Numerical study of coastal hydrodynamics using a coupled model for Hudhud cyclone in the Bay of Bengal

Accepted Manuscript Numerical study of coastal hydrodynamics using a coupled model for Hudhud cyclone in the Bay of Bengal P.L.N. Murty, Prasad K. Bhaskaran, R. Gayathri, Bishnupriya Sahoo, T. Srinivasa Kumar, B. SubbaReddy PII:

S0272-7714(16)30459-0

DOI:

10.1016/j.ecss.2016.10.013

Reference:

YECSS 5272

To appear in:

Estuarine, Coastal and Shelf Science

Received Date: 6 July 2016 Revised Date:

6 October 2016

Accepted Date: 10 October 2016

Please cite this article as: Murty, P.L.N., Bhaskaran, P.K., Gayathri, R., Sahoo, B., Srinivasa Kumar, T., SubbaReddy, B., Numerical study of coastal hydrodynamics using a coupled model for Hudhud cyclone in the Bay of Bengal, Estuarine, Coastal and Shelf Science (2016), doi: 10.1016/j.ecss.2016.10.013. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Numerical Study of Coastal Hydrodynamics using a Coupled Model for Hudhud Cyclone in the Bay of Bengal

22

ABSTRACT

23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46

The past decade has witnessed an increased intensity of cyclones in the Bay of Bengal region. With higher winds spread over a larger area, the associated risk and coastal vulnerability have increased with wider destructive potential from high waves, storm surges, and associated coastal inundation. The very severe cyclones that made landfall over the Bay of Bengal in the past decade had strong winds in their outer cores, unlike those cyclones that made landfall in previous years. The original parametric wind formulation performs well for more compact cyclones, but at a radial distance far away from the cyclone centre, the winds are under-estimated. Hence, there is a need to revisit and modify this formula for practical applications, and this study attempts to provide a better representation of the overall radial distance in the wind field envelope. The study postulates a 3/5-power law fitted to the original wind formulae, which provides a reasonably good estimate for the surface wind field. The recent very severe cyclones that developed over the Bay of Bengal provided an excellent test-bed to verify this hypothesis, which is supported by validation from six in-situ buoys. The modified wind formula used with a coupled hydrodynamic model (ADCIRC+SWAN) simulated the storm surge and wave characteristics associated with a recent very severe cyclonic storm 'Hudhud' that made landfall in Andhra, located on the east coast of India in 2014. The study also investigated the dependence of coastal geomorphic features and beach slope on the variability of wave-induced setup. Computed significant wave height and storm surge show an excellent match with wave-rider buoy and tide gauge observations.

Earth System Science Organisation (ESSO)- Indian National Centre for Ocean Information Services (INCOIS) Hyderabad-500 090, India 2

Department of Ocean Engineering and Naval Architecture Indian Institute of Technology Kharagpur Kharagpur-721 302, India

Email: [email protected], [email protected] Tel: +91-3222-283772, Fax: +91-3222-255303

AC C

EP

TE D

*

National Centre for Sustainable Coastal Management Chennai-600 025, India

M AN U

3

SC

1

RI PT

Murty P L N1, Prasad K. Bhaskaran*,2, Gayathri R2, Bishnupriya Sahoo2, Srinivasa Kumar T1 and SubbaReddy B3

Keywords: Modified Jelesnianski Winds, Coupled Model, Storm Surge, Significant Wave Height, Wave Setup, Hudhud Cyclone

1

ACCEPTED MANUSCRIPT

1. Introduction

50

Tropical cyclones that generate storm surges, extreme wind-waves, and coastal flooding

51

during landfall cause a major threat to human life, property, damage to the ecosystem,

52

and infrastructure. The destruction depends mainly on cyclone intensity, the maximum

53

radius of curvature of high winds, the location of landfall, and the resulting storm surge

54

and coastal flooding causes catastrophic damage to the coastal community. Hence, there

55

is a need and necessity to model tropical cyclones and understand the role of underlying

56

coastal mechanisms for proper planning and evacuation measures. The total water level

57

elevation near the coast is a cumulative effect resulting from astronomical tides, storm

58

surges, wind-waves, wave induced setup/set-down, and sea-level rise.

M AN U

SC

RI PT

47 48 49

59

The East coast of India located in the North Indian Ocean basin is highly vulnerable to

61

tropical cyclones and coastal flooding. The frequency of cyclones is almost five times

62

higher in the Bay of Bengal compared to the Arabian Sea (Sahoo and Bhaskaran, 2015).

63

The recent decade exhibits a paradigm shift in the frequency of high intense hurricanes

64

and cyclonic storms over the global ocean basins, compared to the past few decades.

65

Emanuel (1987, 2005) examined the dependence of cyclone intensity with climate

66

change, and proposed the Power Dissipation Index (PDI) directly linked with the strength

67

of a cyclone. A recent study on tropical cyclones for the North Indian Ocean basin clearly

68

revealed a manifold increase in PDI (Sahoo and Bhaskaran, 2015). In addition, the

69

present decade also showed an increased intensity and maximum radius of curvature for

70

tropical cyclones. Over the Bay of Bengal region, the cyclones that developed and made

AC C

EP

TE D

60

2

ACCEPTED MANUSCRIPT

landfall had strong winds in the outer core that were unlikely seen for cyclones during the

72

1990s. Hence, there is a need to revisit and modify the widely accepted and used

73

parametric wind models such as Jelesnianski and Holland formulations for practical

74

applications (Jelesnianski and Taylor, 1973; Holland, 1980). The original wind formula

75

needs an appropriate correction for the radial distance of maximum wind speed to

76

provide better input to the hydrodynamic models. It should take into account the

77

maximum radius of curvature of the cyclone over the Indian Ocean basin. The overall

78

decay is exponential in nature from Rmax (radius of maximum winds) towards the outer

79

core as in the original Jelesnianski and Taylor formulation (JT), however, the radial

80

distance of strong wind bands have increased for cyclones in the present decade. A study

81

with 'Phailin' cyclone clearly indicates strong winds in the outer core of the cyclone, as

82

evidenced by tide gauge observation located approximately 280 km (in the outer core)

83

from the landfall location. By using the original JT formula, computed storm surge was

84

under-estimated by nearly 0.3 m. In addition, the wind speed near the tide gauge location

85

using this formulation was also under-estimated about 10 m s-1. Storm surge

86

computations for compact level cyclones as noticed in the past were good enough with

87

the original JT formula. With higher winds spread over a larger area (as seen in present

88

cyclones) the risk and vulnerability of the coastal region to storm surge and associated

89

inundation have increased. Therefore, the authors emphasize that the modified wind

90

formula is useful for surge forecast in low-lying areas influenced by higher winds in the

91

outer core of the cyclone. Accurate representation of the surface wind field envelope and

92

its variability is critical for realistic estimates of storm surge and associated coastal

AC C

EP

TE D

M AN U

SC

RI PT

71

3

ACCEPTED MANUSCRIPT

inundation. Therefore, it is imperative, very crucial, and inevitable for effective and

94

reliable warning system in operational weather centers.

95

The Earth System Science Organization (ESSO) – Indian National Centre for Ocean

96

Information Services (INCOIS) located at Hyderabad, and under the Ministry of Earth

97

Sciences, Government of India provides operational marine weather forecasting service

98

for the Indian seas. In the realm of hydrodynamic and wave modeling efforts and recent

99

developments in computational power, the present state-of-art models have the ability to

100

perform rapid computations for operational needs at high spatial resolutions. It has led to

101

a tremendous boost in the accuracy of computed physics and provided a challenging

102

opportunity to skill assess the performance of model computations. As a result, the

103

accomplishment of model computations connecting complicated physical processes on

104

the non-linear feedback mechanism between waves and currents is now possible. Prior

105

modeling studies conducted elsewhere have treated waves and currents as separate

106

entities. Numerical wave modeling efforts are documented in the published work by

107

Padhy et al. (2008), Chitra et al. (2010), Chitra and Bhaskaran (2012), Nayak et al.

108

(2012), Remya et al. (2012), Chitra and Bhaskaran (2013), Bhaskaran et al. (2013),

109

Nayak et al. (2013a, 2013b), Prasad Kumar et al (2000, 2003, 2004, 2007, 2010), and

110

Sandhya et al. (2014). Also, very few recent studies are reported for Indian Ocean basin

111

coupling waves and currents as a single entity (Bhaskaran et al. 2013, Murty et al. 2014).

112

These studies highlight the importance of nonlinear wave-current interaction and the

113

necessity to treat them as a single entity in storm surge models.

AC C

EP

TE D

M AN U

SC

RI PT

93

114

4

ACCEPTED MANUSCRIPT

Coupling of wave and hydrodynamic models facilitate one to understand precisely the

116

complex non-linear interaction phenomena. The present study uses the coupled ADCIRC

117

(Advanced Circulation Model for Shelves, Coasts, and Estuaries) and SWAN (Simulating

118

Waves Nearshore) model. The ADCIRC (Luettich et al. 1992) is a finite element

119

hydrodynamic model that solves the fully nonlinear shallow water equations in the

120

generalized wave continuity form. The SWAN (Booij et al. 1999) is a third-generation

121

wave prediction model based on action balance equation, having applications to estimate

122

wave parameters in coastal waters and estuaries. The computation of wave setup in

123

coupled model forms an integral part of the total water level elevation caused by storm

124

surge during extreme weather events. In a general sense to understand the process of

125

setup and set-down in near-shore areas, it is imperative to have coupled hydrodynamic

126

models. The scientific document of flood hazard mapping (Dean et al, 2005) highlights

127

the importance of wave-induced setup and associated run-up in the near-shore areas. It

128

has diverse practical applications in coastal and ocean engineering disciplines.

SC

M AN U

TE D

129

RI PT

115

The coastal belt of Andhra Pradesh comprises of diverse geomorphic features

131

predominated by depositional landforms such as beach ridge -swale complexes, mudflats,

132

mangrove swamps, spits, lagoons, barriers, estuaries and tidal inlets. In a few localities

133

either side of Visakhapatnam city, there exist a number of rocky headlands fringed by

134

cliffs, wave-cut benches, and other erosional landforms. These diverse geomorphic

135

features have a direct bearing on the wave-induced setup. The significant wave height

136

and water level elevation computed using a coupled model was validated with in situ

137

wave rider buoy and tide gauge observations located off Visakhapatnam. The subsequent

AC C

EP

130

5

ACCEPTED MANUSCRIPT

sections provide more details on Hudhud cyclone, the parametric wind formulation, and

139

its validation, coastal geomorphic features along Visakhapatnam; the data and

140

methodology; followed by the results and discussion on storm surge and wave induced

141

setup computed for the Andhra Pradesh coast.

2. Details of Hudhud Cyclone

The India Meteorological Department (IMD) reported a low-pressure system that formed

SC

142 143 144 145 146

RI PT

138

over the Tenasserim coast adjoining the Andaman Sea in the early hours on October 6,

148

2014 that transformed into a depression on the next day. Under favorable conditions, this

149

system intensified into a deep depression and progressed in the west-northwestward

150

direction. On 8th October 2014, the cyclonic system further intensified, and IMD named it

151

as ‘Hudhud’. As per the JTWC report on 10th October, it was classified as a Category-I

152

tropical cyclone. The advisory upgraded it to a Category-2 on the later part of the same

153

day. Further, on 11th October 2014 the system reached its peak intensity with a minimum

154

central pressure of 950 mb, and average wind speed of 185 km h-1. The coastal belt of

155

Andhra Pradesh witnessed the worst aftermath at landfall during the noon hours on 12th

156

October near Visakhapatnam, a cosmopolitan city having numerous industrial units and

157

vital infrastructure installations (Figure-1). It was the most severe cyclone that struck

158

Visakhapatnam in the past three decades with the core of hurricane winds. The eye

159

locations and cloud top temperatures (BT4) during October 10, 2014 (at 23:15 h) while in

160

the deep water, and on October 12, 2014 (at 04:00 h) during the day of landfall obtained

161

from the INCOIS Ground Station are shown in Figures-1a and 1b respectively. The

162

recorded maximum wind gust by the Cyclone Warning Center in Visakhapatnam was

AC C

EP

TE D

M AN U

147

6

ACCEPTED MANUSCRIPT

about 210 km h-1, and the Doppler Weather Radar (DWR) reported its eye diameter about

164

66 km. A trail of destruction resulted during and after the landfall. Thereafter, the

165

cyclonic system continued over land for quite some time, and finally weakened into a

166

low-pressure system over eastern Uttar Pradesh before its final dissipation. Figure-1c

167

shows the track of Hudhud cyclone.

168 INSERT FIG 1 HERE

SC

169

RI PT

163

170

As per the media reports, the energy department in Andhra Pradesh sustained a maximum

172

loss of more than 1,000 crores during this event. It was the first post-monsoon cyclone to

173

cross Visakhapatnam after 1985, and interestingly the landfall coincided on the same day

174

as Phailin cyclone in 2013. Operational forecasts were issued to the National and State

175

level disaster authorities reporting hourly updates about its movement and intensity on

176

the day of landfall for emergency preparedness. The IMD also issued warning

177

disseminations to local people in the affected states of India.

TE D

EP

3. Maximum Radius of Curvature for Cyclones

Chavas and Emanuel (2010) analyzed the size of tropical cyclones using QuikSCAT

AC C

178 179 180 181 182

M AN U

171

183

Level 2B climatology dataset. They analyzed near-surface wind vectors (10 m) for the

184

period from 1999-2008 (grid size of 12.5 km × 12.5 km) to develop a climatology for

185

tropical cyclone size, also referred as the 'radius of vanishing winds'. They analyzed 2154

186

tropical cyclone samples to estimate the azimuthally averaged 12 m s-1 winds (r12) along

187

with the outer radius (r0). The parameter r0 was used to determine r12 using the outer wind

7

ACCEPTED MANUSCRIPT

structure model, where deep convection was absent beyond r12. Based on their study

189

(Chavas and Emanuel, 2010) the global median values of r12 and r0 were about 197 km

190

and 423 km respectively with significant statistical variations over different ocean basins.

191

The global distribution of r12 followed a lognormal distribution, whereas r0 also followed

192

a distribution closely to lognormal. In another study for the Atlantic basin, Kimball and

193

Mulekar (2004) signify that during storm intensification the radius of outermost closed

194

isobar remained approximately constant irrespective of changes in the radial structure of

195

intermediate wind fields. These works provide credible evidence that the global

196

distribution of tropical cyclone size exhibits a lognormal distribution. Hence, there is a

197

need to re-visit the original wind formulation and modify them accordingly, keeping in

198

view the maximum radius of curvature of tropical cyclones in a climate change scenario.

M AN U

SC

RI PT

188

199

The quality of surface winds is very crucial and one of the primary requirements to obtain

201

realistic estimates of storm surge and associated coastal flooding. In a forecast mode,

202

meteorological models are not available at very high resolution to provide accurate wind

203

fields for storm surge computation (Fleming et al., 2008). Therefore, parametric wind

204

models which are easy to use and produce better quality winds for storm surge

205

computation have gained importance in storm surge studies (Houston et al., 1999;

206

Mattocks et al., 2006). The present study attempts to modify the parametric dynamical

207

wind model of Jelesnianski and Taylor (1973) taking into consideration of the increased

208

maximum radius of curvature in a changing climate (Figure-2). The objective is to

209

provide a realistic estimate of radial distance having dependence on the exponential

210

decay in wind field profile from the location of cyclone eye.

AC C

EP

TE D

200

8

ACCEPTED MANUSCRIPT

211 212 3.1 Parametric Wind Formulation

215

A stationary symmetric wind profile model is the initial condition used to obtain the

216

dynamic wind profile. The asymmetry due to storm motion is then accounted by

217

approximating the corresponding correction term. The original formulation (JT) uses the

218

vector equation governing the horizontal motion of wind flow near sea-surface given by:

dt

=−

1

ρa

grad P + fVg × k + F

SC

dVg

M AN U

219

RI PT

213 214

[1]

In Eq.[1], k is the vertical unit vector, Vg is the wind velocity, ρ a is the air density, f is

221

the Coriolis parameter, P is the atmospheric pressure, and F is the horizontal frictional

222

force per unit mass. The parameters such as pressure and direction are determined from

223

the wind speed by balancing forces. The adapted relation from Myers and Malkin (1961)

224

takes the form:

225

1 dp ks v 2 dV = −V ρ a dr sin φ dr

226

1 dp V2 dφ cos φ = fV + cos φ − V 2 sin φ + knV 2 r dr ρ a dr

227

where, ρ a is the density of atmosphere considered constant, p(r ) is the pressure, φ (r ) is

228

the inflow angle, and V (r ) is the wind speed dependent on r , the distance from centre of

229

the cyclone. The parameters ks and kn are the empirical coefficients denoting stress

EP

TE D

220

AC C

[2a]

9

[2b]

ACCEPTED MANUSCRIPT

230

coefficients in the opposite and to right direction of the wind respectively. The parameter

231

f denotes the Coriolis term. Eliminating the pressure term from Eq. [2a] and [2b] and

using u = cos φ the resulting formulation becomes:

233

du u  1 dV 1  f = ks −u +  − − kn dr  V dr r  V 1− u2

234

where, u = Vg r cos φ . The right hand side of Eqn. [3] is termed as the 'slope function'. The

235

distribution of P(r ) follows the wind profile relation proposed by Jelesnianski and Taylor

236

(1973) and expressed in the form:

237

 2 R r V (r ) = Vr  2 m 2  Rm + r

238

The modified form of Eqn.[4] as suggested by Jelesnianski and Taylor (1973) is given

239

below and used to generate wind field for the present study.

240

 2 R r V (r ) = Vr  2 m 2  Rm + r

241

In the above equation, V (r ) is the value of maximum wind speed and Rm is the radial

242

distance from the storm center, where the maximum wind speed is concentrated. The

243

value of Rm is usually fixed from synoptic maps. More details are available in the NOAA

244

Technical Memorandum (Jelesnianski and Taylor, 1973). The Eqn.[4] is modified by

245

raising the power of qr as proposed in their work. The present study proposes an

246

optimum value of qr = 3 / 5 based on several numerical experiments. Figure 2b shows the

qr

SC

M AN U

TE D

)

  

[4]

[4a]

AC C

(

)

[3]

EP

(

  

RI PT

232

10

ACCEPTED MANUSCRIPT

comparison of wind speed against the radial distance from the cyclone centre using

248

different values of qr . The importance in selection of optimum qr is with regard to the

249

wind speed in the outer core of the cyclone. It is noteworthy that cyclones that developed

250

over the Bay of Bengal in the present decade have higher outer core winds (as observed

251

from buoy records) compared to the past. The section below provides an elaborate

252

discussion that verifies the correctness of the modified wind formulation.

SC

RI PT

247

253 254 255 256 257

The Figure-2c shows the track of five very severe tropical cyclones (2000 Cuddalore

258

cyclone; 2011 Thane cyclone; and 2013 Lehar, Mahasen, and Phailin cyclones) in the

259

Bay of Bengal that occurred during the past one decade. It also shows the location of six

260

deep-water buoys that monitored the meteorological and oceanographic conditions during

261

the extreme events. These five cyclones had landfall along each of the four maritime

262

states located on the east coast of India, and provided an excellent test-bed to verify the

263

modified wind formulation. The meteorological and oceanographic data reported by the

264

network of in-situ buoys used in this study provided an excellent opportunity to verify the

265

original JT formulae pertaining to the maximum radius of curvature of tropical cyclones

266

in the Bay of Bengal basin. The wind field monitored from in situ buoys during extreme

267

event provided an opportunity to fine-tune and modify the formula. Based on several

268

numerical experiments, the study signifies that the application of 3/5-power law resulted

269

in wind fields that were in a close match with the observed buoy wind speed.

AC C

EP

TE D

M AN U

3.2 Validation of the modified Wind formulation

270 271

INSERT FIG 2 HERE

11

ACCEPTED MANUSCRIPT

Figure-3 shows a comparison of the respective wind field distribution using the original

273

(panel-a) JT and 3/5-power law modified (panel-b) JT wind formulation for each cyclone

274

event. It is clear that the modified JT wind formula covers a larger radius of influence

275

(wind field envelope) shown in Table-1. In the case of Cuddalore cyclone (2000), the

276

radial distance between the maximum wind speed vector of 25 m s-1 and 7 m s-1 was

277

about 239 km estimated using original parametric Jelesnianski formulation, whereas from

278

the modified formula, the estimated radial distance was 394 km. For Thane cyclone

279

(2011) considering the range of wind speed from 35 m s-1 and 9 m s-1 the estimated radial

280

distance were 215 km and 367 km using the two formulations respectively. Similarly, for

281

the Lehar cyclone (2013) the estimated radial distance were 260 km and 425 km

282

respectively for the maximum wind speed ranging between 33 m s-1 and 8 m s-1. In the

283

case of Mahasen cyclone (2013), the estimated radial distances were 186 km and 285 km

284

considering the maximum wind speed about 23 m s-1 and 7 m s-1. Lastly, for the Phailin

285

cyclone (2013), the estimated maximum wind speed was 61 m s-1. The estimated radial

286

distance (from maximum of 10 m s-1) using the two wind formulations were 390 km and

287

739 km. Based on analysis from each of these five cyclones, a strong correlation exists

288

between the maxima-minima range in wind speed (∆ws) and the corresponding radial

289

distance (Rd). The best fit regression between ∆ws and Rd is of the form: Rd = 6.7492

290

(∆ws) with R2 = 0.9511 as the coefficient of determination.

292

SC

M AN U

TE D

EP

AC C

291

RI PT

272

INSERT FIG 3 HERE

293

12

ACCEPTED MANUSCRIPT

The Figure-4a shows a comparison of wind speed between both the formulas against the

295

buoy observed wind speed. It also shows the respective distance (in km) of cyclone eye

296

(time stamp marked along cyclone track in the right panel of Figure-3) from the in-situ

297

buoy location. The overall comparison clearly shows that the modified JT winds using

298

the 3/5-power law performed relatively better than the original parametric wind

299

formulation. Their comparison statistics such as the RMSE (in m s-1) and SI (Scatter

300

Index in %) against buoy observation are shown in Table-2. In general, SI less than 30%

301

is widely accepted by the user community for operational planning (Woodcock et al.,

302

2007). The modified JT winds exhibits SI less than 30%. The Figure-4b shows the

303

comparison of computed storm surge using both wind formulas against buoy observation

304

for the Phailin cyclone (2013). The storm surge validation with the modified wind

305

formula shows an excellent match with observation. It is very clear that the computed

306

storm surge using the original JT wind formula is highly underestimated.

SC

M AN U

TE D

307

RI PT

294

The study brings to light that the 3/5 power law scaling provided the best estimate of the

309

radial wind profile for the Bay of Bengal cyclones in a climate change scenario. Chavas

310

et al (2015) developed a simple model for the complete radial structure of tropical

311

cyclones providing a comparison of the observed structure. Their model merged existing

312

theoretical solutions for radial wind structure on top of the boundary layer in the inner

313

ascending and outer descending regions. The outer region solution was compared with

314

the global database from QuikSCAT satellite for the period from 1999-2009, and

315

solutions for the inner region was compared with HWind database for the period from

316

2004-2012 for the Atlantic and eastern Pacific basins. Their model (Chavas et al, 2015)

AC C

EP

308

13

ACCEPTED MANUSCRIPT

substantially underestimated the wind speeds at larger radii that required further

318

improvement. The scope of the present study is limited to verify the modified wind

319

formulae for the Bay of Bengal cyclones. A detailed study is required to verify the

320

effectiveness of this modified formula for cyclones in other ocean basins such as the

321

Atlantic and Pacific, and planned as a future scope of the work.

322

RI PT

317

INSERT FIG 4 HERE

324 325

A number of environmental factors influence the geomorphologic processes on coastal

326

landforms. It includes geological, climatic, biotic, tidal, and other oceanographic factors

327

including sea level changes. However, these factors vary from one region of the coast to

328

another. The continental shelf on the western side of India is comparatively wider varying

329

from 20 to 160 km, whereas on the eastern side of the head Bay region it is as much as

330

160 km wide. For the Andhra coast the shelf has an average width of about 43 km. This

331

study considers the coastal stretch from Visakhapatnam to Bheemunipatnam that had a

332

direct impact from Hudhud cyclone. The coastal geomorphic setting along this region

333

(Figure-5a) comprises of diverse features (Jagannadha Rao et al., 2012).

335 336

M AN U

TE D

EP

AC C

334

SC

323

INSERT FIG 5 HERE

337

This coastal strip has a length of about 24 km with geologically significant features. The

338

beach width varies from 45 to 60 m. The slopes are higher in the foreshore areas having

339

gradients varying from 4° - 5°. The maximum beach width is around the proximities off

340

Rushikonda, Uppada and Bheemunipatnam. The width of continental shelf varies from

14

ACCEPTED MANUSCRIPT

35 to 40 km. The depth of bottom topography varies up to 40 m between Kutukonda and

342

Uppada near Bheemunipatnam with an average gradient of 1:14. To the north of Uppada,

343

the average gradient is about 1:24. The bottom features shown in Figure-5b are survey

344

records from echograms reported by Rao et al (1980). As seen, the general bottom

345

features comprise of Karstic, Reef, and Terrace structures. Near the Visakhapatnam

346

region, terrace structures are most common and well defined. In contrast to the bottom

347

features

348

Bheemunipatnam is void of the dome shaped structure and is primarily comprised of

349

Karstic pinnacled structures. The near-shore width of the 20 m contour line is wider at

350

Bheemunipatnam (Figure-5b). The steepness is relatively larger at Visakhapatnam having

351

direct implications in modulating the phenomena of wave-induced setup.

352 353

4. Data and Methodology

the

coastal

stretch

northward

SC

Visakhapatnam,

near

M AN U

at

TE D

354

(Figure-5b)

RI PT

341

The parameters pertaining to period, amplitude, wavelength, and direction of storm tides

356

depends upon the geometric property of water body and the characteristics of

357

meteorological parameters (Blain et al. 1994b). For realistic estimates, the model study

358

domain should incorporate complex coastal geometry, accounting for rapid changes in

359

bathymetric gradients of the slope and shelf areas, together with reasonable well-defined

360

boundary conditions. The finite element mesh used in the present study provides a good

361

representation of the observed coastal features. The bathymetric data GEBCO (General

362

Bathymetric Chart of the Oceans) having a grid spacing of 30 arc seconds (Figure-6a),

363

maintained by the British Oceanographic Data Centre (BODC) was used in this study.

364

The Surface Modeling System (SMS) an interface to ADCIRC generates the Triangulated

AC C

EP

355

15

ACCEPTED MANUSCRIPT

Irregular Network (TIN) grids for the computational domain using the GEBCO

366

bathymetry data. The study domain covers the entire Bay of Bengal, and Figure-6a,b

367

shows the bathymetric details along with the grid structure. As seen the flexible

368

unstructured grid resolves the sharp bathymetric gradients found in the shelf region along

369

the east coast of India (enlarged version for Andhra Pradesh coast shown in Figure-6c).

370

The unstructured grid used in this study comprises of 123,594 vertices and 235,952

371

triangular elements. Physical phenomena of tides and storm surge can be resolved using a

372

coarse grid in deep waters, whereas the resolution is critical and needs to be higher in

373

coastal and near-shore waters for better estimates (Blain et al., 1994a; Leuttich et al.,

374

1995; Bhaskaran et al., 2013). The flexibility in grid structure provides an allowance to

375

relaxation of grid resolution in deep waters, and refinement based on bathymetric features

376

in the near-shore areas. The grid used in this study is the most optimized version in the

377

context of computational time. The minimum/maximum grid resolution is ≤ 500 m along

378

the coast in near-shore areas, and relaxing to 30 km along the offshore boundary in the

379

deep ocean. A recent study (Bhaskaran et al., 2013) suggests that a high-resolution

380

flexible mesh in near-shore areas resolves the complex bathymetry, and thereby provides

381

a better resolution for wave transformation. The criterion in the fixing grid resolution not

382

exceeding 1 km near-shore is also justified based on the work by Rao et al., (2009). Their

383

study highlights that a grid resolution of 1 km is sufficient and good enough for precise

384

computation of surge heights along the east coast of India.

AC C

EP

TE D

M AN U

SC

RI PT

365

385

INSERT FIG 6 HERE

386

The bottom friction coefficient used in ADCIRC model was 0.0028 with a time step of 10

387

s. Bottom friction coefficient selected in this study is best suited for the sandy bottom 16

ACCEPTED MANUSCRIPT

environment along the Andhra coast, an optimum configuration with both ADCIRC and

389

SWAN models (Murty et al., 2014). The model run execution was from 8th October, 2014

390

(00 h) when Hudhud was in deep waters, until the time of landfall (forenoon of 12th

391

October, 2014). The total length of the simulation was 120 hours with a ramp function of

392

one day, and computation was performed using the parallel computing architecture at

393

INCOIS utilizing 320 processors. In order to establish a robust coupling procedure, the

394

parameters in both these models were set according to the domain type and study

395

specification. Accordingly, the parameters specifying run time, time step, and coupling

396

time interval, forcing frequencies are included in the model setup, and coupling time step

397

for SWAN was set to 600 s. A recent study by Bhaskaran et al. (2013) advocates that the

398

above mentioned coupling time step very well suffices to understand the non-linear

399

interaction effects arising from changing water levels and currents in the resultant wave

400

field.

401

The implementation of SWAN comprises of 36 directional and 35 frequency bins. These

402

numbers are optimum to resolve the spectral distribution of wave energy propagation,

403

and capture realistically the evolution of wave energy in both geographic space and time.

404

The prescription of wave frequency uses logarithmic frequency bins ranging from 0.04 to

405

1.0 Hz, with an angular resolution of 10°. The physical process of non-linear wave-wave

406

interaction activates using the quadruplet’s discrete interaction approximation (DIA)

407

technique. The bottom friction formulation of Madsen et al. (1988) takes care of the

408

bottom resistance for spatially varying roughness length in near-shore regions. This study

409

uses the Madsen formulation with 0.05 m as the bottom-roughness length scale. The

AC C

EP

TE D

M AN U

SC

RI PT

388

17

ACCEPTED MANUSCRIPT

source/sink functions in SWAN run for wind input and white capping dissipation uses

411

Komen et al. (1984) formulation.

412

The nearest location with an observing system for Hudhud cyclone was Visakhapatnam

413

that recorded wave data from a directional wave rider buoy, and water level observations

414

for storm surge from tide gauge (Figure-6c). The Datawell directional wave rider buoy

415

(DWRB) located off Gangavaram (17.63°N, 83.26° E) is at a water depth of 20 m

416

(Datawell, 2009). Integral wave parameters such as significant wave height, maximum

417

wave height, peak wave period and mean wave period are derived from the wave

418

spectrum. Data reception at INCOIS is in real-time through the Indian National Satellite

419

System (INSAT)/ Global System for Mobile Communications (GSM). The DWRB

420

measures wave height and wave periods ranging between 1.6 to 30 s with an accuracy of

421

0.5% of measured value. The continuous water level monitoring using a tide gauge was

422

also available at Visakhapatnam during the Hudhud event. The tide gauge located off

423

Visakhapatnam (17.683° N, 83.283° E) comprises of three sensors such as the pressure

424

sensor, shaft encoder, and radar gauge. The extreme water level recorded by these sensors

425

through data logger communicates to INCOIS through the INSAT system. In this study,

426

data from these two observational platforms were used to skill assess the performance of

427

the coupled model.

428 429 430

5. The Coupled Hydrodynamic Modeling System

431

The coupled hydrodynamic modeling system used in the present study takes into account

432

the mutual nonlinear interaction between waves, current, and storm surge. Two

AC C

EP

TE D

M AN U

SC

RI PT

410

18

ACCEPTED MANUSCRIPT

hydrodynamic models ADCIRC (Advanced Circulation Model) and SWAN (Simulating

434

Waves Nearshore) both coupled in a ‘tight-coupling mode’ such that the mutual exchange

435

of information of physical process occurs between these two models during the

436

integration process. The tight coupling mode is a grid structure that computes waves and

437

hydrodynamic conditions at similar grid nodes in the study domain. The parametric wind

438

model generates the wind fields for simulation using the modified JT formulation, along

439

with the best track estimates from India Meteorological Department (IMD). The Section-

440

3.1 provides a detailed discussion on the modified JT wind formulation and its

441

implementation in the present study. This study used the unstructured version of SWAN

442

(Version 40.85) implementing an analog to the four-direction Gauss-Seidel iteration

443

technique with unconditional stability (Zijlema, 2010). The model computes the resulting

444

wave action-density spectrum at each vertex of the unstructured grid, and finally provides

445

essential wave parameters such as significant wave height, wave period, and wave

446

direction.

SC

M AN U

TE D

447

RI PT

433

In recent studies using coupled model, location specific grids were used to model VSCS

449

(Very Severe Cyclonic Storm) such as Phailin (Murty et al, 2014), and Thane (Bhaskaran

450

et al, 2013). The development of basin scale, flexible unstructured grid covering the

451

entire Bay of Bengal has a definite advantage to model both wave and hydrodynamic

452

conditions for maritime states along the east coast of India, as well serving other

453

countries surrounding the Indian Ocean rim. This study considers the nonlinear

454

dynamical aspects from the contribution of waves to total water level elevation through

455

physical mechanism of radiation stress; computation of wave induced setup and set-

AC C

EP

448

19

ACCEPTED MANUSCRIPT

down. In addition, the study also investigates the role of storm surge on basin scale wave

457

propagation through physical mechanisms such as refraction, shoaling, wave dissipation,

458

and wave breaking effects; effect of total water level elevation on generation,

459

propagation, and dissipation of wind-waves.

460

RI PT

456

The momentum flux associated with spatial variability of wave action density leads to the

462

radiation stress that eventually contributes to near-shore hydrodynamics usually

463

expressed in terms of wave set-up and set-down. Wave set-up is a transient increase in

464

water levels in the near-shore regions arising from the transfer of wave momentum to the

465

underlying water column from wave breaking process. Estimates of wave setup and set-

466

down process is crucial especially during cyclones in the near-shore region, as the net

467

water level is a cumulative effect from reduced atmospheric pressure, storm surge, tidal

468

effects, and wave induced setup. The cumulative effects of net water level elevation have

469

profound implications on onshore inundation studies.

470 471

For the Indian Ocean region, recent studies by Gayathri et al (2015), Bhaskaran et al

472

(2013, 2014), Murty et al. (2014) demonstrated the application of ADCIRC and coupled

473

ADCIRC-SWAN model for operational use. The depth-averaged version of fully

474

parallelized ADCIRC model used in these studies computes the storm surge, depth

475

averaged currents, and the net water level elevations due to surge and wave along the

476

Andhra coast. In a tight coupling mode (ADCIRC+SWAN), the ADCIRC model in the

477

prescribed coupling time interval uses radiation stress from SWAN to extrapolate

478

forward the wave forcing in time. On completion of the coupling time interval, the

479

ADCIRC exchange the wind velocity, water levels, currents, and roughness length to

AC C

EP

TE D

M AN U

SC

461

20

ACCEPTED MANUSCRIPT

SWAN model. Finally, the updated water level information and corresponding currents

481

computed by ADCIRC shares with SWAN to update the wave transformation processes.

482

Both ADCIRC and SWAN models march ahead with time, through the mutual exchange

483

of information.

RI PT

480

484 485 486 487 488

Many applications utilize the parametric wind formulations to depict the radial profiles of

489

cyclonic wind fields that evolved from the Rankine vortex formulation wherein solid

490

body rotation is assumed in the core region close to the eye wall, with tangential winds

491

decreasing by the radial scaling parameter using a rectangular hyperbola approximation

492

to radial pressure variation (Schloemer, 1954). At present, a new metric termed as TIKE

493

(Track Integrated Kinetic Energy) classify the cyclones (Misra et al., 2013) based on

494

parameters such as cyclone intensity, duration, and size. Other metrics such as

495

Accumulated Cyclone Energy (ACE) and Power Dissipation Index (PDI; Emanuel, 2005)

496

supplements TIKE. In context to cyclones in global ocean basins, these metrics have

497

gained attention at large from the scientific community. There are evidences that indicate

498

a drastic increase in these metrics for recent decade cyclones that form over the north

499

Indian Ocean. This study used the parametric JT wind formulation to force the coupled

500

ADCIRC+SWAN model, based on best track data from the Indian Meteorological

501

Department (IMD). The Figures-7(i) and (ii) show the respective wind fields generated

502

using both the wind formulations for Hudhud episode from October 9, 2014 (00 UTC)

503

until October 11, 2014 (06 UTC). The plots show the minimum wind speed covering the

504

5 m s-1 envelope. This provides an insight into the maximum radius of curvature. As

AC C

EP

TE D

M AN U

SC

6. Results and Discussion

21

ACCEPTED MANUSCRIPT

noticed from these Figures, the modified JT wind formulation (qr = 0.60) provides a

506

better description of the spatial coverage of wind field (left panel in Figure-1) as

507

compared to the original parametric wind formulation. The modified Jelesnianski winds

508

will undoubtedly serve as a better input for storm surge and coastal inundation studies.

509 510

INSERT FIG 7 HERE

SC

511

RI PT

505

Figure-8a shows the computed maximum significant wave height (in m) from the coupled

513

model run. Waves are stronger along the coastal belt north of Visakhapatnam (in excess

514

of 8.0 m) that faces the right side of the track, attributing to strong onshore winds. For

515

regions on the leeward side of Hudhud track, the wave heights are relatively small (less

516

than 4.0 m) as the predominant wind direction is in the offshore direction. The spatial

517

distribution of significant wave heights along the coastal stretch from Visakhapatnam to

518

Bheemunipatnam separated by a distance of about 24 km is almost similar.

TE D

519

M AN U

512

The Figure-8b shows the validation of significant wave height between the coupled

521

model and wave rider buoy recorded at Gangavaram, south of Visakhapatnam. The

522

coupled model performs well representing the variation of extreme waves in the near-

523

shore region off Visakhapatnam quite satisfactorily. The Figure-8c shows the comparison

524

of computed storm surge at the locations off Bheemunipatnam and Visakhapatnam using

525

ADCIRC in standalone mode and coupled to SWAN using both wind formulations. In

526

this figure, for coupled model the surge residual plotted along the Y-axis is the combined

527

effect of storm surge and wave induced setup. During the time of landfall, the maximum

AC C

EP

520

22

ACCEPTED MANUSCRIPT

computed storm surge at Bheemunipatnam was about 1.5 m from the ADCIRC

529

standalone run, and the contribution from wave-induced setup was about 0.5 m. Figure-

530

8d shows the computed surge (in m) at a location 200 km northward (located in the outer

531

core of the cyclone) from the actual landfall location. It is evident that the surge levels are

532

comparatively higher with the modified JT winds. The simulated difference in surge level

533

is approximately 0.22 m using winds generated by both the wind formulation. By using

534

the original JT formula, the contribution from waves are negligible (Figure-8d) as the

535

simulated winds are weak in the outer core of cyclone and hence there is no difference in

536

the computed surge level with standalone and coupled models. On the other hand using

537

the modified wind formula there is a difference of about 0.1 m in the surge level between

538

the standalone and coupled model. However, there are no in situ observations available

539

at this location during this event for validation. However, in Figure-8c both

540

Bheemunipatnam and Visakhapatnam are located in the core area of Hudhud cyclone,

541

and therefore the difference in surge levels simulated using both wind formulations are

542

marginal. The wave setup gradually increased during the approach of cyclonic system

543

and diminished rapidly after its landfall (Figure-8e). In addition, no considerable

544

differences noticed in the surge amplitude at Visakhapatnam comparing the ADCIRC

545

standalone and coupled model runs. The maximum storm surge height predicted by

546

ADCIRC with and without SWAN (Figure-8c) was about 1.05 m. In contrast to the

547

observed wave set-up phenomena at Bheemunipatnam, the location off Visakhapatnam

548

experienced set-down. The long period oscillations in the computed water level elevation

549

require a separate study. The wave setup remained almost invariant at Visakhapatnam

550

followed by set-down during the landfall event. On the other hand, at Bheeminupatnam

AC C

EP

TE D

M AN U

SC

RI PT

528

23

ACCEPTED MANUSCRIPT

the wave setup increased during the approach of Hudhud. There is a steady increase seen

552

until the landfall time, and thereafter the wave set-down attributes from predominant

553

offshore winds at this location.

554

Figure-8f shows the coastal slope and significant wave heights during fair weather

555

condition for the Andhra Pradesh coast covering the coastal belts of Visakhapatnam and

556

Bheemunipatnam. This study pertains to sea-level rise and coastal vulnerability through

557

remote sensing techniques (Nageswara Rao et al., 2008). The vulnerability rank at

558

Visakhapatnam varied from very- low to low. For coastal areas north of Visakhapatnam,

559

the vulnerability rank is either moderate or very high. The vulnerability ranking based on

560

significant wave heights (shown in Figure-8f) corresponds to the fair weather condition

561

along the Andhra Pradesh coast (Nageswara Rao et al., 2008).

M AN U

SC

RI PT

551

562 563

TE D

564

INSERT FIG 8 HERE

Based on this coastal vulnerability map (left panel in Figure-8f) it can be assessed that

566

coastal areas south of Visakhapatnam with headlands such as south Kakinada, off

567

Vijayawada, and areas near Pulicat Lake is highly vulnerable attributed due to mild beach

568

slopes. The vulnerability level considering the significant wave height is high between

569

Visakhapatnam to Bheemunipatnam. The coastal areas south of Visakhapatnam

570

especially south of Kakinada are highly vulnerable. The coastal vulnerability covering the

571

narrow stretch from Visakhapatnam to Bheemunipatnam has resemblance with the

572

observations seen for wave-induced setup. It has a direct bearing on the coastal

573

geomorphic features shown in Figure-5. The bottom features off Bheemunipatnam have

AC C

EP

565

24

ACCEPTED MANUSCRIPT

Karstic pinnacled features both along the mid and shelf edge regions retarding wave

575

propagation towards the near-shore areas, causing piling up of water during extreme

576

weather events. It is unlike the bottom dome-shaped features observed off

577

Visakhapatnam comprising of reef structures with a higher gradient in beach slopes. The

578

wave-induced setup has a direct bearing on beach slopes and bottom features, and the

579

results obtained from this study will be of immense value to coastal zone authorities.

RI PT

574

The study reports on the application of a parallel coupled (ADCIRC+SWAN) model with

584

the modified Jelesnianski wind formulation utilizing the IMD best track estimate for

585

Hudhud event that had landfall on 12th October, 2014 near Visakhapatnam in Andhra

586

Pradesh. The coupled model simulated aspects on coastal hydrodynamics such as extreme

587

waves, wave induced setup, and storm surges. A flexible unstructured finite element

588

mesh generated for this study covers the entire Bay of Bengal region having a resolution

589

of 30 km in the open ocean boundary, and refining to less than 1 km along the near-shore

590

areas. The modified JT wind formula considers the correction for radial distance of

591

exponential wind decay, and its verification for computed storm surge of 2013 Phailin

592

cyclone event exhibits an excellent match with the observed data. Based on several

593

sensitivity experiments conducted with six very severe cyclones that occurred over the

594

Bay of Bengal in the present decade, the study obtains an optimum value as 0.60 for the

595

coefficient for radial wind decay. The overall performance of coupled model run with

596

modified wind showed a good match against observations when compared with the

597

standalone mode. Validation exercise indicates the performance of model computation

598

for storm surge and waves against observations from a wave rider buoy and tide gauge

SC

580 581 582 583

AC C

EP

TE D

M AN U

7. Summary and Conclusions

25

ACCEPTED MANUSCRIPT

located in Visakhapatnam. A comprehensive analysis was carried out to understand the

600

variation of wave induced setup for a narrow coastal stretch between Visakhapatnam to

601

Bheemunipatnam, separated by a distance of almost 24 km. Interestingly, the wave

602

induced setup at both these locations was quite different, directly linked to the coastal

603

geomorphic features and beach slopes at these locations. The study also signifies the

604

importance of wave induced setup and its role in determining the total water level

605

elevation along the coastal stretch of Andhra Pradesh state. This information is quite vital

606

and very useful in the development of an operational forecasting system, and information

607

dissemination to coastal zone authorities, and planning departments to render massive

608

evacuation strategies.

SC

M AN U

Acknowledgements

EP

TE D

The authors thank the development team of the coupled model. The data used in this study was obtained from INCOIS, Hyderabad an organization under the Ministry of Earth Sciences, Government of India and 'Data is available upon request' from the author with the email address: [email protected]. This study is conducted under the HOOFS (High-resolution Operational Ocean Forecast and Reanalysis System, vide Sanction No. F/INCOIS/HOOFS-03/2013).

AC C

609 610 611 612 613 614 615 616 617 618 619

RI PT

599

26

ACCEPTED MANUSCRIPT

References

RI PT

Bhaskaran, P.K., R. Gayathri, P.L.N. Murty, B. Subba Reddy, and D. Sen (2014), A numerical study of coastal inundation and its validation for Thane cyclone in the Bay of Bengal. Coastal Eng. 83, 108-118. Bhaskaran, P.K., S. Nayak, S.R. Bonthu, P.L.N. Murty, and D. Sen (2013), Performance and validation of a coupled parallel ADCIRC–SWAN model for THANE cyclone in the Bay of Bengal. Environ. Fluid Mech. 13(6), 601-623.

M AN U

SC

Blain, C.A., J.J. Westerink, R.A. Luettich, and N.W. Scheffner (1994a), ADCIRC: An advanced three-dimensional circulation model for shelves, coasts, and estuaries, Report No. 4. Hurricane storm surge modeling using large domains. Tech. Report DRP-92-6, August 1994, U.S. Army Engineer Waterways Experiment Station, Vicksburg, Mississipi. Blain, C.A., J.J. Westerink, and R.A. Luettich (1994b), The influence of domain size on the response characteristics of a hurricane storm surge model. J. Geophys. Res. 99(C9), 18467-18479. Booij, N., R.C. Ris, and L.H. Holthuijsen (1999), A third-generation wave model for coastal regions, Part-I Model descriptions and validation. J. Geophys. Res. 104 (C4), 7649-7666.

TE D

Chavas, D.R., and K.A. Emanuel (2010), A QuikSCAT climatology of tropical cyclone size. Geophysical Research Letters, 37, L18816, doi: 10.1029/2010GL044558.

EP

Chavas, D.R., Lin, N., and K. Emanuel (2015). A model for the complete radial structure of the tropical cyclone wind field. Part I: Comparison with observed structure. Journal of Atmospheric Sciences, 72, 3647-3662. Chitra, A., B. Prasad Kumar, I. Jain, A. Bhar, and A.C. Narayana (2010), Bottom boundary layer characteristics in the Hooghly estuary under combined wave-current action. Mar. Geodesy 33, 261-281.

AC C

620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664

Chitra, A., and P.K. Bhaskaran (2012), Parameterization of bottom friction under combined wave-tide action in the Hooghly estuary, India. Ocean Eng. 43, 43-55. Chitra, A., and P.K. Bhaskaran (2013), Numerical modeling of suspended sediment concentration and its validation for the Hooghly estuary, India. Coastal Eng. 55(2), 1-23. Datawell BV (2009). Datawell waverider reference Zomerluststraat 4, 2012 LM, The Netherlands, pp. 123.

27

manual.

Datawell

BV,

ACCEPTED MANUSCRIPT

Dean, B., I. Collins, D. Divoky, D. Hatheway, and N. Scheffner (2005), Wave Setup: FEMA Coastal Flood Hazard Analysis and Mapping Guidelines Focussed Study Report, 29.

RI PT

Emanuel, K.A. (1987), The dependence of hurricane intensity on climate. Nature, 326, 483-485. Emanuel, K.A. (2005), Increasing destructiveness of tropical cyclones over the past 30 years. Nature, 436, 686-688.

SC

Fleming, J., C. Fulcher, R. Luettich, B. Estrade, G. Allen, and H. Winer (2008), A Real Time Storm Surge Forecasting System Using ADCIRC. Estuarine and Coastal Modeling, 893-912.

M AN U

Gayathri, R., Murty, P.L.N., Bhaskaran, P.K., and Srinivasa Kumar, T (2015). A numerical study of hypothetical storm surge and coastal inundation for AILA cyclone in the Bay of Bengal. Environ. Fluid Mech., DOI: 10.1007/s10652-015-9434-z. Holland, G (1980), An analytic model of the wind and pressure profiles in hurricanes. Monthly Weather Review 108, 1212-1218. Houston, S.H., W.A. Shaffer, M.D. Powell, and J. Chen (1999), Comparisons of HRD and SLOSH surface wind fields in hurricanes: implications for storm surge modeling. Weather Forecast 14, 671-685.

TE D

Jagannadha Rao, M., A.G. Greeshma Gireesh, P. Avatharam, N.C. Anil, and R. Karuna Karudu (2012), Studies on Coastal Geomorphology along Visakhapatnam to Bhimunipatnam, East Coast of India. Jour. Ind. Geophys. Union 16(4), 179-187.

EP

Jelesnianski, C.P., and A.D. Taylor (1973), A preliminary view of storm surges before and after storm modifications. NOAA Technical Memorandum. ERL WMPO-3, National Oceanic and Atmospheric Administration, U.S. Department of Commerce, pp.33. Kimball, S.K., and M.S. Mulekar (2004), A 15-year climatology of North Atlantic tropical cyclones. Part-I: Size parameters, Jour. of Climate, 17, 3555-3575.

AC C

665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710

Komen, G.J., K. Hasselmann, and S. Hasselmann (1984), On the existence of a fully developed wind-sea spectrum. J. Phys. Oceanogr. 14, 1271-1285. Luettich, R.A, and J.J. Westerink (1995), Continental shelf scale convergence studies with a barotropic model. In D. R. Lynch and A. M. Davies (eds.) Quantitative Skill Assessment for Coastal Ocean Models, Coastal and Estuarine Studies Series, No 47, A G U 349-371. Madsen, O.S., Y.K. Poon, and H.C. Graber (1988), Spectral wave attenuation by bottom friction - Theory. In: Proceedings of the 21st ASCE Coastal Eng. Conf., 492-504.

28

ACCEPTED MANUSCRIPT

Mattocks, C., C. Forbes, and L. Ran (2006), Design and implementation of a real-time storm surge and flood forecasting capability for the State of North Carolina. UNC-CEP Tech. Report, November 2006, 103.

RI PT

Misra, V., S. DiNapoli, and M. Powell (2013), The Track Integrated Kinetic Energy of Atlantic Tropical Cyclones. Monthly Weather Review, 141, 2383-2389. Murty, P.L.N., K.G. Sandhya, P.K. Bhaskaran, F. Jose, R. Gayathri, T.M. Balakrishnan Nair, T. Srinivasa Kumar, and S.S.C. Shenoi (2014), A coupled hydrodynamic modeling system for PHAILIN cyclone in the Bay of Bengal. Coastal Eng. 93, 71-81.

SC

Myers, V.A., and W. Malkin (1961), Some properties of hurricane wind fields as deduced from trajectories. National Hurricane Research Project Rep. 49, NOAA, U.S. Dept. of Commerce, p. 43.

M AN U

Nageswara Rao, K., P. Subraelu, T. Venkateswara Rao, B. Hema Malini, R. Ratheesh, S. Bhattacharya, A.S. Rajawat, and Ajai (2008), Sea-level rise and coastal vulnerability: an assessment of Andhra Pradesh coast, India through remote sensing and GIS. J.Coast Conserv. 12, 195-207. Nayak, S., P.K. Bhaskaran, and R. Venkatesan (2012), Near-shore wave induced setup along Kalpakkam coast during an extreme cyclone event in the Bay of Bengal. Ocean Eng. 55, 52-61.

TE D

Nayak, S., P.K. Bhaskaran, R. Venkatesan, and Sikha Dasgupta (2013a), Modulation of local wind-waves at Kalpakkam from remote forcing effects of southern ocean swells. Ocean Eng. 64, 23-35.

EP

Nayak, S., and P.K. Bhaskaran (2013b), Coastal vulnerability due to extreme waves at Kalpakkam based on historical tropical cyclones in the Bay of Bengal. Int. Journal of Climatology, DOI: 10.1002/joc.3776. Padhy, C.P., D. Sen, and P.K. Bhaskaran (2008), Application of wave model for weather routing of ships in the North Indian Ocean. Nat. Hazards 44, 373-385.

AC C

711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754

Powell, M.D., S.H. Houston, L.R. Amat, and N. Morisseau-Leroy (1998), The HRD realtime hurricane wind analysis system. Jour. of Wind Eng. and Industrial Aerodynamics 77&78, 53-64. Prasad Kumar, B., R. Kalra, S.K. Dube, P.C. Sinha, A.D. Rao, R. Kumar, and A. Sarkar (2000), Extreme wave conditions over the Bay of Bengal during severe cyclone – simulation experiment with two spectral wave models. Mar. Geodesy 23, 91-102.

29

ACCEPTED MANUSCRIPT

Prasad Kumar, B., I.C. Pang, A.D. Rao, T.H. Kim, J.C. Nam, and C.S. Hong (2003), Seastate hindcast for the Korean seas with a spectral wave model and validation with buoy observation during January 1997. J. Korean Earth Sci. Soc. 24(1), 7-21.

RI PT

Prasad Kumar, B., R. Kalra, S.K. Dube, P.C. Sinha, and A.D. Rao (2004), Sea-state hindcast with ECMWF data using a spectral wave model for typical monsoon months. Nat. Hazards 31, 537-548. Prasad Kumar, B., and G.W. Stone (2007), Numerical simulation of typhoon wind forcing in the Korean seas using a spectral wave model. J. Coast. Res. 23(2), 362-373.

SC

Prasad Kumar, B. (2010) Reliability based design method for coastal structures in shallow seas. Indian Geo-Mar. Sci. 39(4), 605-615.

M AN U

Raghavan, S., and S. Rajesh, (2003) Trends in Tropical Cyclone Impact – A study in Andhra Pradesh, India. Bulletin of Amer. Meteorological Society, American Meteorological Society, 635-644. Rao, T.C.S., X.T. Machado, and K.S.R. Murthy (1980), Topographic features over the Continental Shelf off Visakhapatnam. Mahasagar 13(1), 23-28. Rao, A.D., J. Indu, M.V. Ramana Murthy, T.S. Murty, and S.K. Dube (2009), Impact of cyclonic wind field on interaction of surge-wave computations using finite-element and finite-difference models. Nat. Hazards 49, 225-239.

TE D

Remya, P.G., R. Kumar, S. Basu, and A. Sarkar (2012), Wave hindcast experiments in the Indian Ocean using MIKE21SW model. J. Earth Syst. Sci. 121(2), 385-392.

EP

Sandhya, K.G., T.M. Balakrishnan Nair, B. Prasad Kumar, L. Sabique, N. Arun, and K. Jeykumar (2014), Wave forecasting system for operational use and its validation at coastal Puducherry, east coast of India. Ocean Eng. 80, 64-72. Sahoo, B., and P.K. Bhaskaran (2015), Assessment on historical cyclone tracks in the Bay of Bengal, east coast of India. Int. Journal of Climatology, doi:10.1002/joc.4331.

AC C

755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800

Schloemer, R.W. (1954), Analysis and synthesis of hurricane wind patterns over Lake Okeechobee. NOAA Hydrometeorology Report 31, Department of Commerce and U.S. Army Corps of Engineers, U.S. Weather Bureau, Washington, D.C., 49. Woodcock, F., and D.J.M. Greenslade (2007), Consensus of numerical model forecasts of significant wave heights. Weather Forecast, 22, 792-803. Zijlema, M. (2010), Computation of wind wave spectra in coastal waters with SWAN on unstructured grids. Coastal Eng. 57, 267-277.

30

ACCEPTED MANUSCRIPT

Legend to Tables

RI PT

Table-1: Radial distance (in km) between the original and modified Jelesnianski winds for severe cyclones in the Bay of Bengal (Note: the radial distance covers the range from maximum to minimum wind speed. The minimum wind speed is ≈ 10 m s-1 and not taken as zero to highlight cyclone size).

SC

Table-2: Statistics between modeled and observed wind speed for severe cyclones in the Bay of Bengal

M AN U

Legend to Figures

Figure-1: Satellite Imageries for the Eye locations of Hudhud Cyclone (a) during October 10, 2014 (at 23:15 h), (b) during October 12, 2014 (at 04:00 h) the day of landfall, and the corresponding cloud top temperatures (BT4) obtained from INCOIS Ground Station, (c) Track of Hudhud cyclone based on IMD best track estimates.

TE D

Figure-2: (a) Comparison of radial profile of cyclonic wind speeds with the original and modified version of Jelesnianski parametric wind model, (b) same with different values of power law, (c) Tracks of severe cyclones and locations of deep water in situ buoys that recorded the meteorological and oceanographic conditions.

EP

Figure-3: Comparison between the original and modified Jelesnianski and Taylor wind formula for the five severe cyclones in the Bay of Bengal. Figure-4: (a) Validation of wind speed (in m s-1) between original and modified parametric wind formula against buoy observation for the severe cyclones in Bay of Bengal during the present decade, (b) Computed storm surge (in m) and its validation between observed, original, and modified Jelesnianski winds for 2013 Phailin cyclone at Paradip location.

AC C

801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846

Figure-5: (a) Geomorphic Features between Visakhapatnam and Bheemunipatnam, East Coast of India (source: Jagannadha Rao et al., 2012), and (b) Bottom topographic features between Kutukonda to Bheemunipatnam based on echogram reports (source: Rao et al., 1980). Figure-6: (a) Bathymetry of the study region from GEBCO, (b) flexible finite element mesh for the study area, and (c) zoomed version of (b) for the Andhra Pradesh coast along with the location of in-situ observation.

31

ACCEPTED MANUSCRIPT

Figure-7: (i) Time series plot of the wind speed envelope from original Jelesnianski and Taylor parametric wind formulation, (ii) same using the modified Jelesnianski and Taylor parametric wind formulation.

EP

TE D

M AN U

SC

RI PT

Figure-8: (a) Model computed maximum significant wave height (in m) for Hudhud event, (b) Significant wave height validation between model and wave rider buoy off Visakhapatnam (arrow indicates the landfall time), (c) Validation of storm surge against tide gauge observation near Bheemunipatnam and Visakhapatnam, (d) Storm surge at 200 Km northward of landfall point (e) Comparison of wave induced setup (in m) for same locations (as in c), and (f) Coastal vulnerability rank based on coastal slopes and significant wave height during fair weather condition for the Andhra Pradesh coast (source: Nageswara Rao et al., 2008).

AC C

847 848 849 850 851 852 853 854 855 856 857 858

32

ACCEPTED MANUSCRIPT

Cuddalore (2000)

02

Thane (2011)

03

Mahasen (2013)

04

Phailin (2013)

05

Lehar (2013)

AC C

EP

01

Radial distance (in km) from minimum to maximum wind speed

M AN U

Cyclonic Event

Modified Wind

Un-modified Wind

394

239

367

215

285

186

739

390

425

260

TE D

S.No.

SC

RI PT

Table-1: Radial distance (in km) between the un-modified and modified Jelesnianski winds for severe cyclones in the Bay of Bengal (Note: the radial distance covers the range from maximum to minimum wind speed. The minimum wind speed is ≈ 10 m s-1 and not taken as zero to highlight cyclone size).

ACCEPTED MANUSCRIPT

Buoy used for validation

RMSE (in m/s) between model wind and buoy observation

DS03 BD13 BD11

Cuddalore (2000) Thane (2011)

Unmodified

33 22 25

40 43 52

4.14 1.87 2.85 3.10

7.33 4.34 6.39 6.29

42 17 25 27

74 40 57 56

BD11 BD13

4.63 3.06

5.38 6.32

48 26

56 55

BD09 BD08 BD11

3.02 4.76 3.01

9.56 14.27 7.89

15 18 25

49 42 76

TE D

AC C

Phailin (2013)

Modified

3.06 5.25 6.30

EP

Lehar (2013)

Scatter Index (%)

2.58 2.75 3.13

BD11 BD10 BD09 BD08

Mahasen (2013)

Unmodified

M AN U

Modified

SC

Cyclone Event

RI PT

Table-2: Statistics between modeled and observed wind speed for severe cyclones in the Bay of Bengal

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT

Highlights

AC C

EP

TE D

M AN U

SC

RI PT

• Numerical Modeling of Peak Storm Surge for Hudhud Cyclone with modified parametric wind formulation • Modeling of Wave induced setup along Andhra Pradesh coast • Modeling of Significant Wave Heights for Hudhud Cyclone • Investigate dependence of coastal geomorphic features on wave induced setup • Validation of near-shore surge residuals and significant wave height with field data