Accepted Manuscript

Numerical simulation of typhoon-induced storm surge along coast, Part II: Calculation of storm surge

Jin-hai Zheng, Jin-cheng Wang, Chun-yan Zhou, Hong-jun Zhao, Sang Sang

PII: S1674-2370(17)30033-9 DOI: 10.1016/j.wse.2017.03.011 Reference: WSE 94

To appear in: Water Science and Engineering

Received Date: 22 June 2016

Accepted Date: 15 December 2016

Please cite this article as: Zheng, J.-h., Wang, J.-c., Zhou, C.-y., Zhao, H.-j., Sang, S., Numerical simulation of typhoon-induced storm surge along Jiangsu coast, Part II: Calculation of storm surge, Water Science and Engineering (2017), doi: 10.1016/j.wse.2017.03.011.

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. 1 Numerical simulation ofA CCEPTEDtyphoon-induced MANUSCRIPT storm surge along Jiangsu 2 coast, Part II: Calculation of storm surge 3 Jin-hai Zheng a,b,*, Jin-cheng Wang a,b, Chun-yan Zhou a,b, Hong-jun Zhao a,b, Sang Sang a,b

4 a Key Laboratory of Coastal Disaster and Defence (Hohai University), Ministry of Education, Nanjing 210098, ; 5 b College of Harbor Coastal and Offshore Engineering, Hohai University, Nanjing 210098, China

6 Received 22 June 2016; accepted 15 December 2016 7 Available online ***

8 Abstract 9 The Jiangsu coastal area is located in central-eastern China and is well known for complicated dynamics with large-scale radial 10 sand ridge systems. It is therefore a challenge to simulate typhoon-induced storm surges in this area. In this study, a 2D 11 astronomical tide and storm surge coupling model was established to simulate three typical types of typhoons in the area. The 12 Holland parameter model was used to add the background wind field and the offshore boundary information was provided by the 13 improved Northwest Pacific Ocean Tide Model. Typhoon-induced storm surges along the Jiangsu coast were calculated based on 14 analysis of wind data from 1949 to 2013 and the spatial distribution of maximum storm surge levels under different typhoon types, 15 providing references for the design of sea dikes and planning for control of coastal disasters. 16 17 Keywords: Jiangsu coast; Typhoon-induced storm surge; Numerical simulation; Holland parameter model; ADCIRC 18

19 1. Introduction

20 Storm surges induced by typhoons are among the most popular research topics in flood prevention and 21 engineering design along coastal areas (Chen et al., 2009; Zhang et al., 2013). For example, according to the 22 nuclear safety guide HAF0111: The Confirmation of Design Basis Flood on Site Elevation of Seaside Nuclear 23 Power Station , the probable maximum storm surge of the sea area around the nuclear power station is the 24 essential element that determines the design basis of the flood. The design high tide level of nuclear power 25 stations under construction or soon to be under constructionMANUSCRIPT is determined as the sum of the maximum 26 astronomical tide level and the probable maximum storm surge according to international standards (Liang and 27 Zou, 2004). 28 In terms of modeling storm surges along the Jiangsu coast, probable maximum storm surges of Haizhou 29 Bay in the Jiangsu coastal area have been calculated by the numerical model of the storm surge, governed by a 30 depth-averaged flow equation in spherical coordinates, and verified by five cases of remarkable extratropical 31 storm surges with a maximum storm surge of 3.36 m (Yu et al., 2002; Wu et al., 2002). A high-resolution storm 32 surge model along the Jiangsu coast was built using the explicit difference (Zhang, 2008), so the stability was 33 low and the computation time was long. Several typhoons striking Jiangsu Province were simulated using the 34 weather research and forecasting (WRF) model and Delft 3D model in order to investigate the influence of 35 storm surges on the radial sand ridges off the Jiangsu coast with the sea level rising (Yu, 2013). The 36 hydrodynamics in the Jiangsu sea waters during Typhoon Damrey were simulated and a good fit was generated 37 between the simulated and the measured values of the typhoon data (Wang et al., 2015), where the gradient 38 wind field of Typhoon Damrey was computed by a Jele wind parameter model, then compounded with the 39 ambient wind field fromACCEPTED National Centers for Environmental Prediction (NCEP). 40 In this study, hydrodynamic simulations were carried out to investigate the storm surge along the Jiangsu 41 coast during different typhoons, based on the track and parameters of the theoretical typhoons. The calculated 42 surge levels will then show the spatial distribution of maximum storm surge levels under different typhoon 43 types, providing valuable references for planning for control of coastal disasters. 44 ————————————— This work was supported by the National Science Fund for Distinguished Young Scholars (Grant No. 51425901) and the National Natural Science Foundation of China (Grants No. 41606042). * Corresponding author. E-mail address: [email protected] (Jin-hai Zheng).

45 2. Method ACCEPTED MANUSCRIPT 46 2.1. Model description

47 The water levels of tide gauges consist of the astronomical tide level and storm surge level during the 48 typhoon period. This study built an astronomical-tide and storm-surge coupling model and validated the 49 reasonability of model topography, boundary conditions, and the wind field. The surges caused by hypothetical 50 typhoons in the Jiangsu coastal area were calculated using the storm surge model. 51 The numerical model included three sub-models, which were the wind field and wind pressure model, the 52 Northwest Pacific Ocean Tide Model (a large model), and the two-dimensional (2D) typhoon storm surge model 53 for the Jiangsu coast (a small model). The Holland parameter model was used to simulate the wind field and 54 pressure field of the typhoon. The improved Northwest Pacific Ocean Tide Model (Zhang et al., 2012) provided 55 the open boundary water levels for the 2D typhoon storm surge model, and the latter calculated the surge along 56 Jiangsu coastal waters. 57 The Northwest Pacific Ocean Tide Model uses the 2D tidal propagation equation in spherical coordinates 58 (Zhang, 2013), with a domain covering the East China Sea, South China Sea, Philippine Sea, Sea, Sulu 59 Sea, and nearby Pacific Ocean areas. The initial flow velocity is zero. When computing the coupled water level 60 of the astronomical tide level and storm surge level, the open boundary condition for the large model is

61 provided by NAO99 with 16 short-period (M2, S2, K1, O1, N2, P1, K2, Q1, M1, J1, OO 1, 2N 2, Mu 2, Nu 2, L2, and

62 T2) and 7 long-period tidal constituents (Mtm , Mf, MSf , Mm, MSm , Ssa , and Sa), and the boundary condition for the 63 small model is the coupled water level of the astronomical tide level and storm surge level derived from the 64 large model. While the storm surge is simulated alone, the open boundary water level for the large model is set 65 as zero, so that the large model only provides the storm surge for the small model. The model was validated 66 with four main constituents (M2, S2, K1, and O1) of 435 tide gauges listed in the Admiralty Tide Table and the 67 good agreements indicated that the model is able to simulate the tidal system of the , the Huanghai 68 Sea, and the East China Sea. 69 2.1.1. Wind field and wind pressure model MANUSCRIPT 70 The wind field and wind pressure model adopted the Holland parameter model, which is the most popular 71 method to simulate typhoon storm surges. best track dataset, published by the China 72 Meteorological Administration-Shanghai Typhoon Institute (CMA-STI), was used as the input conditions for the 73 model to simulate typhoons (Ying et al., 2014). 74 The Holland parameter model relies on the primary assumption of a radially symmetric pressure field, but 75 with a modified rectangular hyperbola to give the pressure p at any radius ( r) from the center as follows: R  B −max  ()()= + ∆ r  (1) pr pc pe 76 ( ) where p r is the surface pressure (in Pa) at a distance r from the typhoon center; pc is the central pressure; 77 ∆ ∆ = − p is the difference between the ambient pressure ( pn ) and the central pressure ( p pn p c ), and the value of 78 the first anti-cyclonically curved isobar is pn in practice; Rmax is the radius to the maximum wind speed (m), 79 referring to the distance from the typhoon center to the region of maximum wind speed; and B is the so-called 80 profile peakedness used to characterize the shape of the radial pressure profile. 81 Gradient winds inACCEPTED the upper atmosphere are derived from the balance between the centrifugal and Coriolis 82 forces acting outwards and the pressure force acting inwards: V2 ( r ) 1 dp( r ) g +fV() r = (2) g ρ ra dr 83 ( ) where Vg r is the gradient wind speed (m/s) at a distance r from the typhoon center; f is the Coriolis parameter − ϕ 84 and f = 2ω sin ϕ ; ω is the angular speed of Earth’s rotation, set to be 7.2722054× 10 5 rad/s; is latitude; 85 ρ 3 and a is air density (assumed to be constant at 1.15 kg/m in the Holland parameter model).

2 Water Science and Engineering

86 Substituting pressure field Eq. (1)A intoCCEPTED the force balance MANUSCRIPT Eq. (2) will yield the gradient wind field Vg : 1 B B Rmax  2  2 BR  −   rf  rf Vr() =max  () ∆ pe r  +−   (3) g ρ r  2   2 a  87 In the region of maximum winds, the Coriolis force is small compared to the pressure gradient and centrifugal 88 forces, so the Coriolis force may be ignored and Eq. (3) can be simplified into 1 B B Rmax   2 B R  −  Vr() =max  () ppe − r   (4) g ρ r  n c  a 

89 According to the distribution of typhoon wind speed, the maximum wind speed Vmax occurs at Rmax . Then,

90 replacing r in Eq. (4) with Rmax will lead to B∆ p V ≈ max ρ (5) e a 91 According to Eq. (5), the parameter B in the Holland parameter model can be calculated as follows: V2 ρ e B = max a − (6) pn p c 92 The wind field derived from the Holland parameter model is symmetrical with respect to the center of the 93 typhoon. Considering the asymmetry of the real wind field, the Holland model should be amended, taking into 94 account some additional effects. 95 The first is the amendment of the angle of inflow ( +β for the southern hemisphere, and −β for the 96 northern hemisphere). Surface friction and continuity demand that the wind flow inward across the isobars. The 97 angle of inflow is taken to be approximately 25° in the outer region, but decreases to zero near the radius of 98 maximum winds. The β proposed by Harper et al. (2001) was used in this study.    ×r <  10  r R max  Rmax   MANUSCRIPT  r  β =+×10 75  R ≤< r 1.2 R max max  Rmax  (7)  25 r> 1.2 R  max 

99 Typhoon forward motion is another important factor producing complex changes to the surface wind field 100 and asymmetric wind field of tropical cyclones. Thus, this asymmetry should be achieved by adding the forward

101 speed vector (Vt ) to the surface wind speed. The effect of forward motion weakens with distance from the eye of

102 typhoon, so Vt should multiply a weight coefficient, which decreases with distance from the cyclone center. The 103 formula proposed by Miyazaki (1977) is employed as the forward wind field: −π V  = r  x Vt exp     (8) 500000  Vy  104 where V and V are the components of forward speed of the typhoon center, with V being positive eastward x y ACCEPTED x 105 and Vy being positive northward. 106 Combining Eq. (3) and (8) yields the model’s wind field: −sin (θ + β )  = + VcVM1g  cV 2t (9) cos ()θ+ β  β 107 where c1 and c2 are the correction coefficients, set to be as 0.8 and 1.0, respectively, is the angle between 108 the gradient wind and sea surface wind, and θ is the angle between the x axis and the line connecting the 109 computing point and typhoon center.

Water Science and Engineering 3

ACCEPTED MANUSCRIPT 110 The input parameters other than Rmax for the Holland model can be derived from the CMA-STI Best Track

111 Dataset for Tropical Cyclones. Rmax can be calculated as follows (Carr and Elsberr, 1997):   0.0873()ϕ − 28 + 12.22  R =28.52tanh  + 0.2V + 37.22 max − t (10) pn p c   exp   33.86   112 The range of influence of typhoons is hundreds of kilometers in general. Parametric models are good at 113 describing the wind field and wind pressure within the range affected by typhoons, but the surrounding ambient 114 wind field differs greatly from the parametric model result due to the influence of other meteorological systems. 115 This difference can be eliminated by considering the background wind field (Zhao, 2010). The Japanese 55-year 116 reanalysis (JRA-55), the second generation reanalysis carried out by the Japan Meteorological Agency (Harada et 117 al. 2016; Kobayashi and Iwasaki, 2016), was employed as the background wind field. The JRA-55 includes 118 reanalyzed meteorological data from the period of 1958 to 2012, and has 640 and 320 grids in the longitude and 119 latitude directions, respectively, with a resolution of 0.5625° and a time interval of 6 h, providing the typhoon 120 background wind field and pressure field. 121 The background wind field and typhoon model wind field can be combined as follows: = − + VC(1 eV ) M eV Q (11)

122 where VC is combined wind field, VQ is background wind field, and e is the weight coefficient used to smooth 123 the two wind fields’ connection, e = c4/(1 + c4 ), c = r/(10 × R) . 124 The model results were verified with wind speeds and wind directions of meteorological stations along the 125 Jiangsu coast during typhoon No. 9711, and the good agreement indicated that parameters adopted by the model 126 and the simulation method were appropriate.

127 2.1.2. Typhoon storm surge model along Jiangsu coast-ADCIRC

128 ADCIRC (ADvanced CIRCulation) is a 2D finite-element hydrodynamic model, developed by the University 129 of North Carolina in the United States, and has very broad applicationsMANUSCRIPT (Luettich et al., 1992). 130 Fig. 1 shows the mesh of a typhoon storm surge mode l along the Jiangsu coast. The offshore boundary is 131 basically parallel to the Jiangsu coastline. The mesh is finer near the coast with the minimum grid distance of 600 132 m, and coarser in the open sea with a grid distance of 10 km. Local mesh refinement was carried out in the radial 133 sand ridge area, which has complex topography and many tidal channels, and the minimum grid distance was 134 about 170 m (Fig. 1). There were 45872 elements and 23485 nodes in the mesh. 135

ACCEPTED

136 137 Fig. 1. ADCIRC model mesh and local mesh refinement in the radial sand ridges area 138 The model time step was three seconds, the horizontal viscosity diffusion coefficient was 10 m2/s, and the 139 bottom friction coefficient adopted the hybrid friction formulation (Luettich and Westerink, 2015):

4 Water Science and Engineering

γ Af CCEPTED MANUSCRIPT θ θ H  f  f = + BREAK (12) Cdb2d C fmin 1    H  

140 where Cdb2d is the bottom friction coefficient, Cfmin is the minimum friction coefficient (in deep waters), θ 141 H BREAK is the control water depth, and f and rf are coefficients. The friction coefficient satisfies the < = ()γ 142 Manning formula in shallow water ( H H BREAK ), close to Cdb2d CH fmin BREAK H , and satisfies the > 143 Chezy formula in deep water ( H H BREAK ), close to Cfmin . In this study, Cfmin was set to be 0.0015, θ 144 H BREAK was set to be 1.0 m, and f and rf were set to be 10 and 4/3, respectively. 145 The surface wind stress was calculated as follows: τ ρ sx = Ca W W ρds ρ 10 10 −x (13) 0 0 τ ρ sy = Ca W W ρds ρ 10 10 − y (14) 0 0

146 where the surface wind drag coefficient Cds was calculated as follows (Luettich and Westerink, 2015): = ×( + ) Cds 0.001 0.75 0.067 W 10 (15) 147 The dry-wet boundary was determined by a node-element combination approach. When all the nodes of one 148 element were wet , the element was included in the calculation. The dryness-wetness of a node was determined

149 mainly by two parameters. First, H0 , the minimum wetness height (between 0.01 and 0.10 m in general, set to be

150 0.1 m here), was used to determine wetness or dryness for a node. Second, U min , the minimum wetting velocity, 151 set to be 0.05 m/s. When a node satisfied both the water level and wetting velocity conditions simultaneously, it 152 turned into wetting node. The dry-wet elements information was first updated before each calculation, demanding 153 that the time step of ADCIRC should not be too large. However, the matrix was solved by the finite element 154 method, and the calculation still proceeded very quickly. 155 2.2. Model verification MANUSCRIPT 156 The model coupling astronomical tide and storm surge was verified with three typhoons, No. 9711, No. 0012, 157 and No. 1210. The Northwest Pacific Ocean Tide Model (large model) provided the coupled water level at the 158 open boundary of the small model. The verification with typhoon No. 0012 in 2000 (Fig. 2) shows that the model 159 results agree with the measured data at the and Lüsi tide stations along the Jiangsu coast. The other 160 verifications are described in Wang (2015). The good agreement indicated that the boundary conditions, wind field 161 data, and parameters adopted by the model were appropriate.

ACCEPTED

Water Science and Engineering 5

ACCEPTED MANUSCRIPT

162 163 Fig. 2. Tide level comparison for typhoon No. 0012 in 2000.

164 2.3. Storm surge simulation along Jiangsu coastal area

165 The Holland parameter model calculated the wind field and pressure field of a hypothetical typhoon, 166 providing the atmospheric drive force for the other two sub-models. By not including the astronomical tide at the 167 open boundary, the large model calculated the pure storm surgeMANUSCRIPT under the effect of the wind field and pressure 168 field of the hypothetical typhoon, which then provided the pure storm surge as the water level at open boundaries 169 of the small model. Finally, the ADCIRC model computed the pure storm surge along the Jiangsu coast. Tidal 170 stations are shown in Fig. 3. 171 During the hypothetical typhoon, there is no measured background wind field to be used to amend the 172 ambient wind field. Thus, the surge process curve is different to some extent, with the result considering the 173 background wind field.

ACCEPTED

174 175 Fig. 3. Tidal stations along the Jiangsu coast 176 3. Results and discussions

177 The typhoons affecting the Jiangsu coast can be classified into three types: typhoons making straight

6 Water Science and Engineering

178 landfall, typhoons active in the offshoreACCEPTED area, and typhoons MANUSCRIPT moving northward after landfall, corresponding to 179 typhoon No. 1210, No. 0012, and No. 9711, respectively (Fig. 5 in Part I). The storm surge along the Jiangsu 180 coastal area during these three types of typhoons was modeled and investigated as follows. 181 3.1. Typhoons making straight landfall

182 The typhoon making straight landfall can be divided into three types according to the path (Fig. 6 in Part I). 183 The typhoon with path A1 makes landfall near the Yanwei station, and hits the Haizhou Bay area directly. The 184 typhoon with path A2 makes landfall near the Chenjiawu, which is between Haizhou Bay and the radial sand ridge 185 area. The typhoon with path A3 makes landfall at the north bank of the Yangtze Estuary. 186 The model simulated the typhoons from 00:00 on August 1, 2012 to 00:00 on August 4, 2012 (Greenwich 187 Mean Time). The time series of surge levels are shown in Fig. 4. For the typhoon making straight landfall, the 188 largest surge occurred at the right side of landfall point (the right side of the typhoon heading direction) when the 189 typhoon was making landfall. The surge duration and length are listed in Table 1 when the surge induced by the 190 three typhoons was larger than 1 m or 2 m. 191 In terms of surge area and duration, the surge area induced by the typhoon making landfall at Haizhou Bay 192 was relatively larger, mainly due to the coastline and topography of Haizhou Bay. Based on the surge amplitude, 193 with a probable surge level of 3.4 m, the typhoon with path A3 caused the most severe threat to the Jiangsu coast 194 of all the straight landfall typhoons. 195 Table 1 Surge duration and length caused by typhoon making straight landfall Path Criterion Duration Coastal section Length Rizhaogang-Abandoned Yellow > 1 m 6-7 h 200 km River Estuary A1 Lanshangang-Abandoned Yellow > 2 m 2-3 h 140 km River Estuary Abandoned Yellow River > 1 m 3-7 h 180 km Estuary-Jianggang A2 Abandoned Yellow River > 2 m 2 h 60 km Estuary-Xinyanggang > 1 m 2-9 h Jianggang-Xinyanggang-Lüsi 100 km A3 > 2 m 2-4 h Jianggang-Xinyanggang-LüsiMANUSCRIPT 60 km 196 197

ACCEPTED

Water Science and Engineering 7

ACCEPTED MANUSCRIPT

MANUSCRIPT

198 199 Fig. 4. Time series of surge levels at coastal observation stations for typhoons making straight landfall. 200 3.2. Typhoons active in offshore area

201 The occurrence frequency of typhoons active in the offshore area is in the middle of that of all the three types 202 of typhoons. Based on typhoon No. 0012, three theoretical typhoons active in the offshore area with different 203 paths were modeled (Fig. 8 in Part I). 204 The model time was three days from 12:00 on August 29, 2000 to 12:00 on September 1, 2000 (Greenwich 205 Mean Time). The time series of surge levels during these three theoretical typhoons are shown in Fig. 5. 206 ACCEPTED

8 Water Science and Engineering

ACCEPTED MANUSCRIPT

MANUSCRIPT

207 208 Fig. 5. Time series of surge levels at coastal observation stations for typhoons active in offshore area 209 The surge curves of the coastal observation stations were different for the typhoons active in the offshore 210 area. It is worthwhile to notice that when the typhoon left the Jiangsu coastal area and arrived at the Korean 211 Peninsula or Liaodong Peninsula, the surge that propagated from the Peninsula would arrive at the 212 Jiangsu coastal area and would continue to propagate southward. This surge oscillation could reach 0.8 m 213 sometimes, a fact that should be investigated and noted for the purpose of coastal disaster prevention and relief. 214 In general, when the typhoons were active170 km away from the Jiangsu coastline, the induced surge level in 215 the Jiangsu coastal area was not very high, mostly less than 1 m. When the typhoons were closer to the Jiangsu 216 coastline, the induced surges were larger. When the typhoon was moving just along the coastline, the induced 217 surge was quite significant,ACCEPTED up to 3.47 m. Moreover, the surge caused by this type of typhoon was not large at 218 the straight coastline section, but very large at Haizhou Bay and the radial sand ridge area. 219 3.3. Typhoons moving northward after landfall

220 The occurrence frequency of typhoons moving northward after landfall is the largest of the three types of 221 typhoons. On the basis of typhoon No. 9711, this type of typhoon is classified based on four paths (Fig. 9 in Part 222 I). The model run duration was seven days from 00:00 on August 15, 1997 to 00:00 on August 22, 1997 223 (Greenwich Mean Time). The time series of surge levels during these four theoretical typhoons are shown in Fig. 224 6. Water Science and Engineering 9

225 The range of influence of typhoonsACCEPTED is generally hun MANUSCRIPTdreds of kilometers. Thus, the wind speed of a typhoon 226 with path C1 at the Jiangsu coastal area becomes very small without consideration of the background wind field, 227 and the induced surge was almost zero. However, in the real process of typhoon No. 9711, there was strong 228 landward and northeast wind lasting a long time throughout the Jiangsu coastal area, which could not be 229 reproduced in the theoretical typhoon. 230 In terms of the surge area, the surge induced by a typhoon with path C2 mainly occurred at Haizhou Bay, and 231 the coastline section of 120 km between Rizhaogang and Yanweigang station had a surge level larger than 1 m. A 232 surge induced by a typhoon with path C3 also mainly occurred at Haizhou Bay. The surge induced by a typhoon 233 with path C4 mainly occurred at the coastline section between the Chenjiawu and L üsi stations, and the length of 234 the coastline section was about 150 m. 235 In terms of surge amplitude, the typhoon exiting the sea near the radial sand ridge area had the most 236 significant influence on the Jiangsu coast of all the typhoons moving northward after landfall, and the typhoon 237 with path C4 can cause a surge level up to 2.09 m in this area.

MANUSCRIPT

ACCEPTED 238 239 Fig. 6 Time series of surge levels at coastal observation stations for typhoons moving northward after landfall 240 3.4. Spatial distribution of storm surge

241 In terms of spatial distribution of storm surges caused by the hypothetical typhoons, the Jiangsu coastal area 242 was divided into the radial sand ridge area, the middle straight coastline area, and the Haizhou Bay area from 243 south to the north. The storm surge level caused by the typhoon making straight landfall can be large, up to 2.5 m 244 in the Haizhou Bay area, 2.6 m in the middle straight coastline area, and 3.4 m in the south radial sand ridge area. 10 Water Science and Engineering

245 The storm surge level induced by the typhoonACCEPTED active in MANUSCRIPTthe offshore area was large, up to 3.47 m, occurring at the 246 radial sand ridge area. The storm surges level caused by typhoons moving northward after landfall were as large 247 as 1.32 m and 2.09 m in the Haizhou Bay area and the radial sand ridge area, respectively. Generally speaking, the 248 storm surge level caused by each type of typhoon was largest at the radial sand ridge area and decreased 249 northward. 250

251 4. Conclusions

252 A 2D astronomical tide and storm surge coupling model was established using the Holland parameter model, 253 Northwest Pacific Ocean Tide Model, and ADCIRC model, in order to simulate the typhoon storm surges along 254 the Jiang coastal area. Based on the 65-year Best Track Dataset for Tropical Cyclones, the typhoons affecting 255 Jiangsu Province were investigated and the parameters of various typical typhoons were obtained. The surge 256 distribution of the Jiangsu coastal area under theoretical typhoons was modeled in order to obtain the storm surge 257 caused by the theoretical typhoons along the Jiangsu coast. 258 The frequency of typhoons making straight landfall was smallest but the induced surge was relatively large. 259 The surge caused by the typhoon active in the offshore area was lowest in general, but the extreme high storm 260 surge could occur in the rare event that the typhoon moved just along the coastline. The frequency of typhoons 261 moving northward after landfall was the highest, and the induced surge was relatively large. 262 In terms of spatial distribution, the radial sand ridge area was more seriously affected by the three types of 263 typhoons and the storm surge was the largest. The storm surges level caused by typhoons making straight landfall, 264 by typhoons active in the offshore area, and by typhoons moving northward after landfall were large, up to 3.4 m, 265 3.47 m, and 2.09 m, respectively, and all occurred in the radial sand ridge area. The storm surge caused by each 266 type of typhoon decreased northward. 267 The astronomical tide and storm surge are nonlinearly coupled. Their interaction is complicated, and thus it is 268 necessary to study the nonlinear pattern in terms of coupled tide-storm surge along the Jiangsu coast in the future 269 research. MANUSCRIPT 270 References

271 Carr L.E., Elsberr Y.L., 1997. Models of tropical cyclone wind distribution and beta-effect propagation for application to the 272 tropical cyclone track forecasting. Monthly Weather Review, 125, 3190-3209. 273 http://dx.doi.org/10.1175/1520-0493(1997)125<3190:MOTCWD>2.0.CO;2. 274 Harada, Y., Kamahori, H., Kobayashi, C., Endo, H., Kobayashi, S., Ota, Y., Onoda, H., Onogi, K., Miyaoka, K., Takahashi, K., 275 2016. The JRA-55 Reanalysis: Representation of Atmospheric Circulation and Climate Variability. Journal of the 276 Meteorological Society of Japan, Ser. II 94(3), 269-302. http://doi.org/10.2151/jmsj.2016-015. 277 Harper B, Hardy T, Mason L., 2001. Queensland Climate Change and Community Vulnerability to Tropical Cyclones: Ocean 278 Hazards Assessment Stage 1, Department of Natural Resources and Mines, Queensland. 279 Kobayashi, C., Iwasaki, T., 2016 Brewer-Dobson circulation diagnosed from JRA-55. Journal of Geophysical Research: 280 Atmospheres, 121(4), 1493-1510. http://doi.org/10.1002/2015JD023476. 281 Liang, H., Zou, X., 2004. Calculation of overland risk of storm surge off the Haikou Bay. Marine Science Bulletin, 23(3), 20-26. 282 http://dx.doi.org/10.3969/j.issn.1001-6392.2004.03.003 (in Chinese). 283 Luettich, R.A., Westerink, J.J., Scheffner, N.W., 1992. ADCIRC: An Advanced Three-dimensional Circulation Model for Shelves 284 Coasts and Estuaries, Report 1: Theory and Methodology of ADCIRC-2DDI and ADCIRC-3DL, Dredging Research 285 Program Technical ACCEPTEDReport DRP-92-6, U.S. Army Engineers Waterways Experiment Station, Vicksburg. 286 Luettich, R.A., Westerink, J.J., 2015. Advanced Circulation Model for Oceanic, Coastal and Estuarine Waters, User’s Manual-v51. 287 Miyazaki, M., 1977. Marine Physics . Tunghai University Press, Taizhong. 288 Wang J., 2015. Typhoon Storm Surge Simulation and Possible Maximum Storm Surge along Jiangsu Coast. Master Dissertation. 289 Hohai University, Nanjing (in Chinese). 290 Wang, X., Zhai, J., Dong, P., Fan, F., 2015. A study of storm surge in Jiangsu sea waters based on a typhoon parameter model. 291 Hydro-Science and Engineering, (2), 61-66. http://dx.doi.org/10.16198/j.cnki.1009-640x.2015.02.010 (in Chinese). 292 Wu, S., Wang, X., Yu, F., Dai, M., Ye, L., Chen, X., Yin, J., 2002. Extratropical storm surge and calculation of probable maximum 293 extratropical storm surge in Lianyungang. Acta Oceanologica Sinica, 24(5), 8-18. 294 http://dx.doi.org/10.3321/j.issn:0253-4193.2002.05.002 (in Chinese).

Water Science and Engineering 11

295 Ying, M., Zhang, W., Yu, H., Lu, X.Q., Feng,ACCEPTED J.X., 2014. An MANUSCRIPT overview of the China Meteorological Administration tropical 296 cyclone database. Journal of Atmospheric and Oceanic Technology, 31(2), 287-301. 297 http://dx.doi.org/0.1175/JTECH-D-12-00119.1. 298 Yu, F., Wang, X., Dai, M., 2002. Analysis and numerical simulation of some severe extraropical strom surges influenced on 299 Lianyungang. Marine Forecast, 19(1), 113-122. http://dx.doi.org/10.3969/j.issn.1003-0239.2002.01.018 (in Chinese) 300 Yu, L., Chen, K., Lu, P., Zhang, J., 2014. A study of storm surge in radial sand ridges under sea level rising. Hydro-Science and 301 Engineering, (6), 52-57. http://dx.doi.org/doi: 10.16198/j.cnki.1009-640x.2014.06.003 (in Chinese). 302 Zhang, C., Zheng J.H., Dong X.W., Cao K., Zhang J.S., 2013. Morphodynamic response of Xiaomiaohong tidal channel to a 303 coastal reclamation project in Jiangsu Coast, China. Journal of Coastal Research, SI65, 630-635. 304 http://dx.doi.org/10.2112/SI65-107.1. 305 Zhang, W., Chen, C., Li, X., Zhang, J., 2012. The impact of coastal engineering on high water level of storm surges in Bohai Bay, 306 The Ocean Engineering, 30(2), 72-78. http://dx.doi.org/10.16483/j.issn.1005-9865.2012.02.018 (in Chinese). 307 Zhang, X., 2008. Numerical Model Study on the Storm Surge along the Jiangsu Coast. Ph. D. Dissertation. Hohai University, 308 Nanjing (in Chinese). 309 Zhao, H., Song, Z., Xu, F., Cheng, C., 2010. Numerical simulation of typhoon waves in the South China Sea: A case study of 310 typhoon Chanchu. The Ocean Engineering, 28(3), 128-134. http://dx.doi.org/10.3969/j.issn.1005-9865.2010.03.021 (in 311 Chinese).

MANUSCRIPT

ACCEPTED

12 Water Science and Engineering