2{1 )formoreinformationonthesewindelds). Figures 5{5 through 5{6 showcomparisonbetweenmeasuredwindspeedanddirectionattheCapeLookoutandDuckPierstations,andwindspeedanddirectionobtainedfromWINDGENandWNAwindelddatasetsduringHurricaneIsabel.Thecomparisonforallwindstationswithinthecomputationaldomain,listedinTable 5{2 ,isshowninAppendix D . AthorougherroranalysisofWNAandWINDGENwindspeedandwinddirectionisshowninTable 5{6 .FormulasusedtocalculatetheerrorscanbefoundinAppendix B .

PAGE 128

Figure5{5: WINDGENandWNAvs.measuredwindspeedanddirectionatCapeLookout,NCduringHurricaneIsabel. Figure5{6: WINDGENandWNAvs.measuredwindspeedanddirectionatDuckPier,NCduringHurricaneIsabel.

PAGE 129

Table5{6:ErrorsofWNAandWINDGENwindspeedanddirectioncomparedwithmeasuredatwindstationsduringHurricaneIsabel. WNA WNA WINDGEN WINDGEN speed(m/s) dir(deg) speed(m/s) dir(deg) CapeLookout RMSerror 2.20 18.25 2.48 20.09 Meanabsoluteerror 1.56 8.95 1.96 13.79 Maxabsoluteerror 8.14 128.72 8.08 129.38 onJulianDay 261.860 261.740 261.650 261.740 Duck RMSerror 2.00 6.28 2.56 13.22 Meanabsoluteerror 1.50 4.95 2.15 11.34 Maxabsoluteerror 6.18 13.71 5.74 28.30 onJulianDay 261.890 261.110 260.330 260.750 ChesapeakeLight RMSerror 3.45 8.98 6.02 15.90 Meanabsoluteerror 2.80 8.16 5.34 14.33 Maxabsoluteerror 10.36 16.39 13.65 31.24 onJulianDay 262.220 261.860 261.950 260.780 ChesapeakeBayBridge RMSerror 3.27 7.64 4.33 8.88 Meanabsoluteerror 2.00 5.72 3.72 7.07 Maxabsoluteerror 12.52 20.88 11.98 25.32 onJulianDay 261.980 261.860 262.040 260.150 Kiptopeke RMSerror 6.70 11.11 3.86 14.32 Meanabsoluteerror 5.72 9.62 3.06 11.05 Maxabsoluteerror 14.34 25.80 9.10 33.10 onJulianDay 261.830 262.580 261.710 262.580 GloucesterPoint RMSerror 6.20 69.18 4.09 71.21 Meanabsoluteerror 5.58 56.46 3.29 56.10 Maxabsoluteerror 12.51 173.00 10.97 174.20 onJulianDay 261.620 260.390 262.160 260.390 MoneyPoint RMSerror 9.08 17.65 6.41 17.64 Meanabsoluteerror 8.26 14.72 5.60 13.98 Maxabsoluteerror 19.23 70.77 14.36 76.06 onJulianDay 261.860 260.180 261.830 260.180 HPLWS RMSerror 6.17 31.23 4.93 31.24 Continuedonnextpage

PAGE 130

WNA WINDGEN WINDGEN speed(m/s) dir(deg) speed(m/s) dir(deg) Meanabsoluteerror 5.90 29.17 4.42 28.92 Maxabsoluteerror 10.91 63.70 8.89 57.60 onJulianDay 261.650 260.000 262.130 260.00 ChoptankRiver RMSerror 2.63 28.60 2.54 30.90 Meanabsoluteerror 2.25 24.28 2.00 25.96 Maxabsoluteerror 5.77 90.62 7.64 112.02 onJulianDay 261.890 262.550 262.130 262.550 Lewisetta RMSerror 6.03 14.38 3.35 12.69 Meanabsoluteerror 5.34 12.23 2.77 9.46 Maxabsoluteerror 12.64 44.60 6.46 45.50 onJulianDay 261.620 262.730 261.530 260.000 NorthBay RMSerror 8.64 45.90 5.91 49.02 Meanabsoluteerror 7.81 34.27 5.15 34.80 Maxabsoluteerror 15.89 118.14 12.79 134.64 onJulianDay 261.770 262.070 261.770 262.070 BothwinddatasetscomparewellwiththedataovertheOuterBanksandnearthemouthoftheChesapeakeBay,withtheWNAwindbeingslightlymoreaccurate.InsidetheChesapeakeBay,theWINDGENwindismoreaccuratethanWNAbuttheoverallaccuracyismuchworsewhencomparedtothewinddataovertheOuterBanks.Evidently,boththeWNAandWINDGENwindmodelsdonotperformwelloverthelandandtheChesapeakeBayareaastheydoovertheopenwater.ThisisdisappointingbecausesomeofthewaterelevationstationsliewithintheareawheretheWNAandWINDGENwindswerenotadequate,andthusitwouldbehardtojudgethequalityoftheperformanceofthestormsurgemodelbasedonthepoorwindaccuracy. Waveboundaryconditionswereobtainedfromtheregionalwavemodel,WAVEWATCH-III.Figures 5{7 and 5{8 showacomparisonbetweensignicantwaveheightandpeakwaveperiodobtainedfromtheWAVEWATCH-IIImodel

PAGE 131

Figure5{7: SignicantwaveheightandpeakwaveperiodobtainedfromWAVEWATCH-IIIcomparedwithmeasuredwaveheightatNDBCstation41001. Figure5{8: SignicantwaveheightandpeakwaveperiodobtainedfromWAVEWATCH-IIIcomparedwithmeasuredwaveheightatNDBCstation41002. andmeasuredatNDBCstations41001and41002,acouplehundredkilometersothecoastofNorthCarolinaduringthemonthofSeptember,2003.ItisinterestingtonotethattheHurricaneIsabeltrackpassedrightbetweenthetwowavebuoys.Theseguresshowthatthesimulatedandmeasuredwaveparametersareingoodagreement.Though,asshowninFigure 5{8 ,the41002signicantwaveheightisunderestimatedrightbeforethewaveheightpeakassociatedwithHurricaneIsabelnearJulianDay260.ThiswillhavesomeeectonwaveheightcalculatedbytheSWANmodelinsideourlocalcomputationaldomain.TheeectisdiscussedinSection 5.1.5 ofthischapter.

PAGE 132

4.4 forwavesetupvalidation).Thus,inordertoobtainaccurateestimateonwavesetup,itisessentialtohaveaccuratesimulationofnearshorewaveelds.TheSWANwavemodel(seeSection 2.3.5 )wasusedtosimulatewaveeldsduringHurricaneIsabel. Threesetsofwavedatawereavailableforcomparisonwithsimulatedwaveresults.TwodatasetscamefromtheFieldResearchFacilities(FRF)atDuck,NC(seeFigure 4{29 fortheentireFRFinstrumentsetup);andthethirdsetwasprovidedbyVirginiaInstituteofMarineScience(VIMS)whichmeasuredwavesatGloucesterPoint,VA(seeFigure 5{9 forthelocationoftheVIMSinstrumentpackage).First,letuscomparethesimulatedandmeasuredwaveparametersatthe Figure5{9: LocationoftheVIMSinstrumentpackageatGloucesterPoint,VA.

PAGE 133

FRFWaveriderapproximately4kmoshore(thesedatawillbefurtherreferredtoastheDuck630data).Thedepthatthelocationis17m.ThemaximumsignicantwaveheightmeasuredattheFRFWaveriderbuoyduringHurricaneIsabelwas8.1m,whilethelargestwave(cresttotrough)recordedonSeptember18at19:11UTC,was12.1m.Theratioofthelargestwavetolocalwaterdepth,12:1=17=0:71,isclosetotherangewhenwavesmightstarttobreak.The\simulatedvs.measured"signicantwaveheight,peakwaveperiodandwavedirectionareshowninFigures 5{10 , 5{11 ,and 5{12 ,respectively.ItshouldbepointedoutthatallsimulatedresultspresentedinthissectionwereobtainedusingtheWNAwind,sincetheWINDGENwindforHurricaneIsabelwasslightlylessaccurate.Thesimulatedwaveheightmatcheswellwiththemeasureddatawith Figure5{10: Simulatedsignicantwaveheightvs.measuredfromtheFRFWa-veriderbuoyduringHurricaneIsabel. slightunderestimationatthepeak.ThereisalsoaphaselagrightbeforethepeakwhichmightbeduetoswellwavesgeneratedoutsideofthecomputationaldomainwhichcouldnotbeproperlysimulatedbytheSWANmodelbecause,aswasdiscussedinSection 5.1.4 ,thewaveheightboundaryconditionsobtainedfrom

PAGE 134

theregionalWAVEWATCH-IIImodelwereslightlyunderestimatedrightbeforeHurricaneIsabelpassedoverthearea. Theconsequenceofsuchwaveheightunderestimationwillmostlikelyresultinlowerthanexpectedwavesetuprightbeforethepeakofthestorm.Despitethat,however,thepeakwaveheightvaluesmatchwell,whichshouldresultinanaccuratecalculatedwavesetupcontributiontotheoverallpeakstormsurgelevel.Thecalculatedpeakwaveperiodmatchesthemeasuredpeakperiodverywell,so Figure5{11: Simulatedpeakwaveperiodvs.measuredfromtheFRFWaveriderbuoyduringHurricaneIsabel. doesthecalculatedwavedirection,especiallyduringthepeakofthestormwhenthedierencewaslessthan5degrees,anexcellentagreement. Wavedirectionisanimportantfactorincalculationofwavesetup.Waveraysapproachingtheshoreperpendicularlywillcausehigherwavesetupasopposedtothecasewhentheyapproachtheshoreinanobliquemanner( ShengandAlymov , 2002 ).Wavedirectioncanalsoinuencetheestimationofcross-shoreandlongshorecurrents.Figure 5{13 showsresultsobtainedfromatestcasesimulationofthe StiveandWind ( 1982 )laboratoryexperiment,withashorelineonthe

PAGE 135

rightandincidentwavesontheleftsidewhereHsig=0.226mandT=1.79sec.Thetoppanelplotshowsthecalculatedwavesetupandcurrentscausedbywavesapproachingtheshoreata45oangle(fromsouth-westtonorth-east).Thecalculatedwavesetupandcurrentscausedbywavesapproachingtheshoreat-45o(fromnorth-westtosouth-east)areshowninthebottompanelplot.Northerlyandsoutherlylongshorecurrentsareformedinbothplots.Sincenootherforcingmechanism(windortide)wasconsideredinthissimulation,currentsaregeneratedbywaveactiononlythroughradiationstressesandwavesetup. Figure5{12: Simulatedwavedirectionvs.measuredfromtheFRFWaveriderbuoyduringHurricaneIsabel. Now,letuscomparesimulatedandmeasuredwaveparametersattheendoftheFRFpierapproximately600moshore(thesedatawillbereferredtoastheDuck625data).Thedepthatthislocationis8.4mandthemaximummeasuredsignicantwaveheightduringHurricaneIsabelwas3.7m.The\simulatedvs.measured"signicantwaveheightandpeakwaveperiodcomparisonsareshowninFigure 5{14 .Therewasnorecordofwavedirectionatthatlocation.ThetrendinsimulatedsignicantwaveheightattheFRFpierissimilartothatattheDuck

PAGE 136

Figure5{13: Atestcase:wavesetupandcurrentsinducedbywavesapproach-ingtheshorefromsouth-westtonorth-east(toppanel),andfromnorth-westtosouth-east(bottompanel).

PAGE 137

Figure5{14: Simulatedsignicantwaveheightandpeakwaveperiodvs.measuredfromtheFRFpierduringHurricaneIsabel. 630location.Thepeakvaluesmatchwellbutthesimulatedwaveheightdidnotcapturetheincreaseofthemeasureddatawaveheightapproximatelyonedaybeforetheactualstormcamein.ThisismorelikelyduetotheinabilityoftheSWANmodeltoaccountforlargeswellwavesthatarecomingfromoutsideofthecomputationaldomain. Aswaspointedoutearlierinthissection,theunderestimationofwaveheightbeforethepeakofthestormwillresultinlowerthanexpectedwavesetupduringthattime.Butduetothefactthatthepeakwaveheightvaluesmatchwell,thecontributionofthecalculatedwavesetuptotheoverallstormsurgeduringitspeakattheendoftheFRFpiershouldbeaccurate.Thecalculatedpeakwaveperiodisagainingoodagreementwiththeobservedone. Now,letuscomparesimulatedandmeasuredwaveparametersatGloucesterPoint,VA(thesedatawillbereferredtoastheVIMSdata).Thedepthatthelocationisaround8.5m,whilethemaximummeasuredsignicantwaveheightduringHurricaneIsabelwas1.7m.The\simulatedvs.measured"signicantwaveheightandpeakwaveperiodcomparisonsareshowninFigure 5{15 .Therewasnorecordofwavedirectionatthatlocation.Thecalculatedsignicantwaveheightisslightlyoverestimatedbasedoncomparisonwithmeasuredwaveheight,although

PAGE 138

Figure5{15: Simulatedsignicantwaveheightandpeakwaveperiodvs.measuredatVIMSduringHurricaneIsabel. thepeaksagreewell.ThemainreasonbehindthisisthestrongWNAwindwhichinrealitywasslightlyweaker(seeFigure D{7 inAppendix D ).TheresultoftheslightwaveheightoverestimationatVIMSwillbeaslightlyhigherthanexpectedwavesetupinthatlocation.Thecalculatedpeakwaveperiodagreeswellwithobserved. 5{16 through 5{22 .AlltheothersimulatedversusmeasuredwaterelevationresultsduringHurricaneIsabelareshowninAppendix F . ItcanbenotedthatthepeakvaluesofthecalculatedwatersurfaceelevationmatchthemeasuredwaterelevationwellatDuck,ChesapeakeBayBridge,andGloucesterPoint.ThesimulatedwindattheOuterBanksandlowerChesapeakeBayagreewellwithdataand,asaresult,thesimulatedwaterelevationsagreewell

PAGE 139

Figure5{16: Comparisonofsimulatedvs.measuredwaterelevationatBeaufort,NC.Twosimulatedresultsareshown:oneusingWNAwindandtheotherusingWINDGENwind. Figure5{17: Comparisonofsimulatedvs.measuredwaterelevationatDuck,NC.Twosimulatedresultsareshown:oneusingWNAwindandtheotherusingWINDGENwind.

PAGE 140

Figure5{18: Comparisonofsimulatedvs.measuredwaterelevationatChesapeakeBayBridge,VA.Twosimulatedresultsareshown:oneusingWNAwindandtheotherusingWINDGENwind. Figure5{19: Comparisonofsimulatedvs.measuredwaterelevationatGloucesterPoint,VA.Twosimulatedresultsareshown:oneusingWNAwindandtheotherusingWINDGENwind.

PAGE 141

Figure5{20: Comparisonofsimulatedvs.measuredwaterelevationatMoneyPoint,VA.Twosimulatedresultsareshown:oneusingWNAwindandtheotherusingWINDGENwind. Figure5{21: Comparisonofsimulatedvs.measuredwaterelevationatKiptopeke,VA.Twosimulatedresultsareshown:oneusingWNAwindandtheotherusingWINDGENwind.

PAGE 142

Figure5{22: Comparisonofsimulatedvs.measuredwaterelevationatLewisetta,VA.Twosimulatedresultsareshown:oneusingWNAwindandtheotherusingWINDGENwind. withmeasureddata.Rightbeforethepeak,aslightunderestimationcanbeob-servedatallthestations.Aswaspointedoutintheprevioussection,thecomputedwaveheightbeforethepeakwasunderestimatedduetotheunderestimatedwaveboundaryconditionsprovidedtoSWANbytheregionalWAVEWATCH-IIImodel.Thisunderestimationresultedinlowerthanexpectedwavesetuprightbeforethepeakofthestorm.Thewaveheightatthepeakofthestormwasaccuratelysimulated,whichresultedinadequatecontributionofcalculatedwavesetuptothesimulatedwaterlevelatthattime. Maximumsimulatedwaterelevations(includingtide,surgeandwavesetup,andrelativetoNAVD88)intheOuterBanksandChesapeakeBayduringIsabel,calculatedusingWNAwind,areshowninFigure 5{23 .HighwaterelevationscanbeseennotonlyintheareawhereIsabelmadelandfallbutalsothroughouttheentireOuterBanksandlowerChesapeakeBay.Thehighestvaluereached4.0matthetipoftheOuterBanksnearBuxton,NC.Thecomputedwaterlevelreached3.2

PAGE 143

mjustsouthofPamlicoSoundneartheupperSouthRiver,NC.InVirginia,thehighestcalculatedwaterlevelof3.5mcanbeobservedinCobbBaynearCheritonandButlerCreeknearBraysLanding.ThemaximumwaterlevelintheupperJamesRiver,VAreached3.0m. Figure 5{24 showsthemaximumwavesetupelevationcalculatedduringIsabelusingWNAwind.Ascanbeseeninthisgure,wavesetupreached1.2mandoodedpartsofthechainofemergentbarrierislandsinthesouthernOuterBanks.Thisdemonstratesthatwavesetupcanbeasignicantfactorandcontributortostormsurgelevelandinundation. Figure 5{25 showsthestormsurgeatallsevenstationswithtidessubtractedfromthesimulatedwaterlevelwhichincludesallforcingmechanisms(tide,wind,wavesetup,waveenhancedsurfaceandbottomfriction). Thecalculatedstormsurgeatthesevenstationsshowsawiderange:from1.3mnearDuck,NCto2.0mnearGloucesterPoint,VA. Itshouldbenotedthatduringthepeakofthesurge,thetidewasatitshighstage.Infact,atDuck,NCthetidewasatitshighestlevelof0.45m.BythetimeIsabelapproachedChesapeakeBay,thetidehadalreadypasseditspeakandstartedtorecede. PlotsshowingcomparisonbetweenthesimulatedstormsurgeelevationandmeasuredstormsurgeelevationcanbefoundinAppendix F . Throughoutthemodeldomain,ittook19to26hoursfromthepointwhenthesurgelevelstartedtorisetothepointwhenthesurgeretreated,withshorterperiodsoccurredovertheOuterBanksandlongeronesinthelowerChesapeakeBay.Forexample,althoughDuck,NCandGloucesterPoint,VAstartedtoexperiencetheincomingstormsurgeatapproximatelythesametime,evenafterithadrisentoitsmaximumlevelatDuck,thesurgekeptonrisingatGloucesterPoint.

PAGE 144

Figure5{23: MaximumwaterelevationrelativetoNAVD88(includestide,surgeandwavesetup)calculatedduringsimulationofHurricaneIsabelintheOuterBanks/ChesapeakeBayusingWNAwind.

PAGE 145

MaximumwavesetupelevationrelativetoNAVD88calculatedduringsimulationofHurricaneIsabelinthesouthernpartofOuterBanksusingWNAwind. Figure5{25: Simulatedstormsurge(waterlevelminustide)atthesevenstationsthroughouttheOuterBanks/ChesapeakeBayusingWNAwind.

PAGE 146

Ithastobepointedoutthatmeasuredwaterelevationconsistsoftwoparts:tideandstormsurgeitself.Theoretically,thestormsurgepartcanalsobesplitintotwoconstituents,oneduetowindactionandtheotherduetowavesetup.Butpracticallyitisimpossibletolteroutthewavesetupfromthestormsurgeandthereforeitwouldbereasonableifweleftthestormsurgeelevationintact.Tide,ontheotherhand,canbelteredoutusingthe DoodsonandWarburg ( 1941 )39-hourlyaveragetidallterasdescribedby Groves ( 1955 ). Anerrorbetweenmeasuredandcalculatedwaterelevationcanbeattributedtoeithertideorstormsurgeortheircombination.Inordertoestimatethecontributionofeachsourceoferror,a\pure"tidesimulationwasperformedanditsresultswerecomparedwithtidalelevationwhichwaslteredoutfrommeasuredwaterelevationusing DoodsonandWarburg ( 1941 )39-hourlyaveragetidallter.\Pure"tidesimulationmeansthattheonlyboundaryforceimposedduringthatsimulationwastideandallotherforcingmechanismssuchaswindandwavewerenotconsidered. Inordertoweightheeectofeachcomponentinvolvedinthenon-linearinteractionbetweenthesurge,tide,wind,andwave,severalsimulationsweremadebyincludingdierentcomponentcombinations.Turningthewetting-and-dryingfeatureonandowasanoptionaswell.Table 5{7 speciesthesixspecicfeaturesincludedinvesimulations.

PAGE 147

Table5{7:Alistofsimulationswithvariouscombi-nationsofsixmodelfeatures(psymboldenotesthefeaturewasincludedduringthesimulation). Factors Sim1 Sim2 Sim3 Sim4a Tide p p p p p p p p p p p p p p p p p p p p p Table 5{8 showstheRMS,MeanAbsolute,andMaximumAbsoluteerrors(seeAppendix B fordenitions)ofcalculatedwaterelevationduringHurricaneIsabel.Errorsofpeakvalues(measuredpeakelevationminussimulatedpeakelevation)and\timing"errors(thetimewhenmeasuredpeakelevationoccurredminusthetimewhensimulatedpeakelevationoccurred)arealsoshown.Aseparatecolumninthetabledisplaystheerrorsattributedto\pure"tide.Tidalrangeisalsoshownforeachstation. ( 1993 )formulationwasused2 ( 1979 )formulationwasused3 ( 1989 )formulationwasused

PAGE 148

Table5{8:ErrorsofwaterelevationattidestationsduringHurricaneIsabel.Themodelresultsaredatawerecomparedevery15minutes. WNAwind WINDGENwind Tide Sim1 Sim2 Sim3 Sim4a Sim1 Sim2 Sim3 Sim4a Station:Beaufort(depth=4.0m;tidalrange=115cm) RMSError(cm) 3.8 24.8 18.0 18.3 19.4 19.3 19.0 26.2 20.1 20.8 21.8 21.9 20.4 MeanAbsError(cm) 3.1 15.1 11.1 11.4 13.5 12.4 12.4 16.4 12.4 12.9 14.8 13.9 13.3 MaxAbsError(cm) 8.9 72.0 61.9 61.5 66.1 66.5 68.3 77.9 64.6 61.8 62.4 62.4 66.7 Meas.SurgePeak(cm) 107 ErroratPeak(cm) 55 37 35 41 40 26 54 42 35 41 47 38 TimingatPeak(min) 1 0 16 -14 7 19 -30 -26 -32 -42 -32 5 Station:Duck(depth=5.8m;tidalrange=140cm) RMSError(cm) 3.6 13.1 10.5 10.0 10.1 10.0 10.2 15.8 11.1 9.9 10.0 9.5 11.0 MeanAbsError(cm) 2.8 9.5 7.8 7.5 7.6 7.5 7.5 10.8 8.4 7.6 7.7 7.3 8.5 MaxAbsError(cm) 11.0 42.7 36.4 34.6 34.3 34.5 35.6 51.7 32.4 27.3 27.5 27.1 30.6 Meas.SurgePeak(cm) 171 ErroratPeak(cm) 27 10 -1 0 0 1 36 19 11 11 9 15 TimingatPeak(min) 66 46 -16 -16 -16 71 59 45 45 45 45 43 Station:ChesapeakeBayBridge(depth=10.6m;tidalrange=110cm) RMSError(cm) 4.1 13.1 11.2 8.9 9.8 9.1 9.7 18.1 14.6 12.9 13.7 13.1 13.0 MeanAbsError(cm) 3.4 9.1 7.7 6.7 7.2 6.8 7.1 11.9 9.8 8.9 9.6 9.0 8.5 MaxAbsError(cm) 9.4 45.0 38.3 27.5 32.9 28.3 32.0 51.5 49.0 42.3 45.2 42.5 30.6 Meas.SurgePeak(cm) 168 ErroratPeak(cm) 34 33 20 27 21 19 35 35 24 31 29 27 TimingatPeak(min) 9 1 0 0 0 6 14 10 16 10 10 12 Continuedonnextpage

PAGE 149

WINDGENwind Tide Sim1 Sim2 Sim3 Sim4a Sim1 Sim2 Sim3 Sim4a Station:GloucesterPoint(depth=8.5m;tidalrange=80) RMSError(cm) 5.7 13.1 11.2 10.5 11.0 10.2 10.0 22.3 19.3 15.1 18.4 16.3 15.3 MeanAbsError(cm) 4.5 9.9 8.5 7.9 8.2 7.9 7.3 14.6 13.1 11.1 13.1 11.9 10.5 MaxAbsError(cm) 16.1 42.2 40.7 32.7 35.3 28.9 41.3 71.5 63.4 45.0 57.2 50.8 52.1 Meas.SurgePeak(cm) 199 ErroratPeak(cm) 30 25 -10 13 -6 -23 58 60 42 57 46 44 TimingatPeak(min) -118 -104 -143 -120 -115 -131 39 16 24 -6 3 36 Station:MoneyPoint(depth=13.1m;tidalrange=100cm) RMSError(cm) 9.8 22.2 20.7 26.1 24.8 25.2 16.3 31.7 25.5 31.5 27.5 29.4 23.7 MeanAbsError(cm) 8.0 14.1 13.5 16.1 14.8 16.3 10.3 19.3 16.7 17.8 17.5 17.4 15.4 MaxAbsError(cm) 21.4 115.1 106.7 145.0 131.7 136.3 93.1 127.4 108.8 151.9 115.3 133.0 101.7 Meas.SurgePeak(cm) 192 ErroratPeak(cm) 23 31 20 32 24 14 46 36 25 36 29 32 TimingatPeak(min) 92 53 42 56 45 78 33 26 26 26 10 36 Station:Lewisetta(depth=3.0m;tidalrange=45cm) RMSError(cm) 5.6 19.8 16.9 14.7 19.7 15.5 16.1 19.6 18.3 16.6 21.3 17.8 15.5 MeanAbsError(cm) 4.7 14.2 12.3 10.7 14.0 11.4 11.9 14.2 13.3 12.3 15.3 13.3 11.2 MaxAbsError(cm) 13.3 65.2 56.1 50.6 64.6 52.6 43.7 69.1 67.7 64.5 73.5 65.2 55.7 Meas.SurgePeak(cm) 131 ErroratPeak(cm) 55 48 43 60 46 13 53 47 38 52 42 33 TimingatPeak(min) -154 -198 -198 -142 -168 -259 -60 -115 -100 -118 -122 -113 Station:Kiptopeke(depth=2.4m;tidalrange=95cm) RMSError(cm) 5.4 17.7 15.4 16.7 19.4 17.5 12.4 16.9 13.4 13.0 15.4 14.0 12.1 MeanAbsError(cm) 4.8 11.5 9.8 10.2 12.4 10.8 9.6 11.7 9.5 9.3 11.2 10.2 9.6 Continuedonnextpage

PAGE 150

WINDGENwind Tide Sim1 Sim2 Sim3 Sim4a Sim1 Sim2 Sim3 Sim4a MaxAbsError(cm) 11.6 64.5 57.8 64.1 74.0 67.1 40.0 56.7 50.0 48.7 56.2 50.9 41.6 Meas.SurgePeak(cm) 138 ErroratPeak(cm) 58 54 55 66 58 31 46 44 42 52 45 32 TimingatPeak(min) 60 60 60 60 60 -185 72 72 72 58 71 69 AvgRMSError(cm) 5.4 17.7 14.8 15.0 16.3 15.3 13.4 21.5 17.5 17.1 18.3 17.4 15.9 AvgErr.atPeak(cm) 40 34 26 34 28 19 47 40 31 40 35 32 AvgTimingError(min) 71 66 72 58 59 107 44 44 45 44 42 45

PAGE 151

Basedonthiserroranalysis,especiallyontheaverageRMSerrorsandaverageabsoluteerrorsatthepeakwaterelevation,itcanbeconcludedthattheWNAwindproducedslightlybettersurgesimulationresultsthantheWINDGENwind.ThisisconsistentwiththeoverallanalysisofwinddatashowninTable 5{6 ,whichjustiestheimportanceofwindaccuracy:moreaccuratewindproducesmoreaccuratewaterelevationresults. TheaccuracyofthesimulatedwaterelevationdependedontheaccuratesimulationoftidewhichwasaccurateacrosstheOuterBanksuptothemouthoftheChesapeakeBay,withtheaverageRMSerrorofapproximately4cm.Insidethebay,theaccuracyofthecalculatedtideworsened,withtheaverageRMSerrorincreasingto6cm.Sodidtheaccuracyofthesimulatedwaterelevation,whichwasalsoaccompaniedwiththeworseaccuracyoftheWNAwindinsidetheChesapeakeBayasopposedtothatovertheOuterBanks. Overall,Simulation3producedbetterresultsintermsofsmallerRMSerrorsandbettercomparisonwithmeasuredwatersurfaceelevationatitspeak.Waterlevelcalculatedusingwaveenhancedbottomfrictionbasedonthe ShengandVillaret ( 1989 )formulation(Simulation4b)wasslightlyworse.Simulation4awhichaccountedforwaveenhancedbottomfriction( GrantandMadsen ( 1979 )formulation)wasevenworseduetooverestimatedbottomfriction.Simulation5whichdidnotaccountforwetting-and-dryingimprovedtheresultsinsideChesapeakeBay,buttheonlyreasonforthisimprovementwasthatthecalculatedwaterelevationwassignicantlyunderestimatedthere,sowhentheoodingwasturnedo,thewaterelevationwasabletogain20-30cmmore,thusmakingitlookbetter. The\timing"errorsareverysmall(basedonSimulation4busingWNAwind,0to16min)forthethreeOuterBankstations(Beaufort,Duck,andChesapeakeBayBridge).FortheotherfourstationsinsideChesapeakeBaytheerrorsaremuch

PAGE 152

larger(45mintoalmost3hours).ThisisconsistentwiththefactthattheWNAwindwasmoreaccurateintheOuterBanksareacomparedwiththewindintheChesapeakeBayarea.Also,moreaccurateovertheOuterBanksareaWNAwindproducedsmaller\timing"errorscomparedwithlessaccurateWINDGENwindinthatarea.AndmoreaccurateovertheChesapeakeBayWINDGENwindproducedsmaller\timing"errorscomparedwithlessaccurateWNAwindinthatarea. Table 5{9 showspeakwaterelevationvaluescalculatedduringthesesimu-lationsalongwiththemeasuredvalue.Thetablealsodisplayspercentincreaseordecreaserelativetothesimulationintheprevioussimulationcolumn(e.g.,atGloucesterPoint,theSimulation2value"3%relativetotheSimulation1value,andtheSimulation4aand4bvalues#12%and#2%relativetotheSimulation3value,respectively).Percentvalueforeachstationwasnormalizedbythemeasuredvalueatthisstation. Includingwaveandradiationstressterms(Simulation2)increasedthecalculatedwaterlevelby1-17%.MoresignicantwavesetupeectwasobservedalongtheOuterBanksandlesssignicantinsideChesapeakeBay.Thisisinaccordancewiththewaveheightdistributionoverthearea:largerbreakingwavesresultedinhigherwavesetup.IntheChesapeakeBay,waveswerenotashighasneartheOuterBanks,thusthewavesetupwaslower. Addingwaveenhancedsurfacestresshelpedfurtherincrease(inmostcases)thecomputedwaterelevationby5-18%.AtBeaufort,theincreasewasnotverysignicant(2%),andatKiptopeke,a1%decreasewasobserved.Thishappenedbecauseduringthepeakofthestorm,thewindwasblowingprimarilyoshore.Wavesincreasedthestressthuspushingmorewateroshoreandcreatingaslightset-down.Accountingforwaveenhancedbottomfrictionusingthe GrantandMadsen ( 1979 )modeldecreasedthecalculatedwaterelevationby1-14%.Theimportanceofthewaveenhancedbottomfrictionwasmoresignicantinside

PAGE 153

shallowerChesapeakeBay.Itisknownthatthe GrantandMadsen ( 1979 )modeltendstooverestimatethewaveenhancedbottomstress( Sheng , 1982 ),althoughtherewasinsucientdataduringHurricaneIsabeltoconrmthis.Toimprovetheuncertaintyofthewave-enhancedbottomstress,aone-dimensionalturbulentboundarylayermodelwithTKEclosure( ShengandVillaret , 1989 )wasusedtodevelopa\look-uptable"(seeSection 2.3.4.3 )forcalculatingwaveenhancedbottomstressinawiderangeofcombinedwave-currentboundarylayerows. Whenwetting-and-dryingschemewasnotactivatedduringthecalculation,thepeakwaterlevelvalue,ingeneral,grewanextra2-20%,althoughtheelevationactuallydroppednearDuck.AfteranalyzingtheoodmapobtainedduringSimulation4bwhenwetting-and-dryingwasactivated,itturnedoutthatthispartofthegridhadhighenoughtopographicelevationsoitdidnotgetoodedduringIsabel.Therefore,whenwetting-and-dryingwasdisabled,thecalculatedwaterelevationwouldnotbeaectedmuchanyway.Whydidtheelevationactuallydrop?IthappenedbecausewhentheCH3Dmodelrunsin\nowetting-and-drying"mode,aminimumdepthhastobespecied.DuringSimulation5,thisdepthwassetto10m,meaningthatanygridcellwhosedepthwaslessthan10mwasforcedtobe10m.Normally,theminimumdepthshouldbesetaslowaspossibletoreducethenumberofcellstobeaectedbythecut-odepth.Pickingasmallvaluemightcausethemodeltobecomeunstable,especiallyunderseverewindconditions,andblow-upwhichdidhappenwhentheminimumdepthwassettolessthan10m.TheactualdepthatDuckislessthan6m,soduringSimulation5,itwassetto10m.Suchdeepeningofthebathymetrycausedaslightdecreaseinthecalculatedwatersurfaceelevation.Thiscasebringsoutaproblemrelatedtostormsurgemodelswithoutwetting-and-drying.Mostsuchmodels,includingADCIRC,haveacut-odepth.Thismayresultininaccuratestormsurgesimulations.

PAGE 154

Table5{9:MeasuredpeakwaterelevationsatsevenstationsduringHurricaneIsabelusingWNAwindandvariouscombinationsofstormsurgemodelfeatures. Sim1 Sim2 Sim3 Sim4a Meas. cm cm cm cm cm cm cm Beaufort 52 70"17% 72"2% 66#6% 67#5% 81"13% 107 Duck 144 161"10% 172"6% 171#1% 171#1% 166#3% 171 Ches.BayBr. 134 135"1% 148"8% 141#4% 147#1% 149"1% 168 GloucesterPt. 169 174"3% 209"18% 186#12% 205#2% 208"10% 199 MoneyPoint 169 161#4% 172"6% 160#6% 168#2% 178"5% 192 Kiptopeke 80 84"3% 83#1% 72#8% 80#2% 99"14% 138 Lewisetta 76 83"5% 89"5% 71#14% 85#3% 96"8% 131

PAGE 155

Table5{10:Calculatedpeakstormsurge(withtidessubtracted)atsevenstationsduringHurricaneIsabelusingWNAwindandvariouscombinationsofstormsurgemodelfeatures. Sim1 Sim2 Sim3 Sim4a Meas. cm cm cm cm cm cm cm Beaufort 19 34"18% 38"5% 39"1% 39"1% 38#1% 83 Duck 103 120"13% 131"8% 1310% 1310% 126#4% 131 Ches.BayBr. 97 111"10% 118"5% 115#2% 117#1% 126"6% 142 GloucesterPt. 165 176"6% 203"16% 183#12% 195#5% 200"3% 170 MoneyPoint 127 132"3% 133"1% 127#4% 130#2% 146"9% 171 Kiptopeke 81 96"14% 98"2% 88#9% 94#4% 109"14% 108 Lewisetta 80 89"8% 95"5% 76#16% 91#3% 105"12% 118

PAGE 156

Table 5{10 showspeakstormsurgevalues(tidewassubtractedfromwaterelevation)computedduringthesesimulations.Percentincreaseordecreasewascalculatedthesamewayitwascalculatedintheprevioustable. Theresultsdemonstratethatincludingwavesetupincreasedandimprovedthecalculatedsurge,especiallyatBeaufortandDuck.Itseectrangedbetween3-18%.Wave-enhancedwindalsomadethecomputedsurgelookbetterbyincreasingthevaluesby1-16%.Thewave-enhancedbottomstresscalculatedusingthe GrantandMadsen ( 1979 )model,mostlymadethesurgelookworsebydecreasingthepeaksurgevaluesby0-16%.Waveenhancedbottomstresscalculatedusingthe ShengandVillaret ( 1989 )formulation(Simulation4b)decreasedthesurgelessdramatically,by0-5%.Whenthewetting-and-dryingfeaturewasinactive,thesurgelevelinsideChesapeakeBaygrewby3-14%. Anotherinterestingfactcanbededucedfromcomparingwaterelevationcal-culatedbylinearsuperpositionofseparatelysimulatedtide,wavesetup,andsurgewithwaterelevationcalculatedthroughthedynamiccoupling.Figure 5{26 showstheindividuallycalculatedtide,wavesetup,andsurge,andtheirlinearsuperposi-tionatDuck.Figure 5{27 showshowthelinearlycoupledwaterelevationatDuckstacksupagainstthewaterelevationwhichwascalculatedthroughdynamiccou-plingofthecirculationandwavemodels.Themeasuredwaterelevationisshownaswell.Similarly,Figure 5{28 showshowlinearlyanddynamicallycoupledresultscompareatGloucesterPoint. Ascanbeseenfromthesegures,atthepeakthelinearlycoupledwaterelevationisonlyslightlyhigherthanthedynamicallycoupledone.Doesthismeanthatthedynamiccouplinginstormsurgemodelingdoesnotbringinasignicantimprovement?Notnecessarily.Thismightbethecasewhenoneisnotconcernedaboutooding,otherwisethemajordierencebetweenlinearanddynamiccouplingcanbeobservedoverinundatedareas.Indeed,Figures 5{29 through 5{31 show

PAGE 157

Figure5{26: Separatelysimulatedtide,wavesetup,andsurge,andtheirlinearsuperpositionatDuck. Figure5{27: Linearlycoupledwaterelevationvs.waterelevationcalculatedthroughdynamiccouplingatDuck,NC.

PAGE 158

Figure5{28: Linearlycoupledwaterelevationvs.waterelevationcalculatedthroughdynamiccouplingatDuck,NC. waterelevationcalculatedlinearlyanddynamicallyatthreeinlandlocationsthatwereoodedduringIsabel:1)southofPamlicoSoundneartheSouthRiver,2)ononeoftheemergentislandsoftheOuterBanks,and3)nearGloucesterinChesapeakeBay.Thetopographicelevations(relativetoNAVD88)atthethreelocationsare1.23m,0.77m,and0.87m,respectively. Ascanbeseenfromthegures,atthebeginningallthreelocationsaredry.Then,inthemiddleofthestorm,theygetinundatedandafterthestormpassestheygetdryagain.Thedierencebetweenwaterelevationscalculatedlinearlyanddynamicallyissignicantduringthepeakofthestorm.Thereasonbehindthatisthatwhentideandwavesetupwerecalculatedseparately,theirelevationwasnothighenoughtooodthoselocations.Thus,onlyseparatelycalculatedsurgewasabletooodthoseareasbutitwassignicantlylowerthantheonecalculatedthroughthedynamiccoupling.Also,whencoupledlinearly,thelocationnearSouthRiverwasooded10hoursaftertheareawasoodedusingthedynamiccoupling.

PAGE 159

Figure5{29: Linearlycoupledwaterelevationvs.waterelevationcalculatedthroughdynamiccouplingneartheSouthRiver,NC.ThelocationisinitiallydryandgetsoodedduringIsabel.Afterthesurgerecedes,itbecomesdryagain. Figure5{30: Linearlycoupledwaterelevationvs.waterelevationcalculatedthroughdynamiccouplingononeoftheemergentislandsoftheOuterBanks,NC.ThelocationisinitiallydryandgetsoodedduringIsabel.Afterthesurgerecedes,itbecomesdryagain.

PAGE 160

Figure5{31: Linearlycoupledwaterelevationvs.waterelevationcalculatedthroughdynamiccouplingnearGloucester,VA.ThelocationisinitiallydryandgetsoodedduringIsabel.Afterthesurgerecedes,itbecomesdryagain. 5{32 through 5{37 showthehighestsimulatedinundationcausedbyHurricaneIsabelduringitspassageovertheOuterBanksandChesapeakeBayusingWNAwind.Theplotsalsoidentifythetimewhenthehighestoodleveloccurred.Similarly,Figures 5{35 through 5{37 showthemaximumsimulatedinundationusingWINDGENwind. Simulation3(seeTable 5{7 )usingWNAwindproducedthebestcomparisonbetweenmeasuredandsimulatedwaterelevationsatallthewaterelevationdatastationsthroughoutthecomputationaldomain.Therefore,thissimulationwasconsideredasthebasesimulationforestimatingtheamountofinundationcausedbyIsabel.Theoodedareaaected7675km2ofland,mostlythesurroundingsoftheCroatan-Albemarle-PamlicoEstuarySystem.

PAGE 161

Figure5{32: MaximumsimulatedinundationinthesouthernpartoftheOuterBanksduringHurricaneIsabelusingWNAwind(toppanel).Thebottompanelshowsthetimeduringwhichthemaximumoodoccurred.

PAGE 162

Figure5{33: MaximumsimulatedinundationintheeasternpartoftheOuterBanksduringHurricaneIsabelusingWNAwind(toppanel).Thebottompanelshowsthetimeduringwhichthemaximumoodoccurred.

PAGE 163

Figure5{34: MaximumsimulatedinundationintheChesapeakeBayduringHurri-caneIsabelusingWNAwind(toppanel).Thebottompanelshowsthetimeduringwhichthemaximumoodoccurred.

PAGE 164

Figure5{35: MaximumsimulatedinundationinthesouthernpartoftheOuterBanksduringHurricaneIsabelusingWINDGENwind(toppanel).Thebottompanelshowsthetimeduringwhichthemaximumoodoccurred.

PAGE 165

Figure5{36: MaximumsimulatedinundationintheeasternpartoftheOuterBanksduringHurricaneIsabelusingWINDGENwind(toppanel).Thebottompanelshowsthetimeduringwhichthemaximumoodoccurred.

PAGE 166

Figure5{37: MaximumsimulatedinundationintheChesapeakeBayduringHurri-caneIsabelusingWINDGENwind(toppanel).Thebottompanelshowsthetimeduringwhichthemaximumoodoccurred.

PAGE 167

Therewasnodirectinundationdataavailableagainstwhichtocompareouroodmaps.TheonlydatathatcouldbefoundtocomparethecalculatedinundatedareasareairphotostakenbeforeandafterIsabelpassedoverthedomain.Figures 5{38 and 5{39 showpre-stormandpost-stormairphotostakenattwolocations,inthesouthernandeasternOuterBanks,respectively(seeFigure 5{2 forlocationinformation).Byexaminingthepre-stormandpost-stormphotos,itcanbeseenthattheseareasweresubjecttoextensiveinundationduringIsabel.Thebottompanelsintheguresdemonstrateclose-upsofourcomputedinundationmapswhichalsodisplaythepresenceofwaterovertheland.Thisvalidationcannotbeconsideredcompleteandisratherqualitativebecausetheactualmeasuredoodlevelisunknownbut,despitethat,itshowsagoodpotentialofthestormsurgemodelingsystemforpredictinginundation.

PAGE 168

Figure5{38: Pre-storm(top)andpost-storm(middle)airphotostakeninthesouthernOuterBanks(courtesyofUSGS,

PAGE 169

Figure5{39: Pre-storm(top)andpost-storm(middle)airphotostakenintheeast-ernOuterBanks(courtesyofUSGS,

PAGE 170

DuringthepassageofHurricaneIsabel,currentprolesweremeasuredattwolocations.TherstlocationisnearKittyHawk,NC(seeFigure 5{40 )approximately3kmoshorein8.8mdepth.Thislocationisawayfrominletsandanyotherfreshwatersources.Therefore,currentstherearemostlydrivenbytide,windandwave.ThesecondlocationisinsideChesapeakeBaynearGloucesterPoint,VA(seeFigure 5{41 )wherecurrentsareaectednotonlybytheactionoftidebutsalinitystraticationaswell.Thedepthatthelocationis8.5m. Figure5{40: LocationofKittyHawk,NCwherecurrentsweremeasured. ItisofinteresttocomparemeasuredcurrentstothosecalculatedbyourstormsurgemodelingsystemingeneralandtheCH3Dmodelinparticular,primarily

PAGE 171

Figure5{41: LocationofGloucesterPoint,VAwherecurrentsweremeasured. toexaminehowrobustthecurrent-wavecouplingalgorithmsinCH3Dare.Thecouplingaccountsforwavesetupthrougheitherdepthuniform( Longuet-HigginsandStewart , 1964 )orverticallyvarying( Mellor , 2003 )radiationstresses,wave-enhancedsurfacestressandwave-inducedbottomfriction. Thecurrentswereclassiedasthe\SouthtoNorth"currentand\WesttoEast"current.Thelattercanbelookedatasthealong-channelcurrentatGloucesterPointandasthecross-shorecurrentnearKittyHawk(althoughnotexactlyduetotheshorelineorientationinthatarea)andtheformercanbeconsidered,againtosomedegree,asthelongshorecurrentnearKittyHawk.EightverticallayerswereusedintheCH3DmodeltosimulatecurrentsduringIsabel.Comparisonbetweenmeasuredandcalculated\SouthtoNorth"and\WesttoEast"currentsatKittyHawkisshowninFigures 5{42 and 5{43 .TheGloucesterPointcomparisonisshowninFigures 5{44 and 5{45 . SimulatedcurrentsatKittyHawkagreewellwithmeasureddata,althoughatthepeakofthestormthe\WesttoEast"simulatedcurrentisoverestimated.SimulatedcurrentsatGloucesterPointareinworseagreementwithmeasured

PAGE 172

Measured(left)andsimulated(right)\SouthtoNorth"currentatKittyHawk,NCduringHurricaneIsabel. Measured(left)andsimulated(right)\WesttoEast"currentatKittyHawk,NCduringHurricaneIsabel. data,especiallyintheWest-Eastdirection.Thisispartiallyduetothecloseproximityofthecurrentmetertotheshoreline.Neartheshoreline,owsimulatedbythemodelcloselyfollowstheshorelineorientation.Inordertoeectivelyaccountforhorizontaldiusion,therehastobeenoughgridcellsbetweentheinstrumentlocationandtheshoreline.ReningthegridintheareahelpedincreasetheaccuracyofsimulatedcurrentsbutitwasnotenoughtoachieveasaccurateagreementwithmeasureddataasitwasachievedatKittyHawk.

PAGE 173

Measured(left)andsimulated(right)\SouthtoNorth"currentatGloucesterPoint,VAduringHurricaneIsabel. Measured(left)andsimulated(right)\WesttoEast"currentatGloucesterPoint,VAduringHurricaneIsabel.

PAGE 174

5.2.1DescriptionAccordingtoNHC 5{46 .ThestormmovedrapidlyacrosstheCaribbean, Figure5{46: BesttrackofHurricaneCharley(courtesyofNOAANHC). andreachedhurricanestrengthonAugust11,150kmsouthofKingston,Jamaica.HurricaneCharleythenpassedjustsouthofJamaica,andthenextmorningpassedbetweenGrandCaymanandLittleCayman.OnthenightofAugust12,

PAGE 175

CharleypassedjusteastoftheIsleofYouth,thenovermainlandCuba,justwestofdowntownHavana. AfterpassingoverCuba,CharleycrossedtheStraitsofFlorida.Around13:00UTC,CharleypassedovertheDryTortugas.Tropicalstormforcewindsof41mph(65km/h)wererecordedatKeyWestInternationalAirport,115kmeast. ThecourseCharleytookatthistimecaughtmanybysurprise.InsteadoffollowingthepredictedtrackthroughtheTampa-St.Petersburgarea,Charleymadeanabruptturntothenortheast,headingforFortMyersandSanibelIsland.Nevertheless,thistrackwaswellwithintheocialforecast'smarginoferror. Atthesametimeasitturned,Charleyrapidlystrengthened,goingfromaCategory2stormat110mph(170km/h)toaCategory4stormat145mph(235km/h)inonlythreehours.Thisrapidintensicationwasoutsidetheocialforecast,whichcalledforonlyaslightstrengtheningbeforelandfall.ThechangeinstrengthwassodrasticthattheNHCissuedaspecialhurricaneadvisoryoutsideofitsnormalschedule.Itispossiblethatthewindswereevenstrongeratlandfall,possiblyatornearCategory5strength(155mphor250km/h),basedonlaterimagesandassessments. CharleybecamethesecondtropicalstormtostrikeFloridain24hourswhenTropicalStormBonniestrucktheFloridapanhandleinApalachicolaat14:00UTConAugust12,22hoursbeforeCharleywentovertheDryTortugas.Thismade2004therstyeartwonamedstormshavestruckthesamestateinthesame24-hourperiodsince1906.Mainlandlandfalloccurredonly29hoursapart. OnAugust13at19:45UTC,CharleymadelandfallatCayoCosta,northofFortMyers.CharleymovedinlandnearCharlotteHarborshortlyafterwards. Nearmidnightlocaltime(August14,04:00UTC),Charleybeganmovingbackoverwater,exitingFloridanearDaytonaBeach.Itreturnedtolandaround15:00UTCnearNorthMyrtleBeach,SCstillretaininghurricanestrength.Charley

PAGE 176

continuedtorunoandonlanduptheEastCoastoftheUnitedStates,anddissipatednearCapeCodaroundmid-dayonAugust15.Charley'sstrongestgustsweremeasuredat180mph(290km/h)atPuntaGorda. OnedeathinJamaica,fourdeathsinCuba,and10deathsintheUnitedStatesweredirectlyattributedtoCharley.PropertydamagefromCharleywasestimatedbytheNHCat$14billion.ThismakesCharleythesecondmostcostlyhurricaneinAmericanhistory,behindHurricaneAndrew's$26billionin1992,andaboveHurricaneHugo's$7billion($9.4billionin2000dollars)in1989. 5{47 )containsthreeopenboundaries.ThesouthernopenboundarystartsjustnorthofNaples,FLandextends60kmoshore.ThenorthernopenboundarystartsnearVenice,FLandstretches50kmoshore.Thelengthofthewesternopenboundaryis107km.Theareaofthecomputationaldomainisapproximately10,350km2withthetotalnumberofcomputationalgridcellsof22,419.10,127(45%)ofthosecomputationalcellsarewatercells.Gridspacingvariesfrom2kmneartheoshoreopenboundaryto200mwithinSanCarlosBayandEsteroBay.ThegridcoverstheentireCharlotteHarborwithallofitsriverbasinssuchasCaloosahatchee,Peace,andMyakkarivers.Thegridcontainslandcellstoaccountforwetting-and-drying.TheUSGSNationalElevationDataset(

PAGE 177

islands.Thegridextendsfarinlandsothatitwouldbepracticallyimpossibleforthewatertoreachtheinlandgridboundariesduringcoastalinundation. TheCharlotteHarborgriddomain. Both,windspeedanddirectionweremeasuredattwostations:onestationatFtMyersandanotheronelocatedatNaples.TheNaplesstationislocatedapproximately10kmsouthofourmodeldomain'ssouthernopenboundary.Such

PAGE 178

proximityallowedustomakeuseofthewinddataforcomparisonwithWNAandWINDGENwind.Figures 5{48 through 5{51 demonstrateacomparisonanalysisbetweenmeasuredwindspeedanddirectionatthetwowindstationsvs.thoseobtainedusingWNAandWINDGENwinddata.TheFtMyersstationwasinoperativefor2.5hoursduringthetimewhenCharleywaspassingoverthearea.ThewinddirectionmeasuredatFtMyerswasalmostconstantallthetimewhichwasalittlestrange.Thesametrendwasobservedafteranalyzingthedataovera3-monthperiodpriortoCharley.SincewinddirectionmeasuredatNapleswasingoodagreementwithWNAandWINDGENwinddirection,winddirectiondataatFtMyerswereconsiderederroneousandwerenotused. TheWINDGENwindatthetwostationswasstrongerthantheWNAwind,thoughitwashardtojudgewhichoneofthetwowasbetterbecausethestationatFtMyersbrokedownbeforethewindreacheditsmaximumandtheNaplesstationwasalittletoofarfromthecenterofthehurricane. Figure5{48: WMeasuredwindspeedvs.WINDGENandWNAwinddataatFtMyers,FLduringHurricaneCharley.

PAGE 179

Figure5{49: Measuredwinddirectionvs.WINDGENandWNAwinddataatFtMyers,FLduringHurricaneCharley. Figure5{50: Measuredwindspeedvs.WINDGENandWNAwinddataatNaples,FLduringHurricaneCharley.

PAGE 180

Figure5{51: Measuredwinddirectionvs.WINDGENandWNAwinddataatNaples,FLduringHurricaneCharley. 5{52 through 5{55 .AlltheothersimulatedversusmeasuredwaterelevationresultsduringHurricaneCharleyareshowninAppendix G . Ascanbeseeninthegures,simulatedwaterelevationusingWNAwindwassignicantlyunderestimatedatallfourlocations.Incontrasttotheseresults,simulatedwaterelevationusingWINDGENwindlookedmuchbetter,matchingthesurgepeakverywellatBigCarlosPass,slightlyoverestimatingthesurgepeakatbothlocationsinEsteroBay,andunderestimatingthepeaksurgevalueatFtMyers.SuchdiscrepancyinthesimulatedwaterelevationsusingthetwowindsisaresultoflargedierencebetweenWNAandWINDGENwindeldsduringthe

PAGE 181

Figure5{52: Comparisonofsimulatedvs.measuredwaterelevationatBigCarlosPass.Twosimulatedresultsareshown:oneusingWNAwindandanotherusingWINDGENwind. Figure5{53: Comparisonofsimulatedvs.measuredwaterelevationatEsteroBay,location1.Twosimulatedresultsareshown:oneusingWNAwindandanotherusingWINDGENwind.

PAGE 182

Figure5{54: Comparisonofsimulatedvs.measuredwaterelevationatEsteroBay,location2.Twosimulatedresultsareshown:oneusingWNAwindandanotherusingWINDGENwind. Figure5{55: Comparisonofsimulatedvs.measuredwaterelevationatFtMyers.Twosimulatedresultsareshown:oneusingWNAwindandanotherusingWIND-GENwind.

PAGE 183

timewhenthestrongestwindswereexperiencedinthearea.Inordertoinvestigatethisproblem,windsnapshotsintheEsteroBayareaweretakenatthreeinstantsshowninFigure 5{56 :(1)beforethestormpeak(Aug-1316:50UTC,JulianDay=226.701),(2)rightatthepeak(Aug-1320:40UTC,JulianDay=226.862),and(3)afterthestormpeak(Aug-1401:20UTC,JulianDay=227.055).Figure 5{56 showsthemeasuredandsimulated(usingWNAandWINDGENwind)watersurfaceelevationatEsteroBaylocation1.Italsospeciesthetimeswhenthewindsnapshotsweretaken.FromFigures 5{57 through 5{62 thatdepictwindspeedand Figure5{56: Simulatedvs.measuredwaterelevationatEsteroBay,location1.Dashedlinesspecifythethreetimeinstantswhenwindsnapshotsweretaken. directionalongwithtotalwaterdepthcontoursintheEsteroBayarea,themostsignicantdierencebetweenWNAandWINDGENwindeldscanbefoundinsnapshot2,whichwastakenatthepeakofthestormsurgeinEsteroBay.Thereisanevidentdiscrepancyinwinddirectionwhichisresponsibleforthesignicantdierenceinwaterelevation. Nowthequestionis:HowcondentareweinpickingWINDGENwindoverWNAwindeventhoughwaterelevationsimulatedusingWINDGENwind

PAGE 184

Figure5{57: WNAwindeldsnapshot1(Aug-1320:55,JulianDay=226.872)alongwithtotaldepthcontoursintheEsteroBayarea. Figure5{58: WINDGENwindeldsnapshot1(Aug-1320:55,JulianDay=226.872)alongwithtotaldepthcontoursintheEsteroBayarea.

PAGE 185

Figure5{59: WNAwindeldsnapshot2(Aug-1320:55,JulianDay=226.872)alongwithtotaldepthcontoursintheEsteroBayarea. Figure5{60: WINDGENwindeldsnapshot2(Aug-1320:55,JulianDay=226.872)alongwithtotaldepthcontoursintheEsteroBayarea.

PAGE 186

Figure5{61: WNAwindeldsnapshot3(Aug-1401:20,JulianDay=227.055)alongwithtotaldepthcontoursintheEsteroBayarea. Figure5{62: WINDGENwindeldsnapshot3(Aug-1401:20,JulianDay=227.055)alongwithtotaldepthcontoursintheEsteroBayarea.

PAGE 187

compareswithmeasuredwaterelevationmuchbetterthanthatcalculatedusingWNAwind.Figure 5{63 showsacomparisonbetweenWNAandWINDGENwinddirectionvs.actualwinddirectionmeasuredatNaples,FLwhichisapproximately30kmsouthofEsteroBay.Ascanbeseenfromthegure,atthetimeofhighsurge(whensnapshot2wastaken),WINDGENwinddirectioncomparesverywellwiththemeasuredwinddirectionandtheWNAwinddirectionisobyapproximately30degrees.ThisleadstotheconclusionthatWINDGENwindeldismorereliablewhenitmatteredmost(i.e.whenthestormwasatitspeak)andthesignicantunderestimationofstormsurgecalculatedusingWNAwindisduetotheincorrectwindinformationprovidedtothemodelingsystembytheWNAwind.ThisshouldnotbeasurprisebecauseWNAwindisnotknowntoresolvethehurricanewindeldaccurately.MaximumwaterelevationrelativetoNAVD88 Figure5{63: WNAandWINDGENwinddirectionvs.measuredwinddirectionatNaples,FL.Atthepeakofthestorm(denotedbynumber2),WINDGENwinddirectionmatchesverywellwiththemeasuredwinddirectionandWNAwinddirectionisobyapproximately30degrees. (includestide,surgeandwavesetup)calculatedduringsimulationofHurricaneCharleyinCharlotteHarborusingWINDGENwindisshowninFigure 5{64 .As

PAGE 188

canbeseeninthegure,themaximumvaluesreached1.9mnearEsteroBaycausingextensiveinundationinthatarea.Figure 5{65 showssimulatedstormsurgeatallfourstationswhichwasobtainedbysubtractingsimulatedtide(nootherforcingmechanismwasincluded)fromsimulatedwaterlevelswhichincludeallforcingmechanisms(tide,wind,wavesetup,waveenhancedsurfaceandbottomfriction).TheresultsshowthatthecalculatedstormsurgewithinEsteroBaywasaround1.2m.Itisworthnotingthatduringthepeakofthestorm,thetidewasinitsoutgoingstageatalevelofapproximately-0.1m.Therefore,ifthestormweretohaveoccurredathightidethesurgelevelwouldhavebeen0.5to0.6mhigher. Ittookapproximately11hoursfromthepointwhenthesurgelevelstartedtorisetothepointwhenthesurgereceded.Forcomparison,ittooknearly30hoursforstormsurgecausedbyHurricaneFrancesintheTampaBayareatorecede.CharleywasafastmovinghurricanewhereasFranceswasmovingveryslowly.

PAGE 189

Figure5{64: MaximumwaterelevationrelativetoNAVD88(includestide,surgeandwavesetup)calculatedduringsimulationofHurricaneCharleyinCharlotteHarborusingWINDGENwind.

PAGE 190

Figure5{65: Simulatedstormsurge(waterlevelminustide)atthefourstationsusingWINDGENwind.

PAGE 191

5.1.7 ,erroranalysisisagoodwayofcomparingcalculatedwaterelevationtomeasuredwaterlevel.Inthatsectionitwasalsopointedoutthatanerrorbetweenmeasuredandcalculatedwaterelevationcanbeattributedtoeithertideorstormsurgeortheircombination.Anestimationofthecontributionoftidetothetotalerrorwasdonebyperforminga\pure"tidesimulationandcomparingitsresultswithtidalelevationwhichwaslteredoutfrommeasuredwaterelevationusing DoodsonandWarburg ( 1941 )39-hourlyaveragetidallter. Inordertoweightheeectofeachcomponentinvolvedinthenon-linearinteractionbetweenthesurge,tide,wind,andwave,severalsimulationsweremadebyincludingdierentcomponentcombinations.Turningthewetting-and-dryingfeatureonandowasanoptionaswell.Table 5{11 specieswhichsimulationhadwhatfeatures. Table5{11:Alistofsimulationswithvariouscombi-nationsofsixmodelfeatures(psymboldenotesthefeaturewasincludedduringthesimulation). Factors Sim1 Sim2 Sim3 Sim4a Tide p p p p p p p p p p p p p p p p p p p p p ( 1993 )formulationwasused2 ( 1979 )formulationwasused3 ( 1989 )formulationwasused

PAGE 192

Table 5{12 showstheRMS,MeanAbsolute,andMaximumAbsoluteerrorsofcalculatedwaterelevationduringHurricaneCharley(seeAppendix B forformulasusedtocalculatetheerrors).Errorsofpeakvalues(measuredpeakelevationminussimulatedpeakelevation)andtimingerrors(timewhenmeasuredpeakelevationoccurredminustimewhensimulatedpeakelevationoccurred)arealsoshown.Aseparatecolumninthetabledisplaystheerrorsattributedto\pure"tide.Tidalrangeisalsoshownforeachstation.

PAGE 193

Table5{12:ErrorsofwaterelevationattidestationsduringHurricaneCharley. WNAwind WINDGENwind Tide Sim1 Sim2 Sim3 Sim4a Sim1 Sim2 Sim3 Sim4a FtMyears(depth=3.2m;tidalrange=50cm) RMSError(cm) 5.9 20.7 20.4 21.3 21.4 20.6 17.2 13.7 13.3 11.5 13.2 13.6 11.5 MeanAbsError(cm) 5.0 11.9 11.8 12.2 12.3 12.1 11.3 10.0 9.8 8.9 9.7 10.0 9.4 MaxAbsError(cm) 13.4 76.1 75.2 79.8 80.4 77.6 53.5 43.8 41.9 34.6 43.2 46.8 32.0 Meas.SurgePeak(cm) 99 ErroratPeak 43 42 36 47 44 -5 TimingatPeak(min) 32 26 5 -13 7 -66 BigCarlosPass(depth=3.8m;tidalrange=90cm) RMSError(cm) 10.3 23.7 23.3 22.2 21.5 22.6 21.6 17.1 16.4 14.8 14.3 13.1 19.3 MeanAbsError(cm) 8.5 16.0 15.8 15.4 15.2 14.3 14.4 12.1 11.8 10.9 10.7 10.0 13.1 MaxAbsError(cm) 27.2 97.7 95.6 93.1 90.6 106.0 95.9 59.5 57.3 53.1 48.2 42.6 71.7 Meas.SurgePeak(cm) 119 ErroratPeak 35 32 19 21 21 -1 TimingatPeak(min) 45 45 45 17 45 45 EsteroBay1(depth=2.8m;tidalrange=105cm) RMSError(cm) 9.1 24.5 23.7 22.5 21.8 22.5 22.1 17.7 17.2 16.3 15.9 14.6 21.1 MeanAbsError(cm) 7.4 14.9 14.5 14.2 13.8 14.0 13.3 13.6 13.0 12.7 12.1 11.3 14.3 MaxAbsError(cm) 26.0 109.8 107.6 109.5 106.1 98.5 111.0 55.4 53.6 52.2 47.1 40.2 69.8 Meas.SurgePeak(cm) 112 ErroratPeak 17 12 -4 0 -2 -25 TimingatPeak(min) -10 -10 -10 -10 -10 -12 EsteroBay2(depth=2.2m;tidalrange=110) Continuedonnextpage

PAGE 194

WINDGENwind Tide Sim1 Sim2 Sim3 Sim4a Sim1 Sim2 Sim3 Sim4a RMSError(cm) 9.2 21.6 21.1 19.7 18.9 20.9 19.0 17.9 16.1 15.9 15.6 14.8 19.1 MeanAbsError(cm) 7.5 14.4 14.2 13.7 13.5 14.4 12.7 13.3 12.3 12.2 12.1 11.5 12.8 MaxAbsError(cm) 25.5 86.9 84.5 79.1 78.5 81.0 79.1 57.6 50.9 47.9 46.5 41.3 70.3 Meas.SurgePeak(cm) 102 ErroratthePeak 18 14 -1 0 0 -18 TimingatPeak(min) 25 25 25 24 24 AvgRMSError(cm) 8.6 22.6 22.1 21.4 20.9 21.6 20.0 16.6 15.8 14.6 14.8 14.0 17.7 AvgErr.atPeak(cm) 28 25 15 17 16 12 AvgTimingError(min) 28 27 21 16 21 37

PAGE 195

BasedonRMSerrorsandaverageabsoluteerrorsshowninthetableabove,itcanbeconcludedthattheWINDGENwindgavesignicantlymoreaccuratesurgeelevationsthantheWNAwind.ThepeakwaterelevationscalculatedusingWNAwindwerenotverywellpronounced,thereforeerrorsatthepeakand\timing"errorsatthepeakwerenotcalculatedandarenotshowninTable 5{12 .Theaccuracyofthesimulatedtidewasnotverygood,withtheaverageRMSerrorof9to10cm,whichmightbeduetoasignicantcontributionofnon-lineartidalconstituentswhichwerenotincludedintheboundaryconditionofthismodelsimulation. ThesimulatedwaterelevationatFtMyerswassignicantlyunderestimated,whereasattheotherthreestationstheresultslookedverygood.ItishardtosaywhatcausedtheFtMyerselevationtobesolow.ThestationislocatedontheCaloosahatcheeRiver,approximately25kminland.Thewindstationlocatedtherebrokedownbeforethestormreacheditsmaximumstrengthinthearea,thustheaccuracyofthewinddatausedinthemodelnearFtMyersisuncertain. Overall,excludingtheFtMyersstation,Simulation4bproducedbetterresultsintermsofsmallerRMSerrorsandbettercomparisonwithmeasuredwatersurfaceelevationatitspeak.Simulation5,whenthewetting-and-dryingfeaturewasdisabled,signicantlyworsentheresultsforthethreestationswithinEsteroBay.Ontheotherhand,Simulation5signicantlyimprovedthecomputedwaterlevelatFtMyers.ThishappenedbecausewithoutwettingmorewaterwaspusheduptheCaloosahatcheeRiversignicantlyincreasingthewaterelevationthere. The\timing"errorsareinexcellentagreementwithmeasureddataforallfourstations(5to45min,withtheaverageerroraround20min). Table 5{13 showsthepeakwaterelevationvaluescalculatedduringthesesimulationsalongwiththemeasuredones.Bynormalizingthedierencebetweenconsecutivesimulationsbythepeakmeasuredwaterlevel,wecancalculatepercent

PAGE 196

increaseordecreasethateachcomponentintroducedatthetimewhenthepeakwaterelevationwasobserved(e.g.,atBigCarlosPass,theSimulation2value"3%relativetotheSimulation1value,andtheSimulation4aand4bvalues#7%and#2%relativetotheSimulation3value,respectively). Includingradiationstresstermsincreasedthecalculatedwaterlevelby1-5%.Addingwaveenhancedsurfacestresshelpedfurtherincreasethecomputedwaterelevationby6-15%.Accountingforwaveenhancedbottomfrictiondecreasedthecalculatedwaterelevationby7-16%and1-8%using GrantandMadsen ( 1979 )and ShengandVillaret ( 1989 )formulations,respectively.Whenwetting-and-dryingschemewasnotengagedduringthecalculationthepeakwaterlevelvaluegrewanextra18-49%. Table 5{14 showspeakstormsurgevalues(tidewassubtractedfromwaterelevation)calculatedduringthesesimulations.Percentincreaseordecreasewascalculatedthesamewayitwascalculatedintheprevioustable.

PAGE 197

Table5{13:MeasuredpeakwaterelevationsatfourstationsduringHurricaneCharleyusingWINDGENwindandvariouscombinationsofstormsurgemodelfeatures. Sim1 Sim2 Sim3 Sim4a Meas. cm cm cm cm cm cm cm FtMyers 56 57"1% 63"6% 53#16% 55#8% 104"49% 99 BigCarlosPass 84 87"3% 100"11% 105#7% 98#2% 120"18% 119 EsteroBay#1 95 100"5% 116"14% 114#13% 112#4% 137"22% 112 EsteroBay#2 84 88"4% 103"15% 110#9% 102#1% 120"18% 102 Table5{14:Calculatedpeakstormsurge(withtidessubtracted)atfourstationsduringHurricaneCharleyusingWINDGENwindandvariouscombinationsofstormsurgemodelfeatures. Sim1 Sim2 Sim3 Sim4a cm cm cm cm cm cm FtMyers 43 45"2% 54"9% 48#12% 46#8% 88"42% BigCarlosPass 86 90"3% 105"13% 113#7% 103#2% 118"13% EsteroBay#1 104 109"4% 124"13% 126#11% 120#4% 139"17% EsteroBay#2 86 91"5% 106"13% 118#5% 104#2% 122"18%

PAGE 198

Theresultsrevealthatincludingwavesetupincreasedandimprovedthecalculatedsurgeby2-5%.ThisincreasewasnotassignicantasitwasobservedduringHurricaneIsabel(3-18%)duethereasonthatwaveswerenotashighduringCharleyastheywereduringIsabelandallfourstationswerelocatedinestuaries(EsteroBayandCaloosahatcheeRiver),andthusweresheltered,whereassomeoftheIsabelstationssuchasBeaufort,Duck,andChespeakeBayBridgewerelocatedontheopenoceanfront. Thewave-enhancedwindwasasignicantfactorinincreasingandimprovingthecalculatedstormsurge.Itseectrangedwithin9-13%.Suchsignicancecanbeexplainedbythefactthatthewavesenteringthecomputationaldomainwerenotveryhigh,around4minheightwithapproximately9secperiod(forcomparison,duringIsabelwaveswereashighas15mwith14secwaveperiod).Withstrongwindblowingonshore,thesmall,shortperiodwaveswerenotfullydeveloped,inotherwords,theywerestillyoungandenergetic.Youngseasincreaseseasurfaceroughnessand,asaresult,surfacestressalsoincreases. Wave-enhancedbottomstressbasedonthe ShengandVillaret ( 1989 )formula-tion(Simulation4b)decreasedthepeaksurgevaluesby2-4%nearEsteroBayandby8%atFtMyersstationwherethemoresignicanteectcanbeexplainedbytheremotenessofthestationandshallowdepthsinthearea,2to4m. Whenthewetting-and-dryingfeaturewasinactive,thewaterlevelgrewsignicantlyby13-42%.ThissignicantincreasecanbeexplainedbythefactthatmostofthecalculatedoodingwouldhaveoccurredinthevicinityofEsteroBay,sowhentheoodingfeaturewasinactive,thewatercouldnotpropagateinlandthuspilingupneartheshoreandsignicantlyoverestimatingthecalculatedstormsurge.

PAGE 199

5{66 and 5{67 showmaximumsimulatedinun-dationcausedbyCharleyduringSimulation4(seeTable 5{11 )usingWINDGENandWNAwinds,respectively.Thebottomplotsintheguresidentifythetimewhenthehighestoodleveloccurred.SincetheWNAwindwasweakerthanWINDGEN,itproducedlessinundation.Simulation4usingWINDGENwindproducedthebestcomparisonbetweenmeasuredandsimulatedwaterelevationatthreestationswithinEsteroBaywheremostoftheoodingoccurred.Thus,thissimulationwastakenasthebasesimulationforestimatingtheamountofinundationcausedbyCharley.Theoodedareaaected530km2ofland,mostlythesurroundingsofEsteroBay,SanCarlosBay,SanibelIsland,andPineIsland.

PAGE 200

Figure5{66: MaximumsimulatedinundationinCharlotteHarborusingWIND-GENwind(toppanel).Thebottompanelshowsthetimeduringwhichthemaxi-mumoodoccurred.

PAGE 201

Figure5{67: MaximumsimulatedinundationinCharlotteHarborusingWNAwind(toppanel).Thebottompanelshowsthetimeduringwhichthemaximumoodoccurred.

PAGE 202

Inordertovalidatethecalculatedinundation,airphotostakenbeforeandafterCharleypassedoverthedomainwerecomparedwithourcalculatedoodmaps.Figures 5{68 and 5{69 showpre-stormandpost-stormairphotostakenattwolocations,CaptivaIslandandSanibelIsland,respectively(seeFigure 5{47 forlocationinformation).Ascanbeseenfromthesegures,bothislandsweresubjecttoextensiveinundationduringCharley.Thebottompanelsineachguredisplayclose-upsofourcomputedinundationmapswhichalsoverifythepresenceofwaterovertheislands.Thisvalidationisratherqualitativebutnonethelessimportant. Amorequantitativeoodanalysiswasdonebasedonsomeevidencesobtainedfrom\HurricaneCharleyPost-StormConditionsandImpact"report( ClarkandLaGrone , 2004 ).Inthisreport,fourlocationswereexaminedonpresenceofhighwatermarksleftbytheoodcausedbyHurricaneCharley.Thelocationsare:GasparillaIsland,NorthCaptivaIsland,CaptivaIsland,andEsteroIsland(seeFigure 5{70 forlocations).Inordertoseehowourmodelresultsstackupagainstthereportedstormsurgevalues,twotechniquesofevaluatingwatermarkswereapplied:Technique1usesthestormsurgelevelcomputedfromtheintegratedstormsurgemodelingsystem,andTechnique2addsawaveheighttothestormsurgelevelcalculatedbyTechnique1.Thewaveheightwascalculatedaccordingtothemethodologyof FederalEmergencyManagementAgency(FEMA) ( 1988 )forestimatingoodzones: FloodLevel=Surge+0:7Hcontrolling(5{1) SinceHcontrolling=1:6Hsig Table 5{15 belowshowscomparisonbetweenreportedhighwatermarkvaluesandthosecalculatedbythetwotechniques.

PAGE 203

Figure5{68: Pre-storm(top)andpost-storm(middle)airphotostakenbyCaptivaIsland(courtesyofUSGS,

PAGE 204

Figure5{69: Pre-storm(top)andpost-storm(middle)airphotostakenbySanibelIsland(courtesyofUSGS,

PAGE 205

NauticalchartofcoastalareasintheCharlotteHarborareaimpactedbyHurricaneCharley(from ClarkandLaGrone ( 2004 )). Table5{15: Comparisonbetweenreportedhighwatermarkvaluesandoodlevelscalculatedusingtwotechniques HighWaterMarkValue,ft Location Reported SimulatedbyTechnique1 SimulatedbyTechnique2 GasparillaIsland 4-5 2.1 3.7 N.CaptivaIsland 9 3.3 8.8 CaptivaIsland 7-8 4.8 11.7 EsteroIsland 6-7 4.8 7.0

PAGE 206

Technique2givesresultsclosertothereportedvaluesbutsinceitisnotveryclearhowaccuratelythewatermarksweremeasuredandwhatverticaldatumwasusedtomeasurethewatermarks(seeFigure 5{71 ),thisanalysisanditsresultsaresomewhatuncertain.Despitetheuncertaintiesintheaccuracyofmeasuredhighwatermarks,theresultscalculatedbyTechnique2wouldbeamorefaircomparisonbecausethehighwatermarksweremostlikelyleftbywavecrests,notjuststormsurge. ManpointsatahighwatermarkleftbystormsurgecausedbyHurri-caneCharleyonNorthCaptivaIsland(from ClarkandLaGrone ( 2004 )).

PAGE 207

5.3.1DescriptionAccordingtoNHC 5{72 . Figure5{72: BesttrackofHurricaneFrances(courtesyofNOAANHC). AstrongtropicalwavedevelopedintoatropicaldepressionlateonAugust24,2004.Itwasthen1,400kmwest-southwestofCapeVerde,andabout2,700kmeastoftheWindwardIslands.ThenextdayitwasupgradedandnamedTropicalStormFrances.ThestormwasupgradedtoahurricaneonAugust26.

PAGE 208

Francesstrengthenedrapidly,reachingCategory3intensity24hourslateronthe27thandCategory4thenextday.InitiallyforecasttoturnnorthandpotentiallythreatenBermuda,conditionschangedandFrances'spredictedtrackshiftedwestwardtowardtheBahamas.Frances'sintensityuctuatedasittravelledwestoverthenextseveraldays,droppingbacktoaCategory3stormbeforerestrengthening.Thisdropandsubsequentrestrengtheningwaslikelycausedbyaneyewallreplacementcycle,accordingtotheNationalHurricaneCenter. Overthenextseveraldays,FrancespassedjustnorthoftheAntilles,withonlyitsouterrainbandsaectingtheBritishVirginIslandsandtheDominicanRepublic.OntheeveningofSeptember1,FrancespassedtothenorthofGrandTurkintheTurksandCaicosIslands.AlthoughFrancesdidnotstriketheislanddirectly,hurricaneforcewindswerereportedthere. OnSeptember2,FrancesstrucktheBahamasdirectly,passingdirectlyoverSanSalvadorIslandandveryneartoCatIsland,andpassingoverEleutheraonSeptember3.ReportsfromLongIslandsaidthatpartsoftheislandremainedunderwaterafterthestormhadpassed,withnumeroushomesandotherstructuresdamaged.OnSaturday,September4,theairportatFreeport,GrandBahamawasreportedtobeunder6to8feetofwater.OnedrowningdeathwasreportedinFreeport,GrandBahama. OnSeptember3,FrancesweakenedslightlyasitpassedintothevicinityofAbacoIslandanddirectlyoverGrandBahama.ThestormweakenedfromaCategory3to2priortopassingoverGrandBahamaandalsolessenedinforwardspeed.PartsofSouthFloridabegantobeaectedbysquallsandtheouterrainbandsofthehurricaneatthistime.Gustsaslowas40mph(60km/h)toashighas87mph(140km/h)werereportedfromJupiterInlettoMiami. Francesmovedextremelyslowly,from5to10mph(8to16km/h),asitcrossedthewarmGulfStreambetweentheBahamasandFlorida,leadingto

PAGE 209

fearsitcouldrapidlyrestrengthen.ItremainedstableatCategory2hurricaneandbatteredtheeastcoastofFlorida,especiallybetweenFortPierceandWestPalmBeach,formostofSeptember4.At03:00UTConSeptember5,thewesternedgeofFrances'seyewallbeganmovingonshore.BecauseofFrances'slargeeyeofroughly130kmacrossandslowmotion,thecenterofcirculationremainedoshoreforseveralmorehours.At05:00UTC,thecenterofthebroadeyeofFrancesnallywasoverFlorida,nearSewall'sPoint,Stuart,JensenBeachandPortSalerno. LateonSeptember5,itpickedupspeedandcrossedtheFloridaPeninsula,emergingovertheGulfofMexiconearTampaasatropicalstorm.Afterashorttripoverwater,FrancesagainstrucklandnearSt.Marks,Florida.Francesheadedinland,weakeningtoatropicaldepressionandcausingheavyrainfalloverthesouthernUS.TropicalDepressionFrancescontinuednorth,maintainingitscirculationlongerthanexpected.USforecastersattheHydrometeorologicalPredictionCentercontinuedissuingadvisoriesontheremnantsofFrancesuntilthesystemcrossedtheCanadianborderintoQuebec,whereupto8inches(200mm)ofrainfell,causingsignicantooding. TwodeathshavebeenreportedintheBahamas.32deathsareblamedonthestorminFlorida,twoinGeorgiaandoneinSouthCarolina. TheinsuredclaimsofFranceshavebeendeterminedtobeabout$4billion.SomeareasofFloridareceivedover13inchesofrain.Francesalsospawned117tornadosfromFloridatoasfarnorthasVirginia.Thisamountbeatstherecordnumberoftornadosforahurricane,whichwas115forHurricaneBeulahin1967. 5{73 )containsthreeopenboundaries.ThesouthernopenboundarystartsatVenice,FLandextends28kmoshore.ThenorthernopenboundarystartsnearCrystalBeach,FLandstretches36kmoshore.The

PAGE 210

lengthofthewesternopenboundaryis120km.Theareaofthecomputationaldomainisapproximately7,000km2withthetotalnumberofcomputationalgridcellsof54,476andtheaveragegridspacingofapproximately350m.28,317(52%)ofthosecomputationalcellsarewatercells.ThegridcoverstheentireTampaBaywithallofitsriverbasinsandSarasotaBayaswell.Thegridcontainslandcellsinordertouseitforwetting-and-drying.TheUSGSNationalElevationDataset(

PAGE 211

TheTampaBaygriddomain.

PAGE 212

bayfor20hours,thewaterlevelrosesteadilygainingupto165cm(atStPetestation)fromthepointwhenthelevelreacheditslowestvalue. 5{74 through 5{76 .AlltheothersimulatedversusmeasuredwaterelevationresultsduringFrancesareshowninAppendix H . Figure5{74: Comparisonofsimulatedvs.measuredwaterelevationatClearwater,FL.Twosimulatedresultsareshown:oneusingWNAwindandtheotherusingWINDGENwind. MaximumwaterelevationrelativetoNAVD88(includestide,surgeandwavesetup)calculatedduringsimulationofHurricaneFrancesinTampaBayusingWNAwindisshowninFigure 5{77 .Ascanbeseeninthegure,themaximum

PAGE 213

Figure5{75: Comparisonofsimulatedvs.measuredwaterelevationatStPete,FL.Twosimulatedresultsareshown:oneusingWNAwindandtheotherusingWINDGENwind. Figure5{76: Comparisonofsimulatedvs.measuredwaterelevationatPortMan-atee,FL.Twosimulatedresultsareshown:oneusingWNAwindandtheotherusingWINDGENwind.

PAGE 214

valuesreached2.0minthenortheasternpartofthebay,nearMcKayBayandUpperTampaBayPark,oodingsomeofthelandinthoseareas. Figure 5{78 showssimulatedstormsurgeatallfourstationswhichwasobtainedbysubtractingsimulatedtide(nootherforcingmechanismwasincluded)fromsimulatedwaterlevelswhichincludedallforcingmechanisms(tide,wind,wavesetup,waveenhancedsurfaceandbottomfriction). Theresultsshowthatthecalculatedstormsurgerangedfrom0.6to0.8m.Itshouldbenotedthatduringthepeakofthesurge,thetidewasatitslowstageatanapproximatelevelof-0.3m.Therefore,ifthepeakofthestormweretocoincidewithahightidethesurgelevelwouldhavebeen0.6to0.7mhigher.Aninterestingfactisrevealedbylookingatthedierencebetweenthetimeswhenthewaterelevationandstormsurgelevelreachedtheirpeakvalues:Thesurgereacheditsmaximumapproximately4hoursafterthewaterelevationmaximumoccurred.AnotherinterestingfactisfoundbyobservingthesimulatedstormsurgeatStPeteandPortManateeduringJuliandays249-250:DuringthattimeFranceswascrossingtheFloridapeninsulaapproachingtheTampaBayareawithitswindsblowingprimarilyfromnorthtosouth.Asaresult,thesurgeleveldecreasedatStPetewhichisonthenorthernsideofthebaycausingset-down,andincreasedatPortManateewhichisacrossthebayfromStPete.AfterthehurricaneapproachedthenortheasternpartofTampaBaycomingfromtheeast,thewinddirectionstartedtochangefromnorth-to-southtowest-to-eastcausingthewaterlevelatPortManateetoslightlydecreaseforsometimeuntilthepointwhentheeyeofHurricaneFrancesmovedclosertotheGulfofMexicoslightlyincreasingitsstrengthandstartedpushingthewaterintoTampaBaycausingthemajorstormsurgetoincrease. Despitethefactthatthestormsurgelevelwasnottoohigh(thankstolowtideduringthepeak),ittookapproximately30hoursfromthepointwhenthe

PAGE 215

Figure5{77: MaximumwaterelevationrelativetoNAVD88(includestide,surgeandwavesetup)calculatedduringsimulationofHurricaneFrancesinTampaBayusingWNAwind.

PAGE 216

surgelevelstartedtorisetothepointwhenthesurgereceded.ThemainreasonforthatistheslowmovingHurricaneFrances(5-10mph).Forcomparison,ittookonly11hoursforthestormsurgecausedbyHurricaneCharleytorecedewhenitmadelandfallandwentovertheCharlotteHarborarea. Figure5{78: Simulatedstormsurge(waterlevelminustide)atthethreestationsusingWNAwind. 5.1.7 ,erroranalysisisagoodwayofcomparingcalculatedwaterelevationtomeasuredwaterlevel.Inthatsectionitwasalsopointedoutthatanerrorbetweenmeasuredandcalculatedwaterelevationcanbeattributedtoeithertideorstormsurgeortheircombination.Anestimationofthecontributionoftidetothetotalerrorwasdonebyperforminga\pure"tidesimulationandcomparingitsresultswithtidalelevationwhichwaslteredoutfrommeasuredwaterelevationusing DoodsonandWarburg ( 1941 )39-hourlyaveragetidallter. Inordertoweightheeectofeachcomponentinvolvedinthenon-linearinteractionbetweenthesurge,tide,wind,andwave,severalsimulationsweremade

PAGE 217

byincludingdierentcomponentcombinations.Turningthewetting-and-dryingfeatureonandowasanoptionaswell.Table 5{16 specieswhichsimulationhadwhatfeatures. Table5{16:Alistofsimulationswithvariouscombi-nationsofsixmodelfeatures(psymboldenotesthefeaturewasincludedduringthesimulation). Factors Sim1 Sim2 Sim3 Sim4a Tide p p p p p p p p p p p p p p p p p p p p p Table 5{17 showstheRMS,MeanAbsolute,andMaximumAbsoluteerrorsofcalculatedwaterelevationduringHurricaneFrances(seeAppendix B forformulasusedtocalculatetheerrors).Errorsofpeakvalues(measuredpeakelevationminussimulatedpeakelevation)andtimingerrors(timewhenmeasuredpeakelevationoccurredminustimewhensimulatedpeakelevationoccurred)arealsoshown.Aseparatecolumninthetabledisplaystheerrorsattributedto\pure"tide.Tidalrangeisalsoshownforeachstation. ( 1993 )formulationwasused2 ( 1979 )formulationwasused3 ( 1989 )formulationwasused

PAGE 218

Table5{17:ErrorsofwaterelevationattidestationsduringHurricaneFrances. WNAwind WINDGENwind Tide Sim1 Sim2 Sim3 Sim4a Sim1 Sim2 Sim3 Sim4a ClearwaterBeach(depth=3.5m;tidalrange=110cm) RMSError(cm) 6.0 17.7 16.4 15.6 15.7 15.7 16.1 19.8 18.6 18.3 18.5 18.4 18.4 MeanAbsError(cm) 5.0 14.1 13.2 12.7 12.8 12.8 13.3 15.2 19.2 13.9 14.0 14.0 14.2 MaxAbsError(cm) 15.7 47.1 45.1 41.5 41.6 41.3 43.4 56.9 57.2 50.5 56.8 57.1 54.8 Meas.SurgePeak(cm) 86 ErroratPeak 26 25 23 23 23 21 30 32 33 33 32 30 TimingatPeak(min) 88 72 72 72 73 78 82 78 78 78 75 79 StPete(depth=7.3m;tidalrange=95cm) RMSError(cm) 5.0 19.5 17.6 17.3 19.6 19.0 19.2 25.9 25.4 25.6 25.9 25.6 27.6 MeanAbsError(cm) 4.0 13.7 12.2 12.5 13.5 13.4 14.7 15.9 15.5 16.9 17.1 16.9 18.2 MaxAbsError(cm) 12.8 45.5 44.0 41.4 55.3 47.7 53.9 88.3 84.4 84.1 79.5 81.5 99.0 Meas.SurgePeak(cm) 116 ErroratPeak 33 32 24 39 31 11 55 60 59 64 60 42 TimingatPeak(min) 71 71 58 29 45 117 116 105 85 85 99 163 PortManatee(depth=0.7m;tidalrange=90cm) RMSError(cm) 5.1 16.5 14.3 11.9 13.7 13.0 16.0 22.9 21.8 21.4 22.0 21.8 23.2 MeanAbsError(cm) 4.2 11.7 10.2 9.0 9.7 9.5 12.0 14.9 14.0 14.3 14.8 14.3 15.5 MaxAbsError(cm) 12.8 42.4 42.1 31.9 37.0 34.3 47.1 78.4 75.4 76.8 69.6 73.6 86.7 Meas.SurgePeak(cm) 105 ErroratPeak 32 30 24 36 29 13 49 53 51 56 52 46 TimingatPeak(min) 84 79 65 60 60 107 118 109 90 72 99 151 AvgRMSError(cm) 5.4 17.9 16.1 14.9 16.3 15.9 17.1 22.9 21.9 21.8 22.1 21.9 22.2 Continuedonnextpage

PAGE 219

WINDGENwind Tide Sim1 Sim2 Sim3 Sim4a Sim1 Sim2 Sim3 Sim4a AvgErr.atPeak(cm) 30 29 24 33 28 15 45 48 48 51 48 39 AvgTimingError(min) 81 74 65 54 59 101 105 97 84 78 91 131

PAGE 220

BasedonRMSerrorsandaverageabsoluteerrorsshowninthetableabove,aconclusioncanbedrawnthattheWNAwindgavesignicantlybettersimulatedstormsurgevaluesthantheWINDGENwind.Theaccuracyofthesimulatedtidewasgood,withtheaverageRMSofapproximately5cm. Thesimulatedwaterelevationatallthreestationswasunderestimated,whichmighthavecomeasaresultsoftheunderestimatedWNAwindspeed.Again,evensophisticatedhurricanewindmodelsdonotdoaverygoodjobnearlandoroverestuaries,suchasTampaBay,whichresultsinanunderestimatedstormsurge.Thisseemstobetheweakestlinkingettingveryaccuratestormsurgesimulationresults. Overall,Simulation3producedbetterresultsintermsofsmallerRMSerrorsandbettercomparisonwithmeasuredwatersurfaceelevationatitspeak.Waterlevelcalculatedusingwaveenhancedbottomfrictionbasedonthe ShengandVillaret ( 1989 )formulation(Simulation4b)wasslightlyworse.Waveenhancedbottomfrictionbasedonthe GrantandMadsen ( 1979 )theory(Simulation4a)furtherreducedalreadyunderestimatedsurgewithinTampaBay:atStPeteandPortManateestations,duetoitsrelativeshallowness. The\timing"errorsareacceptable(basedonSimulation4busingWNAwind,45to73min)forallthreestations.MoreaccurateWNAwindproducedsmaller\timing"errorscomparedwithlessaccurateWINDGENwind. Table 5{18 showsthepeakwaterelevationvaluescalculatedduringthesesimulationsalongwiththemeasuredvalues.Bynormalizingthedierencebetweenconsecutivesimulationsbythepeakmeasuredwaterlevel,thepercenterroroftheeectofeachcomponentincludedinthenon-linearinteractionwascalculatedatthetimewhenthepeakwaterelevationwasobserved(e.g.,atStPete,theSimulation2value"1%relativetotheSimulation1value,andtheSimulation4aand4bvalues#13%and#6%relativetotheSimulation3value,respectively).

PAGE 221

Includingradiationstresstermsincreasedthecalculatedwaterlevelby1-2%.Itwassomewhatsurprisingtoseesuchaweakwaveeect,especiallyatClearwaterBeachwhichopenlyfacestheapproachingwaves.Forexample,theanalogouseectduringIsabelwas3-18%.Thismightbereasonedbythefactthatthemaximumwaveheightof1.4moccurredonSep-5at23:00UTCwhilethemaximumwaterelevationvaluewasobservedonSep-6at13:00UTC.Similarly,themaximumcalculatedwaveheightatStPeteandPortManateewas0.65mand0.55m,respectively,andoccurredaroundSep-5,18:00UTCwhilethecalculatedwaterelevationreacheditspeak18hourslater.ThewaveeldswerecalculatedusingtheSWANmodel.Theboundaryconditionswereobtainedfromtheregionalwavemodel,WAVEWATCH-III.Sincetherewasnowavestationinthevicinityofthecomputationaldomain,theaccuracyofthecomputedwavescouldnotbeveried. Addingwaveenhancedsurfacestresshelpedfurtherincreasethecomputedwaterelevationby2-7%.Accountingforwaveenhancedbottomfrictiondecreasedthecalculatedwaterelevationby0-13%and0-6%forSimulations4aand4b,respectively.Whenwetting-and-dryingschemewasnotengagedduringthecalculation,thepeakwaterlevelvaluegrewanextra5-7%. Table 5{19 showspeakstormsurgevalues(tidewassubtractedfromwaterelevation)calculatedduringthesesimulations.Percentincreaseordecreasewascalculatedthesamewayitwascalculatedintheprevioustable.

PAGE 222

Table5{18:MeasuredpeakwaterelevationsatthreestationsduringHurricaneFrancesusingWNAwindandvariouscombinationsofstormsurgemodelfea-tures. Sim1 Sim2 Sim3 Sim4a Meas. cm cm cm cm cm cm cm ClearwaterBch. 60 61"1% 63"2% 630% 630% 67"5% 86 StPete 83 84"1% 92"7% 77#13% 85#6% 85"7% 116 PortManatee 73 75"2% 81"6% 69#11% 76#5% 76"7% 105 Table5{19:Calculatedpeakstormsurge(withtidessubtracted)atthreestationsduringHurricaneFrancesusingWNAwindandvariouscombinationsofstormsurgemodelfeatures. Sim1 Sim2 Sim3 Sim4a Meas. cm cm cm cm cm cm cm ClearwaterBch. 53 56"4% 60"6% 600% 600% 62"3% 69 StPete 73 730% 85"13% 84#1% 84#1% 85"1% 96 PortManatee 62 620% 71"12% 70#1% 70#1% 71"1% 77

PAGE 223

UnlikeHurricanesIsabelandFranceswhenthemaximumsurgeandmaximumwaterelevationwereobservedatapproximatelythesametime,duringFrances,thesurgereacheditsmaximumapproximately4hoursafterthewaterelevationmaximumoccurred.Therefore,wave,wind,andoodconditionswereslightlydierentfromtheconditionsreectedinTable 5{18 .Wavesetup,again,hadaninsignicanteectonstormsurgeatitpeak,0-4%.Waveenhancedsurfacestressimprovedthecalculatedsurgeincreasingitby6-13%.Waveenhancedbottomstresseect(bothformulations)wasveryinsignicantdecreasingthesurgeby1%.Turningothewetting-and-dryingfeatureincreasedthesurgebyonly1-3%,whichcanbeexplainedbythefactthatmostoftheoodingoccurredafewhoursearlierduringthehightidewhenthewetting-and-dryingeectappearedtobemoresignicant,5-7%. 5{79 showsmaximumsimulatedinundationcausedbyFrancescalculatedduringSimulation3(seeTable 5{16 )usingWNAwind.Thebottomplotidentiesthetimewhenthehighestoodleveloccurred.BasedonTable 5{17 ,theWNAwindproducedbettercomparisonbetweenmeasuredandsimulatedwaterelevationatClearwaterBeach,StPete,andPortManateestations.Simulation4hadtoomuchdissipationduetowaveenhancedbottomfrictionwhichsignicantlyreducedthecalculatedwatersurfaceelevationincomparisonwithmeasuredwaterelevation.Thus,Simulation3wastakenasthebasesimulationforestimatingtheamountofinundationcausedbyFrances.Theoodedareaaected187km2ofland,mostlyalongtheeasternshoresofthebay.

PAGE 224

Figure5{79: MaximumsimulatedinundationinTampaBayusingWNAwind(toppanel).Thebottompanelshowsthetimeduringwhichthemaximumoodoccurred.

PAGE 225

FloridaDepartmentofEnvironmentalProtection(FDEP) ( 2003 ),322.3milesofstatewideshorelinearesubjecttocriticalerosionwherenaturalprocessesorhumanactivityhavecausedorcontributedtoerosionandrecessionofthebeachordunesystemtosuchadegreethatuplanddevelopment,recreationalinterests,wildlifehabitat,orimportantculturalresourcesarethreatenedorlost;and111.5milesaresubjecttonon-criticalerosionwheresignicanterosionconditionsprevail,yettheerosionprocessesdonotcurrentlythreatenpublicorprivateinterests.Stormscanonlyworsenthissituation.Accordingto MortonandSallenger ( 2003 ),extremestormsthatstrikecoastalregionscausemorphologicalresponse.Poststormerosionalresponsesincludedunescarps,channelincision,andwashouts,whereasdepositionalresponsesincludeperchedfans,washoverterraces,andsheetwashlineations.Washoverpenetrationinlandcanbeontheorderofafewhundredmetersdependingonstormintensity,stormduration,tide,waveaction,andvariationsinnearshorebathymetry. Thecurrentintegratedstormsurgemodelingsystemcanbefurtherenhancedthroughitscouplingwitharobusterosionmodel(e.g.SBEACH)topredictthepoststormmorphologicalresponses. 202

PAGE 226

theNWSemploysriskfactorsforwaveheight,wind,waveperiodandpreviousripcurrents.AteamofresearchesfromtheUniversityofFloridaconductedastudyinVolusiaCountyandanalyzedripcurrentrescuestatisticsoverasix-monthperiodandfoundthatripcurrentsweremoststronglylinkedwiththewavedirection.Accordingtotheirconclusions,thegreatestthreatassociatedwithripcurrentsoccurswhenwavesapproachatananglenearzero.Wavesapproachingina25-degreebandoneithersideofzero,althoughoccurringonlyabout30%ofthetimeduringthestudyperiod,accountedfornearly60%ofalloftheripcurrentrescuesrecordedduringthestudy.Addingawavedirectionfactortothepredictiveindexresultedinmoreaccurateripcurrentforecastswithfewerfalsealarms.Tidecanbeanimportantfactorinestimatingtheriskofoccurrenceofripcurrents.Aloworoutgoingtidecangreatlyincreasetheripcurrentrisk.Accordingto Engleetal. ( 2002 ),thefrequencyofripcurrentrescuesincreasedmarkedlyduring(1)shore-normalwaveincidence,(2)mid-lowtidalstages,(3)deepwaterwaveheightsof0.5to1.0metersand(4)waveperiodsfrom8to10seconds. OurmethodologywhichcouplesCH3D-SSMSwithawavemodel,SWAN,usinghighresolutioncurvilinear/boundary-ttedgridswillbeabletocapturetheangleatwhichwavesapproachtheshore.Neitherrectilinearnorniteelementmodelsarecapableofdoingthat.Everyshorelinesegment(350m)willcontainthefollowinginformation:waveanglewithrespecttoshorelineorientation(e.g.zeroanglemeansthewaveraysareapproachingperpendiculartotheshore),waveheight,waveperiod,andtidalstage(e.g.low,mean,high).Basedonthisinformationapredictiveripcurrentriskindexcanbecalculatedandassignedtoeachshorelinesegment.Thiswillresultinmoreaccurateripcurrentforecastswithfewerfalsealarms.

PAGE 227

Thegoalofthisstudywastodevelopastormsurgemodelingsystem,CH3D-SSMS,whichaccountsforwaveeects(e.g.,wavesetup,waveenhancedsurfacestress,waveenhancedbottomfriction)andwettinganddrying,andvalidatethemodelusingwaterelevation,wind,andwavedatacollectedduringrecenthurricaneevents(HurricaneIsabel(2003)intheOuterBanks,NCandChesapeakeBay,VA;HurricaneCharley(2004)inCharlotteHarbor,FL;andHurricaneFrances(2004)inTampaBay,FL).ThedevelopedmodelingsystemmakesuseofthewinddataproducedbyWINDGENandWNAatmosphericmodelstosimulatestormsurge.Itdynamicallycoupleslocalscalecirculation(CH3D)andwave(SWAN)models.Openboundaryconditionsforthelocalmodelsareprovidedbycouplingthelocalmodelswithregionalscalecirculation(ADCIRC)andwave(WAVEWATCH-III)models. CH3D'scomponentssuchaspressuregradientterms,radiationstressterms,andwetting-and-dryingschemewerevalidatedbycomparingmodelresultswitheitherlaboratorydata(wavesetup)oranalyticalsolutions(pressuregradientandwetting-and-dryingscheme). Inordertovalidatethemodelingsystem,threeboundary-ttedcurvilineargridsweregeneratedusinghighresolutionGEODASbathymetryandUSGSNEDtopographydata:intheOuterBanksandChesapeakeBayarea,intheTampaBayare,andtheCharlotteHarborarea. AthoroughanalysisoftheWNAandWINDGENwindwasperformedforHurricaneIsabel.TheanalysisrevealedthatresultsfrombothatmosphericmodelscomparedwellwithmeasuredwindspeedanddirectionovertheOuter 204

PAGE 228

BanksandneartheChesapeakeBaymouth,withtheWNAwindbeingslightlybetter.However,insideChesapeakeBay,theaccuracyofbothmodelssignicantlydecreased.ThisissupposedlyduetotheeectoflandsurroundingtheChesapeakeBayand/orthecoarseresolutionoftheatmosphericmodels(22-28km).Similarbutnotascomprehensive,duetothelackofavailablewinddata,analysiswasperformedforHurricanesCharleyandFrances.TheWINDGENwindofCharleywasstrongerinmagnitudeandmoreaccuratethanWNA.UsingtheWINDGENwind,theCH3Dmodelproducedmoreaccuratewaterelevationduringthestorm.ForHurricaneFrances,theWINDGENwindwasmuchweakerandlessaccuratethanWNA.UsingtheWINDGENwind,thesimulatedwatersurfaceelevationwasunderestimated. WaveparameterscalculatedbytheregionalWAVEWATCH-IIIwavemodel,whichwereusedasboundaryconditionsinthelocalwavemodel,SWAN,werevalidatedbycomparingthecalculatedwaveheightandwaveperiodwiththewaveheightandwaveperiodmeasuredfromNDBCmooredbuoysduringHurricaneIsabel.TheSWANmodel,whichprovidedthewaveinformation(waveheight,period,direction)tothelocalcirculationmodel,CH3D,wasalsoveriedusingwavetimeseriesmeasuredatFRFandVIMSfacilitiesduringIsabel. Theeectoflinearandnon-linearinteractionsbetweenstormsurge,tide,windandwavewasinvestigatedinthisstudy.Foreachhurricane,aseriesofsimulationswasconductedbyconsideringvariouscombinationsofthelinearandnon-linearstormsurgeinteractions:(1)windandtide,(2)wind,tide,andwavesetup,(3)wind,tide,wavesetup,andwaveenhancedsurfacestress,(4)wind,tide,wavesetup,waveenhancedsurfacestress,andwaveenhancedbottomstress.WhenIsabel,Charley,andFrancesweresimulatedusingtideandwindonly,thecalculatedpeakwaterelevationwasalwaysunderestimated.

PAGE 229

Theinclusionofwavesetupimprovedthecomputedstormsurgebyupto18%duringIsabel.ThemostsignicantimprovementwasobservedatBeaufortandDuckwherehighbreakingwavescausedasignicantwaterlevelsetup.Forotherhurricanes,CharleyandFrances,thewavesetupeectwaspositivebutnotverysignicant,upto5%and4%,respectively. Accountingforwaveenhancedsurfacestresshadasignicantpositiveeectduringallthreehurricanes.Theinclusionofthisfeatureincreasedthecalculatedstormsurgeby5-16%duringIsabel,9-13%duringCharley,and6-13%duringFrances. Accountingforwaveinducedbottomfrictionhad:(1)moderateeect,upto5%reductionofthestormsurgelevelatthepeak,whenthe ShengandVillaret ( 1989 )formulationwasusedand(2)signicanteect,up16%reductionofthestormsurgelevelatthepeak,duetooverestimatedbottomfrictionwhenthe GrantandMadsen ( 1979 )formulationwasused. Theeectofdynamiccouplingversuslinearlysuperimposedresultsofin-dependentlysimulatedtide,wavesetup,andsurgeistwofold:overopenwater,dynamiccouplingproducesslightlymoreaccuratestormsurge,andoverland,theinundationcalculatedthroughdynamiccouplingoccursearlierandismuchmoresignicant,however,duetothelackofdata,thiswasnotveried. Theeectofexcludingthewetting-and-dryingfeatureduringstormsurgesimulationswasalsoexamined.Thiseectmostlydependsontheextentofinundation.Forinstance,duringCharley,therewassignicantoodingcalculatedinthevicinityofEsteroBay.Whenwetting-and-dryingwasdisabled,waterelevationincreasedby18-22%inthearea,therebysignicantlyoverestimatingthecalculatedwaterlevelatthepeakofthestorm.Ontheotherhand,thecalculatedinundationduringFranceswasnotveryextensiveresultinginonly5-7%increase

PAGE 230

inwaterelevationwhenthewettinganddryingfeatureoftheCH3Dmodelwasinactivated. AninterestingphenomenonwasobservedduringthesimulationofHurricaneFrances.Themaximumstormsurgeoccurredapproximately4hoursafterthepeakwaterlevelwasobservedintheTampaBayarea.Thiswasaresultofslowly(5-10mph)movingFranceswhichcausedthesurgetolastupto30hours.Thehighestwaterleveloccurredduringhightideandwhenthetidewasalreadyreceding,themaximumsurgecamein.ThecalculatedstormsurgeduringFrancesrangedfrom0.6to0.8mbutifthepeakofthestormweretocoincidewithhightidethesurgelevelwouldhavebeen0.6to0.7mhigher. HurricaneCharleyalsomadelandfallduringoutgoingtide.ThecalculatedstormsurgewithinEsteroBaywasaround1.2m.Ifthestormweretooccurduringhightidethesurgelevelwouldhavebeen0.5to0.6mhigher.UnlikeCharleyandFrances,stormsurgeduringHurricaneIsabelhappenedduringhightide,whichonlymadethesituationworse.Thecalculatedstormsurgereached1.8minsideChesapeakeBay. StormsurgeduringIsabellasted19(overtheOuterBanks)to26hours(overtheChesapeakeBay).DuringCharley,whichstrucktheCharlotteHarborarearapidly,thesurgelastedonly11hours. Theextentofthecalculatedoodcausedbythethreehurricanesrangedwidely,from7675km2duringIsabelintheOuterBanksandChesapeakeBayto530km2duringCharleyinCharlotteHarborto183km2duringFrancesinTampaBay.EventhoughCharleywasthestrongestamongthethree,itwasthesmallesttherebyaectinglessterritorycomparedwithIsabelwhoseimpactintermsofinundationwasmoresignicant. Table 7{1 belowsummarizesthethreehurricanessimulatedinthisstudy.

PAGE 231

Table7{1:Summaryofsimulatedhurricanes. Hurricane Isabel Charley Frances Time 9/6-19,2003 8/9-14,2004 8/25-9/8,2004 Aectedarea OuterBanks, Charlotte FLEastCoast, Chesap.Bay Harbor TampaBay Category Max 5 4 4 Atlandfall 2 4 2 Wind(mph) Max 165 150 145 Atlandfall 105 145 105 Rad.ofmax Max 100 20 50 wind(km) Atlandfall 85 9 50 Atm. Min 915 941 935 pres.(mb) Atlandfall 957 942 960 Tidalstage high mean low Surgeduration(hr) 19-26 11 30 Maximumsurge(m) 4.0 1.9 2.0 Maximumwave(m) 19 3.5 4.5 Inundationarea(km2) 7,675 530 183 Damagecost$(rank $3.37B(9) $15B(2) $8.9B(4) Dominantprocesses wavesetup, waveenhanced waveenhanced waveenhanced surf.stress, surf.stress surf.stress wavesetup

PAGE 232

TheSar-SimpsonHurricaneScale TropicalStormWinds39-73mph Category1Hurricanewinds74-95mph:Norealdamagetobuildings.Damagetounanchoredmobilehomes.Somedamagetopoorlyconstructedsigns.Also,somecoastaloodingandminorpierdamage. -Examples:Irene1999andAllison1995 Category2Hurricanewinds96-110mph:Somedamagetobuildingroofs,doorsandwindows.Considerabledamagetomobilehomes.Floodingdamagespiersandsmallcraftinunprotectedmooringsmaybreaktheirmoorings.Sometreesblowndown. -Examples:Bonnie1998,Georges(FL&LA)1998andGloria1985 209

PAGE 233

Category3Hurricanewinds111-130mph:Somestructuraldamagetosmallresidencesandutilitybuildings.Largetreesblowndown.Mobilehomesandpoorlybuiltsignsdestroyed.Floodingnearthecoastdestroyssmallerstructureswithlargerstructuresdamagedbyoatingdebris.Terrainmaybeoodedwellinland. -Examples:Keith2000,Fran1996,Opal1995,Alicia1983andBetsy1965 Category4Hurricanewinds131-155mph:Moreextensivecurtainwallfailureswithsomecompleteroofstructurefailureonsmallresidences.Majorerosionofbeachareas.Terrainmaybeoodedwellinland. -Examples:Hugo1989andDonna1960 Category5Hurricanewinds156mphandup:Completerooffailureonmanyresidencesandindustrialbuildings.Somecompletebuildingfailureswithsmallutilitybuildingsblownoveroraway.Floodingcausesmajordamagetoloweroorsofallstructuresneartheshoreline.Massiveevacuationofresidentialareasmayberequired. -Examples:Andrew(FL)1992,Camille1969andLaborDay1935

PAGE 234

RMSError=s whereSandMaresimulatedandmeasuredvalues,respectively. 211

PAGE 235

ThefollowingtablesshowbesttracksforHurricanesIsabel,Charley,andFrances.TheinformationwasobtainedfromNOAAandcontainstime,lati-tude/longitudeposition,pressureinthemiddleofthestorm,maximumwindspeed,andstormstageaccordingtotheSar-SimpsonHurricaneScale. 212

PAGE 236

TableC{1:BesttrackforHurricaneIsabel,6-19Sep-tember2003. LAT LON Pressure WindSpeed Stage UTC mb mph 13.8 31.4 1009 35 tropicaldepression 06/0600 13.9 32.7 1005 40 tropicalstorm 06/1200 13.6 33.9 1003 45 " 06/1800 13.4 34.9 1000 50 " 07/0000 13.5 35.8 994 60 " 07/0600 13.9 36.5 991 70 " 07/1200 14.4 37.3 987 75 hurricane 07/1800 15.2 38.5 984 80 " 08/0000 15.8 39.7 976 90 " 08/0600 16.5 40.9 966 110 " 08/1200 17.1 42.0 952 130 " 08/1800 17.6 43.1 952 130 " 09/0000 18.2 44.1 948 135 " 09/0600 18.9 45.2 948 135 " 09/1200 19.4 46.3 948 135 " 09/1800 20.0 47.3 948 135 " 10/0000 20.5 48.3 952 130 " 10/0600 20.9 49.4 952 130 " 10/1200 21.1 50.4 948 135 " 10/1800 21.1 51.4 942 140 " 11/0000 21.2 52.3 935 145 " 11/0600 21.3 53.2 935 145 " 11/1200 21.4 54.0 925 155 " 11/1800 21.5 54.8 915 165 " 12/0000 21.6 55.7 920 160 " 12/0600 21.7 56.6 920 160 " 12/1200 21.6 57.4 920 160 " 12/1800 21.7 58.2 920 160 " 13/0000 21.8 59.1 925 155 " 13/0600 21.9 60.1 935 150 " 13/1200 22.1 61.0 935 155 " 13/1800 22.5 62.1 932 160 " 14/0000 22.9 63.3 935 155 " 14/0600 23.2 64.6 939 155 " 14/1200 23.5 65.8 935 155 " 14/1800 23.9 67.0 933 160 " 15/0000 24.3 67.9 937 150 " 15/0600 24.5 68.8 940 145 " 15/1200 24.8 69.4 946 140 " Continuedonnextpage

PAGE 237

Date/Time LAT LON Pressure WindSpeed Stage UTC mb mph 25.3 69.8 949 135 " 16/0000 25.7 70.2 952 120 " 16/0600 26.3 70.5 955 115 " 16/1200 26.8 70.9 959 110 " 16/1800 27.4 71.2 959 110 " 17/0000 28.1 71.5 957 110 " 17/0600 28.9 71.9 957 110 " 17/1200 29.7 72.5 957 105 " 17/1800 30.6 73.0 955 105 " 18/0000 31.5 73.5 953 105 " 18/0600 32.5 74.3 956 105 " 18/1200 33.7 75.2 956 105 " 18/1800 35.1 76.4 958 100 " 19/0000 36.7 77.7 969 75 " 19/0600 38.6 78.9 988 60 tropicalstorm 19/1200 40.9 80.3 997 40 extratropical 19/1800 43.9 80.9 1000 35 " 20/0000 48.0 81.0 1000 30 " 11/1800 21.5 54.8 915 165 minimumpressure 18/1700 34.9 76.2 957 105 landfallatDrumInlet,NC

PAGE 238

TableC{2:BesttrackforHurricaneCharley,9-14August2004. LAT LON Pressure WindSpeed Stage UTC mb mph 11.4 59.2 1010 35 tropicaldepression 09/1800 11.7 61.1 1009 35 " 10/0000 12.2 63.2 1009 35 " 10/0600 12.9 65.3 1007 40 tropicalstorm 10/1200 13.8 67.6 1004 45 " 10/1800 14.9 69.8 1000 50 " 11/0000 15.6 71.8 999 60 " 11/0600 16.0 73.7 999 60 " 11/1200 16.3 75.4 995 70 " 11/1800 16.7 76.8 993 75 hurricane 12/0000 17.4 78.1 992 75 " 12/0600 18.2 79.3 988 85 " 12/1200 19.2 80.7 984 90 " 12/1800 20.5 81.6 980 105 " 13/0000 21.7 82.2 976 105 " 13/0600 23.0 82.6 966 120 " 13/1200 24.4 82.9 969 110 " 13/1400 24.9 82.8 965 125 " 13/1700 25.7 82.5 954 145 " 13/1800 26.1 82.4 947 145 " 14/0000 28.1 81.6 970 85 " 14/0600 30.1 80.8 993 85 " 14/1200 32.3 79.7 988 75 " 14/1800 34.5 78.1 1000 70 tropicalstorm 15/0000 36.9 75.9 1012 45 extratropical 15/0600 39.3 73.8 1014 40 " 15/1200 41.2 71.1 1018 35 " 13/1945 26.6 82.2 941 150 minimumpressure 13/2045 26.9 82.1 942 145 landfallatPuntaGorda,FL

PAGE 239

TableC{3:BesttrackforHurricaneFrances,31August-7September2004. LAT LON Pressure WindSpeed Stage UTC mb mph 20.0 63.4 949 140 hurricane 31/1800 20.3 65.0 942 145 " 01/0000 20.6 66.3 941 140 " 01/0600 21.0 67.9 939 140 " 01/1200 21.4 69.1 937 140 " 01/1800 21.8 70.4 941 140 " 02/0000 22.3 71.4 939 140 " 02/0600 22.7 72.5 937 145 " 02/1200 23.1 73.5 939 140 " 02/1800 23.8 74.4 948 130 " 03/0000 24.2 75.1 948 120 " 03/0600 24.6 75.7 954 115 " 03/1200 25.2 76.4 958 110 " 03/1800 25.7 77.2 960 105 " 04/0000 25.9 77.5 960 100 " 04/0600 26.4 78.0 960 100 " 04/1200 26.7 78.4 962 105 " 04/1800 26.9 79.0 962 105 " 05/0000 27.0 79.4 958 110 " 05/0600 27.2 80.5 960 105 " 05/1100 27.3 80.7 967 100 " 05/1300 27.5 80.9 969 90 " 05/1500 27.7 81.2 973 80 " 05/1800 27.9 81.7 975 70 tropicalstorm 05/2100 28.0 82.2 977 70 " 06/0000 28.1 82.3 978 65 " 06/0600 28.6 83.8 981 65 " 06/1200 29.1 83.6 982 65 " 06/1800 30.1 84.1 982 65 " 07/0000 31.1 84.5 984 40 " 01/0700 21.1 68.1 935 140 minimumpressure 31/1800 20.3 65.0 942 145 maximumwind 05/0430 27.2 80.2 960 105 landfallatPortStLucie,FL 06/1800 30.1 84.0 982 60 landfallnearAucillaRiver,FL

PAGE 240

217

PAGE 241

FigureD{1: WINDGENandWNAvs.measuredwindspeed(top)anddirection(bottom)atCapeLookout,NCduringHurricaneIsabel.

PAGE 242

FigureD{2: WINDGENandWNAvs.measuredwindspeed(top)anddirection(bottom)atDuck,NCduringHurricaneIsabel.

PAGE 243

FigureD{3: WINDGENandWNAvs.measuredwindspeed(top)anddirection(bottom)atChesapeakeLight,VAduringHurricaneIsabel.

PAGE 244

FigureD{4: WINDGENandWNAvs.measuredwindspeed(top)anddirection(bottom)atChesapeakeBayBridge,VAduringHurricaneIsabel.

PAGE 245

FigureD{5: WINDGENandWNAvs.measuredwindspeed(top)anddirection(bottom)atKiptopeke,VAduringHurricaneIsabel.

PAGE 246

FigureD{6: WINDGENandWNAvs.measuredwindspeed(top)anddirection(bottom)atMoneyPoint,VAduringHurricaneIsabel.

PAGE 247

FigureD{7: WINDGENandWNAvs.measuredwindspeed(top)anddirection(bottom)atGloucesterPoint,VAduringHurricaneIsabel.

PAGE 248

FigureD{8: WINDGENandWNAvs.measuredwindspeed(top)anddirection(bottom)atLewisetta,VAduringHurricaneIsabel.

PAGE 249

FigureD{9: WINDGENandWNAvs.measuredwindspeed(top)anddirection(bottom)atHPLWS,VAduringHurricaneIsabel.

PAGE 250

FigureD{10: WINDGENandWNAvs.measuredwindspeed(top)anddirection(bottom)atChoptankRiver,VAduringHurricaneIsabel.

PAGE 251

FigureD{11: WINDGENandWNAvs.measuredwindspeed(top)anddirection(bottom)atNorthBay,VAduringHurricaneIsabel.

PAGE 252

FigureE{1: ComputationalgridnearChesapeakeBaymouth. 229

PAGE 253

FigureE{2: ComputationalgridintheSouthOuterBanksarea.

PAGE 254

TableF{1:Alistofsimulationswithvariouscombi-nationsofsixmodelfeatures(psymboldenotesthefeaturewasincludedduringthesimulation). Factors Sim1 Sim2 Sim3 Sim4a Tide p p p p p p p p p p p p p p p p p p p p p ( 1993 )formulationwasused2 ( 1979 )formulationwasused3 ( 1989 )formulationwasused 231

PAGE 255

FigureF{1: Comparisonofsimulated(top-usingWNAwind,bottom-usingWINDGENwind)vs.measuredwaterelevationatBeaufort,NC.

PAGE 256

FigureF{2: Comparisonofsimulated(top-usingWNAwind,bottom-usingWINDGENwind)vs.measuredwaterelevationatDuck,NC.

PAGE 257

FigureF{3: Comparisonofsimulated(top-usingWNAwind,bottom-usingWINDGENwind)vs.measuredwaterelevationatChesapeakeBayBridge,VA.

PAGE 258

FigureF{4: Comparisonofsimulated(top-usingWNAwind,bottom-usingWINDGENwind)vs.measuredwaterelevationatGloucesterPoint,VA.

PAGE 259

FigureF{5: Comparisonofsimulated(top-usingWNAwind,bottom-usingWINDGENwind)vs.measuredwaterelevationatMoneyPoint,VA.

PAGE 260

FigureF{6: Comparisonofsimulated(top-usingWNAwind,bottom-usingWINDGENwind)vs.measuredwaterelevationatKiptopeke,VA.

PAGE 261

FigureF{7: Comparisonofsimulated(top-usingWNAwind,bottom-usingWINDGENwind)vs.measuredwaterelevationatLewisetta,VA.

PAGE 262

I )comparedwiththesimulatedsurge(simulated"pure"tidewassubtractedformsimulatedwaterelevation)atsevenlocationsintheOuterBanksandChesapeakeBayareasduringHurricaneIsabel.ThesimulatedresultsarebasedonSimulation3usingWNAwind. FigureF{8: Comparisonofsimulatedvs.measuredstormsurgeelevationatBeau-fort,NC.CalculatedresultsarebasedonSimulation3usingWNAwind.

PAGE 263

FigureF{9: Comparisonofsimulatedvs.measuredstormsurgeelevationatDuck,NC.CalculatedresultsarebasedonSimulation3usingWNAwind. FigureF{10: Comparisonofsimulatedvs.measuredstormsurgeelevationatChesapeakeBayBridge,VA.CalculatedresultsarebasedonSimulation3usingWNAwind.

PAGE 264

FigureF{11: Comparisonofsimulatedvs.measuredstormsurgeelevationatGloucesterPoint,VA.CalculatedresultsarebasedonSimulation3usingWNAwind. FigureF{12: Comparisonofsimulatedvs.measuredstormsurgeelevationatMoneyPoint,VA.CalculatedresultsarebasedonSimulation3usingWNAwind.

PAGE 265

FigureF{13: Comparisonofsimulatedvs.measuredstormsurgeelevationatKip-topeke,VA.CalculatedresultsarebasedonSimulation3usingWNAwind. FigureF{14: Comparisonofsimulatedvs.measuredstormsurgeelevationatLe-wisetta,VA.CalculatedresultsarebasedonSimulation3usingWNAwind.

PAGE 266

InthissectionmeasuredwaterelevationrecordedatfourlocationsintheCharlotteHarborareaduringHurricaneCharleyisshownversuswatersurfaceelevationcalculatedbasedonWNAorWINDGENwindandvariousforcingmechanismslistedinthetablebelow. TableG{1:Alistofsimulationswithvariouscombi-nationsofsixmodelfeatures(psymboldenotesthefeaturewasincludedduringthesimulation). Factors Sim1 Sim2 Sim3 Sim4a Tide p p p p p p p p p p p p p p p p p p p p p ( 1993 )formulationwasused2 ( 1979 )formulationwasused3 ( 1989 )formulationwasused 243

PAGE 267

FigureG{1: Comparisonofsimulated(top-usingWNAwind,bottom-usingWINDGENwind)vs.measuredwaterelevationatBigCarlosPass,FL.

PAGE 268

FigureG{2: Comparisonofsimulated(top-usingWNAwind,bottom-usingWINDGENwind)vs.measuredwaterelevationatEsteroBay#1,FL.

PAGE 269

FigureG{3: Comparisonofsimulated(top-usingWNAwind,bottom-usingWINDGENwind)vs.measuredwaterelevationatEsteroBay#2,FL.

PAGE 270

FigureG{4: Comparisonofsimulated(top-usingWNAwind,bottom-usingWINDGENwind)vs.measuredwaterelevationatFtMyers,FL.

PAGE 271

TableH{1:Alistofsimulationswithvariouscombi-nationsofsixmodelfeatures(psymboldenotesthefeaturewasincludedduringthesimulation). Factors Sim1 Sim2 Sim3 Sim4a Tide p p p p p p p p p p p p p p p p p p p p p ( 1993 )formulationwasused2 ( 1979 )formulationwasused3 ( 1989 )formulationwasused 248

PAGE 272

FigureH{1: Comparisonofsimulated(usingWNAwind)vs.measuredwaterele-vationatClearwater,FL.Calculatedresultsarebasedonvesimulations.

PAGE 273

FigureH{2: Comparisonofsimulated(usingWNAwind)vs.measuredwaterele-vationatStPete,FL.Calculatedresultsarebasedonvesimulations. FigureH{3: Comparisonofsimulated(usingWNAwind)vs.measuredwaterele-vationatPortManatee,FL.Calculatedresultsarebasedonvesimulations.

PAGE 274

I )comparedwiththesimulatedsurge(simulated"pure"tidewassubtractedformsimulatedwaterelevation)atthreelocationsintheTampaBayareaduringHurricaneFrances.ThesimulatedresultsarebasedonSimulation3usingWNAwind. FigureH{4: Comparisonofsimulatedvs.measuredstormsurgeelevationatClear-water,FL.CalculatedresultsarebasedonSimulation3usingWNAwind.

PAGE 275

FigureH{5: Comparisonofsimulatedvs.measuredstormsurgeelevationatStPete,FL.CalculatedresultsarebasedonSimulation3usingWNAwind. FigureH{6: Comparisonofsimulatedvs.measuredstormsurgeelevationatPortManatee,FL.CalculatedresultsarebasedonSimulation3usingWNAwind.

PAGE 276

ThislowpassltercodewritteninMatlabwasusedtolteroutnon-tidalcontributiontowatersurfaceelevation,e.g.stormsurge. functionfdata=lplt(measured data,cuto f) %Peformslow-passlterbymultiplicationinfrequencydomain.Usesthree-pointtaper(infrequencyspace)betweenpass-bandandstopband. %CoecientssuggestedbyD.Coats,Battelle,Ventura.Evenbettercoef-cientsforthetapercanbefoundin:Rabiner,L.R.,Gold,B.,andMcGonegal,C.A.(1980).AnApproachtotheapproximationproblemfornonrecursivedigitallters.IEEETran.vol.AU-18(2):83-106.WrittenbyChrisSherwood,BattellePNLModiedbyVadimAlymov:6/8/2005 closeall %dataleconsistsoftwocolumns:rstistimeinjuliandays,secondiswaterelevation time=measured data(:,1); data=measured data(:,2); delta t=(time(2)-time(1))*24.0*60.*60.%juliandayconvertedtosec. n=length(data) mn=mean(data);data=data-mn;P=t(data);N=length(P);lt=ones(N,1);k=oor(cuto f*N*delta t);lt(1:k)=lt(1:k)*1;lt(k+1)=.715;lt(k+2)=.24;lt(k+3)=.024;lt(k+4:N-(k+4))=lt(k+4:N-(k+4))*0.; lt(N-(k+3))=.024;lt(N-(k+2))=.24;lt(N-(k+1))=.715; P=P.*lt; 253

PAGE 277

fdata=real(it(P)); surge=fdata(1:n)+mn; tide=data-surge; gure(1);plot(time,tide,'g-') gure(2);plot(time,surge,'r-') measured surge(:,:)=[time,surge]; measured tide(:,:)=[time,tide]; savemeasured surge.datmeasured surge-ASCII savemeasured tide.datmeasured tide-ASCII

PAGE 278

Arya,S.P.S.(1977).Suggestedrevisionstocertainboundary-layerparameter-izationschemesusedinatmosphericcirculationmodels.Mon.WeatherRev.,105:215{227. Bakker,W.T.andDorn,T.V.(1978).Near-bottomvelocitiesinwaveswithacurrent.Proc.16thInt.Conf.CoastalEng.,ASCE,2:1394{1413. Battjes,J.A.(1975).Modelingofturbulenceinthesurf-zone.SymposiumonmodelingTechniques.AmericanSocietyofCivilEng.,11:1050{1061. Battjes,J.A.andJanssen,J.P.F.M.(1978).Energylossandset-upduetobreakingofrandomwaves.Proc.16thInt.Conf.CoastalEng.,ASCE,1:569{587. Blumberg,A.F.andMellor,G.L.(1987).Adescriptionofathree-dimensionalcoastaloceancirculationmodel.Three-DimensionalCoastalOceanModels,editedbyN.Heaps,AmericanGeophysicalUnion. Bode,L.andHardy,T.A.(1997).Progressandrecentdevelopmentsinstormsurgemodeling.J.ofHydraulicEng.,123/4:315{331. Booij,N.,Ris,R.C.,andHolthuijsen,L.H.(1999).Athird-generationwavemodelforcoastalregions-1.modeldescriptionandvalidation.J.ofGeophys.Res.-Oceans,104:7649{7666. Cardone,V.J.,Greenwood,C.V.,andGreenwood,J.A.(1992).Unitedprogramforthesecicationofhurricaneboundarylayerwindsoversurfacesofspeciedroughness,ContractReportCERC-92-1,CorpsofEngineers,Vicksburg,MS. Carrier,J.R.andGreenspan,H.P.(1958).Waterwavesofniteamplitudeonaslopingbeach.JournalofFluidMechanics,4:97{109. Casulli,V.andCheng,R.T.(1992).Semi-implicitnitedierencemethodsforthree-dimensionalshallowwaterow.Int.J.forNumericalMethodsinFluids,15:629{648. Chow,S.(1971).Astudyofthewindeldintheplanetoryboundarylayerofamovingtropicalcyclone.Master'sThesis,NewYorkUniversity,NewYork. Cialone,M.A.(1991).CoastalModelingSystem(CMS)user'smanual.CERC-91-1,CERC/WES,CorpsofEngineers,Vicksburg,MS. 255

PAGE 279

Clark,R.R.andLaGrone,J.W.(2004).Hurricanecharley,august2004:Post-stormbeachconditions&coastalimpactandrecommendationsforrecovery&modicationsofbeachmanagementstrategies.FloridaDepartmentofEnviron-mentalProtectionBureauofBeachesandCoastalSystems,Report. Davis,J.R.(1996).Hydrodynamicmodelinginshallowwaterwithwettinganddrying.Master'sThesis,CoastalandOceaonographicEngineeringDepartment,UniversityofFlorida,Gainesville. Davis,J.R.(2001).HighperformancemodelingofcirculationandtransportintheIndianRiverLagoon,orida.Dissertation,CoastalandOceaonographicEngineeringDepartment,UniversityofFlorida,Gainesville. Davis,J.R.andSheng,Y.P.(2000).Highperformanceestuarineandcoastalenvironmentalmodeling:TheCH3Dexample.EstuarineandCoastalModeling,6:470{484. Davis,J.R.andSheng,Y.P.(2002).Highperformanceestuarineandcoastalenvironmentalmodeling:PartII.EstuarineandCoastalModeling,7:479{490. Dean,R.G.,Chiu,T.Y.,andWang,S.Y.(1995).CombinedtotalstormtidefrequencyanalysisforPinellasCountyFlorida.Tallahassee,FL:BeachesandShoresResourceCenterInstituteofScienceandPublicAairs. Dean,R.G.andDalrymple,R.A.(1991).Waterwavemechanicsforengineersandscientists.AdvancedSeriesonOceanEng.,Vol.2SO. Donelan,M.A.,Dobson,F.W.,Smith,S.D.,andAnderson,R.J.(1993).Onthedependenceofsea-surfaceroughnessonwavedevelopment.J.ofPhys.Oceanogr.,23:2143{2149. Doodson,A.T.andWarburg,H.D.(1941).Admiraltymanualoftides.Hydrogra-phyDepartment,HerMajesty'sStationaryOce,London. Engle,J.,MacMahan,J.,Thieke,R.J.,Hanes,D.M.,andDean,R.G.(2002).Formulationofaripcurrentpredictiveindexusingrescuedata.Proc.NationalConf.onBeachPreservationTechnology,FSBPA,Biloxi,MS. FederalEmergencyManagementAgency(FEMA)(1988).Waveheightanalysisforoodinsurancestudies.TechnicalDocumentationforWHAFISProgramVersion3.0. FloridaDepartmentofEnvironmentalProtection(FDEP)(2003).Criticalbeacherosionareasinorida.BureauofBeachesandWetlandResources,ReportNO.BCS-99-02. Foreman,M.G.G.(1977).Manualfortidalheightsanalysisandprediction.PacicMarineScienceReport77-10,InstituteofOceanSciences,PatriciaBay,Sidney,B.C.

PAGE 280

Garratt,J.R.(1977).Reviewofdragcoecientsoveroceansandcontinents.MonthlyWeatherReview,105:915{929. Graber,H.C.andMadsen,O.S.(1988).Anite-depthwind-wavemodel.1.modeldescription.J.ofPhys.Oceanogr.,18:1465{1483. Grant,W.D.andMadsen,O.S.(1979).Combinedwaveandcurrentinteractionwitharoughbottom.J.ofGeophys.Res.,84:1797{1808. Grant,W.D.andMadsen,O.S.(1982).Movablebedroughnessinunsteadyocillatoryow.J.ofGeophys.Res.,87:469{481. Groves,G.W.(1955).Numericalltersfordiscriminationagainsttidalperiodici-ties.Transactions,AmericanGeophisicalUnion,36(6):1073{1084. Holland,G.J.(1980).Ananalyticmodelofthewindandpressureprolesinhurricanes.MonthlyWeatherReview,108:1212{1218. Holthuijsen,L.H.,Booij,R.,Haagsma,I.G.,Kieftenburg,A.T.M.M.,andKriezi,E.E.(2000).SWANCycle3version40.11:UserManual.DepartmentofCivilEngineeringDelftUniversityofTechnology. Hubbert,G.D.andMcInnes,K.L.(1999).Astormsurgeinundationmodelforcoastalplanningandimpactstudies.J.ofCoastalRes.,15-1:168. Janssen,P.A.E.M.(1991).Quasi-lineartheoryofwind-wavegenerationappliedtowaveforecasting.J.ofPhys.Oceanogr.,21:1631{1642. Jelesnianski,C.P.,Chen,J.,andShaer,A.W.(1992).SLOSH:Sea,Lake,andOverlandSurgesfromHurricanes.TechnicalReportofNationalWeatherService,SilverSpring,MD,48:71. Johnson,H.K.andKofoed-Hansen,H.(2000).Inuenceofbottomfrictiononseasurfaceroughnessanditsimpactonshallowwaterwindwavemodeling.J.ofPhys.Oceanogr.,30(7):1743{1756. Jonsson,I.G.andCarlsen,N.A.(1979).Experimentalandtheoreticalinvestiga-tionsinanoscillatoryturbulentboundarylayer.J.ofHydraulicRes.,14:45{60. Kaihatu,J.M.andSheremet,A.(2004).Dissipationofwaveenergybycohesivesediments.CoastalEng.2004,Proc.29thInt.Conf.,Lisbon,Portugal,1:498{507. Kirby,J.T.andDalrymple,R.A.(1994).REF/DIF1version:DocumentationandUser'sManual.DepartmentofCivilEngineering.UniversityofDelaware,Newark. Liu,P.L.F.andDalrymple,R.A.(1978).Bottomfrictionalstressesandlongshorecurrentsduetowaveswithlargeanglesofincidence.J.ofMarineRes.,36:357{375.

PAGE 281

Longuet-Higgins,M.S.andStewart,R.W.(1964).Radiationstressesinwaterwaves;aphysicaldiscussion,withapplications.Deep-SeaResearch,11:529{562. Luettich,R.,Westerink,J.J.,andSchener,N.W.(1992).ADCIRC:anadvancedthree-dimensionalcirculationmodelforshelvescoastsandestuaries,report1:theoryandmethodologyofADCIRC-2DDIandADCIRC-3DL.DredgingResearchProgramTechnicalReportDRP-92-6,U.S.ArmyEngineersWaterwaysExperimentStation,Vicksburg,MS,page137. Mastenbroek,C.,Burgers,G.,andJanssen,P.A.E.M.(1993).Thedynamicalcouplingofawavemodelandastorm-surgemodelthroughtheatmosphericboundary-layer.J.ofPhys.Oceanogr.,23:1856{1866. Mellor,G.(2003).Thethree-dimensionalcurrentandsurfacewaveequation.J.ofPhys.Oceanogr.,33:1978{1989. Moon,I.(2000).Develpmentofacoupledoceanwave-circulationmodelanditsapplicationstonumericalexperimentsforwindwaves,stormsurgesandoceancirculationoftheYellowandEastChinaSeas.Ph.D.Dissertation,DepartmentofOceanography,SeoulNationalUniversity. Moon,I.(2005).Impactofacoupledoceanwave-tide-circulationsystemonocastalmodeling.OceanModelling,8:203{236. Morton,R.A.andSallenger,A.H.(2003).Morphologicalimpactsofextremestormsonsandybeachesandbarriers.J.ofCoastalRes.,19:560{573. Mory,M.andHamm,L.(1997).Waveheight,setupandcurrentsaroundadetachedbreakwatersubmittedtoregularorrandomwaveforcing.CoastalEng.,31:77{96. Mukai,A.Y.,Westerink,J.J.,andLuettich,R.A.(2001).Guidelinesforusingtheeastcoast2001databaseoftidalconstituentswithintheWesternNorthAtlanticOcean,GulfofMexicoandCaribbeanSea.CoastalandHydraulicEng.TechnicalNote(IV-XX). NationalOceanicandAtmosphericAdministration(NOAA)(1999).HurricaneBasics,Brochure. NationalOceanicandAtmosphericAdministration(NOAA)(2005).EconomicstatisticsforNOAA.Report,U.S.DepartmentofCommerce. Ochi,M.K.(1998).Probabilitydistributionofheightinnitewaterdepth.CoastalEng.1998,Proc.26thInt.Conf.,Copenhagen,Denmark. Peene,S.,Sheng,Y.P.,andHouston,S.(1993).ModelingtidalandwinddrivencirculationinSarasotaandTampaBay.EstuarineandCoastalModeling,2:357{369.

PAGE 282

Rogers,W.E.(2005).Validationofthreesub-regional-scale,wave-forecastingsystemscreatedbytheNavalResearchLaboratoryfortheNationalWeatherSer-viceCoastalStormsProgram.Report,OceanographyDivision,NavalResearchLaboratory,StennisSpaceCenter,MS. Schoellhamer,D.H.(1993).SimulationandanalysisofsedimentresuspensionobservedinOldTampaBay,Florida.Dissertation,CoastalandOceanographicEngineeringDepartment,UniversityofFlorida,Gainesville. Schwerdt,R.W.,Ho,F.P.,andWatkins,R.R.(1979).Meteorologicalcriteriaforstandardprojecthurricaneandprobablemaximumhurricanewindelds,GulfandEastCoastsoftheUnitedStates.NOAATechnicalReportNWS23,SilverSpring,MD. Sheng,Y.P.(1982).Hydraulicapplicationsofasecond-orderclosuremodelofturbulenttransport.Proc.Conf.ApplyingResearchtoHydraulicPracticeASCE,Jackson,MS. Sheng,Y.P.(1983).Mathematicalmodelingofthree-dimensionalcoastalcurrentsandsedimentdispersion:Modeldevelopmentandapplication.TechnicalReportCERC-83-2,AeronauticalResearchAssociatesofPrinceton,Princeton,NJ. Sheng,Y.P.(1986).Athree-dimensionalnumericalmodelofcoastalandestuarinecirculationandtransportingeneralizedcurvilineargrids.TechnicalReportNo.587. Sheng,Y.P.(1987).Onmodelingthree-dimensionalestuarineandmarinehydrodynamics.Three-dimensionalmodelsofmarineandestuarinedynamics,pages35{54. Sheng,Y.P.(1990).Evolutionofathree-dimensionalcurvilinear-gridhydro-dynamicmodelforestuaries,lakesandcoastalwaters:CH3D.EstuarineandCoastalModeling,pages40{49. Sheng,Y.P.andAlymov,V.(2002).CoastaloodinganalysisofpinellascountyusingALSMdata:AcomparisonbetweenUFs2-Dmethodandresultsvs.FEMAsmethodandresults.DepartmentofCivilandCoastalEngineeringUniversityofFlorida,Gainesville. Sheng,Y.P.,Alymov,V.,Paramygin,V.,andDavis,J.R.(2004).Anintegratedstormsurgemodelingsystem:CH3D-SSMS.DepartmentofCivilandCoastalEngineering,UniversityofFlorida,Gainesville. Sheng,Y.P.andVillaret,C.(1989).Modelingtheeectofsuspendedsedimentstraticationonbottomexchangeprocess.J.ofGeophys.Res.,pages14229{14444.

PAGE 283

Sheng,Y.P.,Zhang,Y.,Alymov,V.,andLee,J.(2002).Implementationofwetting-and-dryingalgorithmtoCH3D.TechnicalReport,DepartmentofCoastalandOceanographicEngineering,UniversityofFlorida,Gainesville. Signell,R.P.,Beardsley,R.C.,Graber,H.C.,andCapotondi,A.(1990).Eectofwave-currentinteractiononsteadywind-drivencirculationinnarrow,shallowembayments.J.ofGeophys.Res.,95:9671{9678. Signell,R.P.andList,J.H.(1997).Eectofwave-enhancedbottomfrictiononstorm-drivencirculationinMassachusettsBay.J.ofWaterway,Port,Coastal,andOceanEng.,123:233{239. Smith,S.D.andBanke,E.G.(1975).Variationofsea-surfacedragcoecientwithwindspeed.Q.J.R.Meteorol.Soc.,101:665{673. Stive,M.J.F.andWind,H.G.(1982).Astudyofradiationstressandset-upinthenearshoreregion.CoastalEng.,6:1{25. Sun,D.(2001).ModelingsuspendedsedimenttransportundercombinedwavecurrentactionsinIndianRiverLagoon.Dissertation,CoastalandOceangraphicEngineeringDepartment,UniversityofFlorida,Gainesville. Sun,D.andSheng,Y.P.(2002).Modelingthree-dimensionalwind-inducedcirculation.J.ofWaterway,Port,Coastal,andOceanEng. Tang,Y.M.andGrimshaw,R.(1996).Theeectofwind-waveenhancementofbottomstressonthecirculationinducedbytropicalcyclonesoncontinentalshelves.J.ofGeophys.Res.-Oceans,101:22705{22714. Tolman,H.L.(1997).UsermanualandsystemdocumentationofWAVEWATCH-IIIversion1.15.NOAA/NWS/NCEP/OMBTechnicalNote151,page97. Tolman,H.L.(1999).UsermanualandsystemdocumentationofWAVEWATCH-IIIversion1.18.NOAA/NWS/NCEP/OMBTechnicalNote166,page110. Vriend,D.andStive,M.J.F.(1987).Quasi-3dmodelingofnear-shorecurrents.CoastalEng.,11:565{601. Wang,Y.H.,Bohlen,W.F.,andO'Donnel,J.(2000).Stormenhancedbottomshearstressandassociatedsedimententrainmentinamoderateenergeticestuary.J.ofOceanogr.,56:311{317. Watson,C.C.(1995).Thearbiterofstorms:ahighresolution,GISbasedstormhazardmodel.NationalWeatherDigests,20:2{9.

PAGE 284

Watson,C.C.andJohnson,M.E.(1999).Design,implementationandoperationofamodularintegratedtropicalcyclonehazardmodel.Proc.23rdConf.onHurricanesandTropicalMeteorology,Dallas,TX. Weaver,R.J.(2004).Eectofwaveforcesonstormsurge.Master'sThesis,DepartmentofCivilandCoastalEngineering,UniversityofFlorida,Gainesville. Wilson,B.W.(1957).HurricanewavestatisticsfortheGulfofMexico.Depart-mentofArmyCorpsofEngineers.TechnicalMemorandum98. Wornom,S.F.,Welsh,D.J.S.,andBedford,K.W.(2001).OncouplingtheSWANandWAMwavemodelsforaccuratenearshorewavepredictions.CoastEng.,43:161{201. Xie,L.,Wu,K.,Pietrafesa,L.,andZhang,C.(2001).Anumericalstudyofwave-currentinteractionthroughsurfaceandbottomstresses:Wind-drivencirculationintheSouthAtlanticBightunderUniformWinds.J.ofGeophys.Res.,106:16,841{16,855. Zhang,M.Y.andLi,Y.S.(1996).Thesynchronouscouplingofathird-generationwavemodelandatwo-dimensionalstormsurgemodel.OceanEng.,23:533{543.

PAGE 285

VadimAlymovwasbornonMarch25th,1974,inAndijan,Uzbekistan.Attheageof11,hemovedwithhisparentstoBarnaul,Russia,wherehenishedhighschoolandenteredtheDepartmentofMathematicsattheAltaiStateUniversity.AfterreceivingaM.S.degreeinappliedmathematicsin1996,hestartedtoworkattheInstituteforWaterandEnvironmentalProblemsoftheRussianAcademyofSciences,inBarnaul.In1997,hejoinedtheDepartmentofCoastalandOceanographicEngineeringattheUniversityofFloridawhereheearnedhisM.S.andPh.D.incoastalengineering,in1999and2005,respectively. 262