ZHANG Mingzong ,ZHOU Chunyan , ,ZHANG Jisheng ,ZHANG Xinzhou,and TANG Zihao

1) Key Laboratory of Ministry of Education for Coastal Disaster and Protection,Hohai University,Nanjing 210024,China

2) College of Harbour, Coastal and Offshore Engineering,Hohai University,Nanjing 210024, China

3) Key Laboratory of Port,Waterway and Sedimentation Engineering of the Ministry of Transport, Nanjing Hydraulic Research Institute,Nanjing 210029, China

4) Department of Civil and Environment Engineering,University of Auckland, Auckland 1023, New Zealand

Abstract In this study,a coupled tide-surge-wave model was developed and applied to the South Yellow Sea.The coupled model simulated the evolution of storm surges and waves caused by extreme weather events,such as tropical cyclones,cold waves,extratropical cyclones coupled with a cold wave,and tropical cyclones coupled with a cold wave.The modeled surge level and significant wave height matched the measured data well.Simulation results of the typhoon with different intensities revealed that the radius to the maximum wind speed of a typhoon with 1.5 times wind speed decreased,and its influence range was farther away from the Jiangsu coastal region;moreover,the impact on surge levels was weakened.Thereafter,eight hypothetical typhoons based on Typhoon Chan-hom were designed to investigate the effects of varying typhoon tracks on the extreme value and spatial distribution of storm surges in the offshore area of Jiangsu Province.The typhoon along path 2 mainly affected the Rudong coast,and the topography of the Rudong coast was conducive to the increase in surge level.Therefore,the typhoon along path 2 induced the largest surge level,which reached up to 2.91 m in the radial sand ridge area.The maximum surge levels in the Haizhou Bay area and the middle straight coastline area reached up to 2.37 and 2.08 m,respectively.In terms of typhoons active in offshore areas,the radial sand ridge area was most likely to be threatened by typhoon-induced storm surges.

Key words Jiangsu coast;South Yellow Sea;extreme weather events;storm surge;numerical experiments

1 Introduction

A storm surge is an abnormal rise in sea level produced by the force of the wind moving shoreward and the low pressure related to intense storms.In the coastal regions,marine disasters primarily consist of coastal erosion,red tide,earthquake tsunamis,oil spills,and storm surges.Nevertheless,storm surges are usually the most severe threat to the coastal economy (Fenget al.,2015;Needhamet al.,2015).Furthermore,the coast of China frequently suffers from storm surges.The Yellow Sea has a continental shelf that is affected by varying typhoons,monsoons,and flow circulation seasonally (Zhouet al.,2015;Yuket al.,2016).Jiangsu Province of China is located on the west shore of the South Yellow Sea (SYS).Storm surges caused by tropical cyclones,extratropical cyclones,and cold waves have struck the Jiangsu coastal region with high frequency.Furthermore,the radial sand ridges in the SYS stretches from the Sheyang River Mouth to the north bank of the Yangtze Estuary.It includes more than 70 sand ridges of various sizes,with a length of approximately 200 km from north to south and a width of approximately 90 km from east to west.Considering the complex impact of the radial sand ridges on the surge level in the SYS and the condition that storm surges possibly overlap with the high astronomical tide,the potential maximum storm surge can bring substantial direct and indirect economic losses of residents within the Jiangsu coastal area.

In the past decades,considerable attention has been paid to the analysis of the generating mechanisms of storm surges and waves in the SYS.Results from an extreme weather event dominated by an extratropical cyclone coupled with a tropical cyclone showed that the change of bed shear stress in the shallow sea is a major contributing factor to the differences in the interaction between astronomical tides and storm surges in the radial sand ridge region(Xionget al.,2019).Numerous published studies have described the convergence and divergence of the SYS amphidromic wave system and the East China Sea forward tidal wave system that occur around Jianggang.In addition,the surge level of the radial sand ridge region generally decreases as the sea level rises (Yuet al.,2013,2014).

In addition to studies on the generating mechanisms and influence of geophysical configuration,studies on the relationship between typhoon tracks and surge levels have attracted considerable interest.According to Zhenget al.(2017a),typhoons threatening the Jiangsu coastal region can be classified into three categories,including typhoons moving northward after landfall,typhoons making straight landfall,and typhoons active in offshore areas.Modeled storm surges under different typhoon tracks suggested that the surge level caused by typhoons active in offshore areas is rather high and decreases from south to north along the Jiangsu coastline (Zhenget al.,2017b).

By contrast,recent evidence has suggested that the nonlinear interaction between astronomical tides and storm surges inhibits storm surges at the high astronomical tide and promotes storm surges at the low astronomical tide (Heet al.,2017).Similarly,the interaction between astronomical tides and storm surges is associated with typhoon tracks,topography,and moving speed (Zhanget al.,2015;Xieet al.,2018;Zhanget al.,2019).By running the advanced circulation model for the SYS,the obtained results could be used in the hindcasting and forecast of storm surges induced by typhoons and extratropical cyclones,respectively (Hanet al.,2019).Whereas,several experimental studies have indicated that the effect of wave-current interaction on storm surges is prominent in shallow water.Considering the impact of wave-current interaction can improve the simulation precision of the storm surge model(Zhang and Li,1996;Ozeret al.,2000;Yuket al.,2016).Moreover,the development of an unstructured mesh wavecurrent coupled model in the northern Jiangsu radial sand ridge areas provides reliable evidence that wave distinctly influences water level (Guanet al.,2019).

Substantial research has been conducted to analyze the impact of the radial sand ridges on storm surges and the interaction between astronomical tides and storm surges in the SYS.However,extratropical cyclones and cold waves also lead to severe storm surges in the Jiangsu coastal area sometimes.Simulations of storm surges caused by extratropical cyclones and cold waves in the SYS,especially that considering the effect of wave-current interaction,are limited.

In this study,a two-dimensional,dynamically coupled tide-surge-wave model was developed.The unstructured mesh MIKE 21 hydrodynamic module (MIKE 21 HD)and the spectral waves module (MIKE 21 SW) were integrated into a strongly coupled model that considers the nonlinear effect of wave-current interaction.This coupled model aimed to simulate storm surges in the SYS induced by extreme weather events that could be classified into four categories:a tropical cyclone,a cold wave,an extratropical cyclone coupled with a cold wave,and a tropical cyclone coupled with a cold wave.Thereafter,numerical scenarios with variations of typhoon track were conducted to analyze the spatial distribution of the probable maximum surge level along the Jiangsu coast.The analysis of the potential extreme value of surge levels provided scientific references for marine structure design and coastal disaster protection.

2 Materials and Methods

2.1 Model Description

The MIKE 21/3 coupled model is used in coastal studies for simulating various hydraulic phenomena including lakes,rivers,estuaries,bays,coastal areas,and seas.MIKE 21 HD and MIKE 21 SW are two modules of the MIKE 21/3 coupled model.MIKE 21 HD is based on the depthintegrated incompressible Reynolds-averaged Navier-Stokes equations and the shallow water equations (DHI,2017a).MIKE 21 SW is used to simulate the growth,decay,and propagation from the deep to the shallow waters of swells and wind waves (DHI,2017b).On the basis of these two modules,an unstructured coupled tide-surge-wave model was developed in this study.

2.2 Model Setups

In this study,the model domain was 27.1?– 41.0?N,117.4?– 128.8?E,covering the Bohai Sea,the Yellow Sea,and the East China Sea (Fig.1(a)).Bathymetry of the area of interest is shown in Fig.1(b).The Jiangsu coastal region was divided into three sections:the Haizhou Bay area,the middle straight coastline area,and the radial sand ridge area.The middle straight coastline area stretched from the Abandoned Yellow River Mouth to the Sheyang River Mouth,the radial sand ridge area ranged from the Sheyang River Mouth to Lüsi,and the radial sand ridge area was divided into the northern and southern radial sand ridge area by Jiangjiasha ridge line.The triangular mesh had 109364 elements and 56248 nodes,with a minimum mesh size of 17.34 m and smallest area of 197.99 m2.Shoreline data were derived from the global self-consistent,hierarchical,high-resolution shoreline database (Wessel and Smith,1996).Bathymetric data around the Jiangsu coastal area were provided by field measurements,and data in the open sea were from the ETOPO1 dataset (Amante and Eakins,2009).

Fig.1 Computational domain and bathymetry:(a) model domain,(b) area of interest.

The time step of the coupled model was 30 s.For the hydrodynamic model,initial values of flow velocity and tidal level were zero.The model applied standard flood and dry,and the density type was set as barotropic.Wind friction at 10 m above sea level was calculated with the drag coefficient (Wu,1994).Sea bed resistance was set to be a Manning number with a value of 45– 65 m1/3s-1.The tidal levels at the open boundary were derived from the DHI Global Tide Model with a resolution of 0.125? ×0.125? (DHI,2017c).

For the wave module,the instationary and fully spectral formulation was used (Komenet al.,1996).The type of spectral directional discretization was set to 360? rose,and 16 directions discretized the wave action spectrum.The frequency was discretized into 25 bins logarithmically in an interval from 0.055 Hz to 0.6 Hz.The model considered the dissipation of energy caused by white capping,bottom friction using Nikuradse roughness,depth-induced wave breaking,and quadruplet-wave interaction (Battjes and Janssen,1978;Weber,1991;Komenet al.,1996).The initial condition was set to JONSWAP fetch growth expression.

Furthermore,MIKE 21 SW produced the wave radiation stress field that influenced the current field of MIKE 21 HD.Next,MIKE 21 SW synchronously simulated waves by considering the impact of the current field and water level from MIKE 21 HD.

2.3 Meteorological Forcing

This section introduces the European Center for Medium-Range Weather Forecasts (ECMWF) dataset and the Holland parameter model for modeling the hydrodynamic process and wave growth in this study.

2.3.1 ECMWF reanalysis data

ECMWF-ERA5 is a new atmospheric reanalysis product,including 137 hybrid sigma levels in the vertical with the top-level at 0.01 hPa (Hersbachet al.,2019).ERA5 data on each level have a horizontal resolution of 0.25? and cover 1979 to present,with a temporal resolution of 1 h.

The wind fields (at 10 m above mean sea level) and the pressure fields (at the mean sea level) from the ERA5 dataset were used as the meteorological forcing of the coupled model.In this research,the ERA5 dataset was used as the meteorological forcing except for the case influenced by typhoons.

2.3.2 Holland parameter model

In terms of simulating wind fields and pressure fields of tropical cyclones,the Holland parameter model exhibits acceptable accuracy and can be used extensively (Holland,1980).In addition,the Tropical Cyclone Best Track Dataset issued by the Shanghai Typhoon Institute of the China Meteorological Administration (CMA) provides typhoon information about the six-hourly track and intensity analyses (Yinget al.,2014).Thus,the meteorological forcing of typhoons can be reproduced by the Holland parameter model and the Tropical Cyclone Best Track Dataset.The gradient pressureP,according to the Holland model,is calculated as

whereP(r) is the gradient pressure at a distancerfrom the cyclone center,Pcis the central pressure,Pnis the ambient surrounding pressure field or neutral pressure,Rmaxis the radius to the maximum wind speed,andBis a shape parameter introduced by Holland to describe the radial pressure profile.

By combining the pressure field with the force balance equation,the gradient wind fieldVg(r) is expressed as

whereVg(r) is the gradient wind at a distancerfrom the cyclone center,fis the Coriolis parameter,andρa(bǔ)is the air density.

In accordance with Willoughby and Rahn (2004),Rmaxin Eqs.(1) and (2) can be given as

whereVmaxis the maximum wind speed,andφis the latitude.

In accordance with Vickeryet al.(2000),theBparameter can be obtained whenVmaxis known,that is,

The Holland model can calculate the wind and pressure data within the range influenced by the tropical cyclone.However,the wind and pressure fields of the Jiangsu coastal area are slightly inaccurate when typhoons have not yet affected this region.Moreover,Zhenget al.(2017a)argued that typhoons striking the region covering latitudes of 28?–38?N and longitudes of 116?–126.1?E are determined to be typhoons influencing the Jiangsu coastal region.When the typhoon had not yet entered the coastal area of Jiangsu,the Holland model was inaccurate in describing the wind and pressure fields in the Jiangsu coastal region.In this study,the wind and pressure fields were replaced by the ECMWF dataset before typhoons affected the Jiangsu coastal area to ensure the accuracy of the simulation.The wind and pressure fields of Typhoon Chan-hom before 12:00 on July 9,2015 were provided by the ERA5 dataset.The Holland model provided the wind and pressure fields of Typhoon Chan-hom after 13:00 on July 9,2015 (Fig.2).The composition of the wind and pressure fields of Typhoon Fung-wong was similar to that of Typhoon Chan-hom.As shown in Fig.2,in the lower right corner of the figure (typhoon center),the wind and pressure fields from the ECMWF dataset were basically the same as those calculated by the Holland model.However,the wind and pressure fields from the ERA5 dataset were more accurate,specifically in the Jiangsu coastal area than those calculated by the Holland model.Thus,the ERA5 dataset provided the wind and pressure fields before typhoons affected the coastal areas of Jiangsu.

Fig.2 Comparison of observed data from the ERA5 dataset and simulated data from the Holland model during Typhoon Chan-hom:(a),wind field from the ERA5 dataset at 12:00 on July 9,2015;(b),wind field from the Holland model at 13:00 on July 9,2015;(c),pressure field from the ERA5 dataset at 12:00 on July 9,2015;(d),pressure field from the Holland model at 13:00 on July 9,2015.

2.4 Jason-2 Altimetry Data

The OSTM/Jason-2 satellite measures every location along the ground track every approximately 9.9 days.The geophysical data record (GDR) products of the Jason-2 satellite were downloaded from the National Oceanic and Atmospheric Administration.The GDR products are fully verified products with increasing latency and accuracy compared with other products.

Measured data from Jason-2 satellite cover 66?S– 66?N,have extensive spatial coverage,and can verify significant wave heights of multiple specific locations in the coupled model.Furthermore,the wave radiation stress exerts a relatively strong impact on the precision of the storm surge simulation (Choiet al.,2013).Consequently,observed data from three Jason-2 passes (62,138,and 153) were used to validate significant wave heights reproduced by wave models (Fig.3).To ensure the accuracy of GDR data,we excluded invalid data by quality and control.In terms of frequency,the GDR products were classified into the Ku band range measurement and the C band measurement.Following Dumontet al.(2017),the Ku band range measurement was more precise than the C band measurement.Thus,the Ku band range measurement was applied in this study.

Fig.3 Ground tracks of Jason-2 satellite located in the study area during the simulation.

2.5 Model Validations for Relatively Calm Weather

This section conducts validation for relatively calm weather to determine how well the tide-wave coupled model could capture the hydrodynamic process and wave variation during calm weather.Fig.4 shows the simulatedversusthe observed data of tidal level,current speed,current direction,and significant wave height at the Liyashan station (red point in Fig.1(b)) from 6:00 on July 23,2013 to 21:00 on July 30,2013 (UTC,the same in the following).

The simulation results reveal that the bias of tidal level,current speed,current direction,and significant wave height were -0.058 m,0.135 m s-1,-38.876?,and -0.034 m,respectively.As shown in Fig.4(d),the maximum values of the observed data did not match the simulated results at some moments,probably due to the low resolution of the wind field.Nonetheless,the variation trend from the coupled simulation generally showed an acceptable agreement with the observations.Substantial evidence indicates that the coupled model could model the hydraulic phenomena and evolution of waves under calm conditions.

Fig.4 Simulated versus observed data at the Liyashan station:(a),tidal level;(b),current speed;(c),current direction;(d),significant wave height.

3 Results

3.1 Model Validations for Extreme Weather

The coupled tide-surge-wave model was developed to reproduce storm surges and waves in the SYS induced by four different types of extreme weather events.Fig.5 shows the tracks of three typhoons influencing the Jiangsu coastal area in this study.

Fig.5 Tracks of Typhoons Fung-wong (No.1416),Chanhom (No.1509),and Malakas (No.1616).

3.1.1 Validation for a tropical cyclone

Typhoon Fung-wong (No.1416) occurred in September 2014 and then intensified to a severe tropical storm.The maximum wind speed reached up to 28 m s-1,and the lowest pressure reached 982 hPa.The wind speed weakened to 23 m s-1,and the pressure increased to 990 hPa when the typhoon made landfall in Shanghai,China.The comparison between the simulated and observed tidal levels and significant wave heights are shown in Fig.6(a).The observed significant wave heights were from Jason-2 cycle 229 pass 62,which the satellite passed through approximately at 5:00 September 22,2014.The significant wave height data in Fig.6(a) were measured by Jason-2 satellite at this specified time and along this appointed track.Moreover,the location of observed data moved along the Jason-2 ground track.

The maximum tidal level at the Liyashan station reached 3.09 m,and the maximum significant wave height from the Jason-2 satellite occurring at the southern part of the East China Sea reached up to 6.84 m (Fig.6(a)).During the passage of Typhoon Fung-wong,the extreme value of surge level reached up to 1.36 m in the radial sand ridge area,whereas the maximum surge levels in the middle straight coastline area and the Haizhou Bay area were lower,that is,up to 0.54 and 0.48 m,respectively.This phenomenon occurred because the moving direction of Typhoon Fung-wong changed from northward to northeastward after the typhoon made landfall in Shanghai.Then,Typhoon Fung-wong passed through the SYS and the southern sea area of the Korean Peninsula (Fig.5).Therefore,the wind speed in the radial sand ridge area was relatively large,and the surge level in the radial sand ridge area was the highest.The northeast wind in the early stage of the simulation was the onshore wind for the Jiangsu coastal area,inducing a storm surge.Afterward,influenced by the moving direction change of the typhoon,the northeast wind gradually turned to northwest wind starting at 7:00 on the 23rd.Then,a negative storm surge occurred in the Jiangsu coastal area due to the offshore wind.For the modeled storm surge and wave caused by Typhoon Fung-wong,the tidal level had a bias of -0.186 m,and the correlation coefficient was 0.981 (Table 1).The significant wave height had a bias of -0.152 m,and the correlation coefficient was 0.943 (Table 2).

Fig.6 Comparison of simulated and observed data from the Liyashan Station and Jason-2 satellite during extreme weather events:(a),Typhoon Fung-wong in September 2014;(b),Typhoon Chan-hom in July 2015;(c),a cold wave in January 2016;(d),an extratropical cyclone coupled with a cold wave in April 2013;(e),Typhoon Malakas coupled with a cold wave in September 2016.

Table 1 Error analysis of the simulated and observed tidal levels at the Liyashan station

Table 2 Error analysis of the simulated and observed significant wave heights from Jason-2 satellite

Typhoon Chan-hom (No.1509) was formed on June 30,2015,in the Northwest Pacific and strengthened into a super typhoon on July 9.The lowest pressure was approximately 935 hPa with 55 m s-1as the maximum wind speed.Moreover,Typhoon Chan-hom made landfall at Zhujiajian Island of Zhejiang Province and weakened into a strong typhoon.The typhoon then moved northward and entered the Korean Peninsula.The comparison between the simulated and observed tidal levels and significant wave heights are shown in Fig.6(b).The observed significant wave heights were from Jason-2 cycle 258 pass 138,which the satellite passed through approximately at 18:00 July 9,2015.

As shown in Fig.6(b),the maximum tidal level at the Liyashan Station reached up to 2.91 m.Influenced by Typhoon Chan-hom,the northeastern wind prevailed and led to a storm surge along the Jiangsu coastline.Starting 16:00 July 11,the northeast wind turned to the westerly wind.Then,the westerly wind,which was offshore wind respective to the Jiangsu coast,induced a negative storm surge in the Jiangsu coastal area.Typhoon Chan-hom made its way northeast after making landfall in Zhujiajian Island and finally weakened into a tropical depression in North Korea.Thus,the typhoon had a great impact on the radial sand ridge area,and the surge level in the radial sand ridge area was the highest.During the passage of Typhoon Chanhom,the probable extreme values of surge level reached up to 1.53 m in the radial sand ridge area,0.75 m in the middle straight coastline area,and 0.70 m in the Haizhou Bay area.For the model results,the tidal level had a bias of -0.022 m,and the correlation coefficient was 0.960(Table 1).The significant wave height had a bias of-0.764 m,and the correlation coefficient was 0.983 (Table 2).

3.1.2 Validation for a cold wave

According to CMA,a weather event called cold wave is a rapid fall of regional averaged temperature by more than 8 ℃ within 24h or by more than 10 ℃ within 48h,and the lowest temperature all drops below 5℃ .The cold waves along the Jiangsu coast usually happen in spring,autumn,and winter.The sustainable wind and pressure drop affect the SYS and bring storm surges in the Jiangsu coastal region.Storm surges and waves caused by cold waves may also impose a considerable threat to offshore facilities and coastal structures.A 7– 8 grade northwest wind occurred in the SYS due to the cold wave from January 22,2016,to January 25,2016,and the maximum wind speed reached 19.5 m s-1.The comparison between the simulated and observed tidal levels and significant wave heights are shown in Fig.6(c).The observed significant wave heights were from Jason-2 cycle 278 pass 138,which the satellite passed through approximately at 1:00 January 24,2016.

Fig.6(c) shows that the maximum tidal level at the Liyashan station reached 2.73 m,and the maximum significant wave height occurring at the northern part of the East China Sea reached up to 7.46 m.At the initial stage of the simulation,the onshore wind caused a storm surge in the Jiangsu coastal area.Nevertheless,the northwest wind,which was offshore wind respective to the Jiangsu coast,prevailed in most of the period of the cold wave and brought a negative storm surge in the Jiangsu coastal region.The surge level caused by the cold wave was relatively low.The extreme values of storm surge were 0.65 m in the radial sand ridge area,0.13 m in the middle straight coastline area,and 0.23 m in the Haizhou Bay area.For the simulation results,the tidal level had a bias of -0.030 m,and the correlation coefficient was 0.981 (Table 1).The significant wave height had a bias of -0.001 m,and the correlation coefficient was 0.972 (Table 2).

3.1.3 Validation for an extratropical cyclone coupled with a cold wave

Although storm surges induced by extratropical cyclones are usually not as severe as those caused by tropical cyclones,they are also possibly destructive in the coastal region (Colleet al.,2015).Thus,extreme weather conditions,such as an extratropical cyclone coupled with a cold wave,deserve to be investigated.

A seven-grade north wind occurred in the SYS from April 5 to April 7,2013,due to an extratropical cyclone coupled with a cold wave.Starting 0:00 April 6,the north wind gradually turned to northwest wind,with a maximum wind speed of 18.4 m s-1.The comparison between the simulated and observed tidal levels and significant wave heights is shown in Fig.6(d).The observed significant wave heights were from Jason-2 cycle 175 pass 138,which the satellite passed through approximately at 18:00 April 7,2013.

Fig.6(d) shows that the extreme level of the tidal level was 2.75 m.Influenced by the extratropical cyclone and the cold wave,the northeast wind turned to northwest wind and then to the southerly wind.Consequently,the northwest wind induced a storm surge in the Jiangsu coastal region first.Then,the southerly wind brought a negative storm surge along the Jiangsu coast.The maximum surge level in the radial sand ridge area was the largest,reaching up to 1.31 m.The potential maximum surge levels were 0.66 and 0.56 m in the middle straight coastline area and the Haizhou Bay area,respectively.For the modeled results,the tidal level had a bias of -0.001 m,and the correlation coefficient was 0.958 (Table 1).The significant wave height had a bias of -0.230 m,and the correlation coefficient was 0.902 (Table 2).

3.1.4 Validation for a tropical cyclone coupled with a cold wave

Typhoon Malakas (No.1616) formed in the Pacific Ocean on September 11,2016,and finally made landfall southwest of Japan (Fig.5);the maximum wind speed reached up to 50 m s-1.Although Typhoon Malakas did not make landfall in the coastal area of China,the interaction of Typhoon Malakas and the cold wave induced a relatively massive storm surge in the offshore area of Jiangsu Province from September 16 to September 19.Typhoon Malakas did not directly influence the Jiangsu coastal area,and the Holland parameter model could not accurately capture the real wind field.Thus,the ERA5 dataset was adopted as the meteorological forcing of this coupled model.The extreme value of surge level at the Lüsi Station reached 1.12 m.The wave height reported in the SYS was 2.5–3.0 m from September 18 to 20.The data collected from the buoy of the State Oceanic Administration of China indicated that the maximum significant wave height was 3.0 m because of the six-grade northeast wind (State Oceanic Administration of China,2017).The comparison between the simulated and observed tidal levels and significant wave heights is shown in Fig.6(e).The observed significant wave heights were from Jason-2 cycle 302 pass 153,which the satellite passed through approximately at 15:00 September 18,2016.

Fig.6(e) shows that the maximum tidal level at the Liyashan station reached up to 4.02 m.The maximum surge levels caused by the tropical cyclone and the cold wave were 1.31,0.80,and 0.67 m in the radial sand ridge area,the middle straight coastline area,and the Haizhou Bay area,respectively.During the entire simulation period,a negative storm surge occurred in the southern radial sand ridge area because of the southwest wind.At the same time,the northeast wind resulted in a storm surge in the northern radial sand ridge area,the middle straight coastline area,and the Haizhou Bay area.From the simulation results,the tidal level had a bias of -0.184 m,and the correlation coefficient was 0.977 (Table 1).The significant wave height had a bias of 0.055 m,and the correlation coefficient was 0.918 (Table 2).

The spatial distribution of the maximum surge level caused by the five extreme weather events is shown in Fig.7.The extreme values of surge level induced by extreme weather events all occurred in the radial sand ridge area.Thus,coastal disasters in the radial sand ridge area should be prevented.Among the five extreme weather events,the surge level caused by Typhoon Chan-hom was the largest,reaching up to 1.53 m.As shown above,the maximum wind speed of Typhoon Chan-hom reached 55 m s-1.The maximum wind speed of Typhoon Fung-wong only reached up to 28 m s-1,and Typhoon Malakas did not make landfall in the coastal areas in China.Surprisingly,the surge level induced by Typhoon Chan-hom was not much larger than the surge levels caused by the other two.This phenomenon occurred because although Typhoon Chan-hom is a super typhoon,it mainly affected the Zhejiang coastal area,and its influence on the Jiangsu coastal region was relatively small.In addition,the storm surge induced by the cold wave and the extratropical cyclone may also result in severe damage.Thus,these two types of storm surges cannot be ignored.

Fig.7 Spatial distribution of the maximum surge level caused by the five extreme weather events.

Tables 1 and 2 show that based on the model results of the five cases,the mean bias and correlation coefficient were -0.085 m and 0.971,respectively,for tidal level (Table 1),whereas the mean bias and correlation coefficient were -0.126 m and 0.944,respectively,for significant wave height (Table 2).The results indicated that the coupled tide-surge-wave model could be used extensively in reproducing storm surges and waves induced by extreme weather events in the SYS.

3.2 Effect of Wave-Current Interaction on Storm Surges

Wave-current interaction usually has a specific impact on the simulation of storm surge evolution.Fig.8 shows the influence of wave radiation stress on the storm surge caused by Typhoons Fung-wong and Chan-hom.The wave radiation stress had a more pronounced influence on the storm surge during Typhoon Chan-hom (Fig.8).Considering the impact of wave-current interaction reduced the average error of storm surge stimulation during Typhoon Fung-wong by 1.2 cm and during Typhoon Chan-hom by 3.8 cm.

Fig.8 Comparison of simulated and observed data from the Liyashan Station during (a) Typhoon Fung-wong and (b) Typhoon Chan-hom.

3.3 Numerical Experiments of Typhoons with Different Intensities

Among all the extreme weather events,typhoons most likely induce substantial storm surges in coastal areas.Numerical experiments of hypothetical typhoons may offer valuable information on preventing and controlling disasters in coastal regions.Moreover,the surge level induced by typhoons active in offshore areas along the Jiangsu coast is the highest (Zhenget al.,2017b).On the basis of Typhoon Chan-hom,typhoons with three different intensity levels were designed to explore surge levels along the Jiangsu coastal area.Table 3 presents a summary of numerical experiments with different intensities.

Table 3 Numerical experiments of typhoons with different intensities

As shown in Fig.9,compared with the surge level caused by Typhoon Chan-hom,the surge level induced by Typhoon A1,which reached up to 0.97 m,was lower.By comparison,the surge level caused by Typhoon A2 reached a maximum of 1.76 m.According to Eq.(3),when the wind speed increases,the radius to the maximum wind speedRmaxdecreases.In addition,Typhoon Chan-hom is a typhoon active in offshore areas.The path of Typhoon A3 is the same as that of Typhoon Chan-hom because the radius to the maximum wind speed of Typhoon A3 decreased to the extent that its influence range was farther away from the Jiangsu coastal region,leading to a weaker impact on the surge levels.Therefore,the maximum surge level caused by Typhoon A3 was the same as that of Typhoon A2.

Fig.9 Spatial distribution of the maximum surge level caused by Typhoon Chan-hom and typhoons with different intensities.

Numerical experiments were conducted on the typhoon tracks because storm surges caused by typhoons are related to the intensity and track of the typhoon.

3.4 Numerical Experiments of Typhoons with Different Tracks

According to Rego and Li (2010),storm surges caused by typhoons depend mainly on the track.Hypothetical typhoons with different tracks were designed on the basis of Typhoon Chan-hom,which is the typhoon active in offshore areas,to investigate differences in storm surges in the SYS due to the effect of various tracks.The central pressure and the maximum wind speed of hypothetical typhoons were the same as those of Typhoon Chan-hom.Furthermore,the radius to the maximum wind speedRmaxand the shape parameterBof hypothetical typhoons were calculated by Eqs.(3) and (4).Eight hypothetical typhoons were obtained by moving the track of Typhoon Chan-hom parallel northward and southward by 1? each time (Fig.10).

Fig.10 Tracks of Typhoon Chan-hom and hypothetical typhoons.

Fig.11 presents an overview of the maximum surge level of Typhoon Chan-hom and hypothetical typhoons in the Haizhou Bay area,the middle straight coastline area,and the radial sand ridge area.The typhoon along path 2 led to the probable extreme value of surge level in the radial sand ridge area,which reached up to 2.91 m.Thus,among all the typhoons in the numerical experiments,the typhoon along path 2 was the most dangerous typhoon track influencing the Jiangsu coastal area.

Fig.11 Spatial distribution of the maximum surge level caused by Typhoon Chan-hom and typhoons with different tracks.

The potential maximum surge level and wind field caused by the typhoon along path 2 along the Jiangsu coast are shown in Fig.12(a).The northeast wind was onshore wind respective to Jiangsu coastal region,and the wind speed in the radial sand ridge area was larger than that of the middle straight coastline area and the Haizhou Bay area.Thus,the surge level increased southward,and the probable maximum surge level occurred in the sea area around the Rudong coast.The significant wave height and wind field at the time of the maximum surge level occurring in the offshore area in Jiangsu Province are shown in Fig.12(b).Given the strong wind,the significant wave height was relatively high in the sea area around Lüsi,and the wave height decreased shoreward.The maximum significant wave height reached 14.34 m in the open sea.

Fig.12 Distribution of (a) surge level and (b) significant wave height at the time of the maximum surge level occurring in the Jiangsu coastal region.

4 Discussion

4.1 Impact of Various Tracks on the Maximum Surge Level

As shown in Fig.11,when the typhoon track moved southward (paths 7 and 8),the maximum surge level was the lowest,i.e.,the threat degree was low because the overlap area for the region affected by the typhoon and the Jiangsu coastal region was small.When the typhoon track moved northward,the overlap area for the region affected by the typhoon and the Jiangsu coastal region also changed.

The wind speed in the Jiangsu coastal region caused by the typhoon along path 1 was greater than that of Typhoon Chan-hom;thus,the typhoon along path 1 induced a higher surge level.Among all the typhoons in the numerical experiments,the typhoon along path 2 resulted in the most severe storm surge in the radial sand ridge area.The typhoon along path 2 caused the largest wind speed near the Rudong coast among all typhoons,and the complex topography of the Rudong coast was conducive to storm surges.The combination of the mighty onshore wind and the unique geophysical configuration of the Rudong coast resulted in the largest surge level,reaching 2.91 m.

The typhoons with paths 3 and 4 also led to severe storm surges on the Rudong coast,but the main area affected by the typhoons with paths 3 and 4 was the southern part of the middle straight coastline area.Therefore,compared with the surge level caused by the typhoon along path 2,the surge levels induced by the typhoons with paths 3 and 4 were higher in the middle straight coastline area and the Haizhou Bay area.

Compared with the typhoon along path 2,the typhoon along path 5 caused a lower surge level in the radial sand ridge area.However,the typhoon along path 5 resulted in higher wind speed in the Haizhou Bay area and the middle straight coastline area.Thus,the typhoon along path 5 caused the largest surge levels in the middle straight coastline area and the Haizhou Bay area,which reached up to 2.08 and 2.37 m,respectively.Fig.10 shows that the overlap area for the region influenced by the typhoon and the Jiangsu coastal region became smaller when moving to path 5 parallel northward.Consequently,the extreme value of the storm surge induced by path 6 was smaller than that caused by path 5.

4.2 Spatial Distribution of the Maximum Surge Level

The spatial distribution of the storm surge in the SYS is still subject to intense debate.As argued by Yuet al.(2013),the spatial distribution of storm surges in the radial sand ridge area is relevant to coastal segments during typhoons active in offshore areas,and the potential maximum surge level occurs near Jianggang.According to Zhenget al.(2017b),three types of typhoons induce the highest surge levels in the radial sand ridge area,and the surge levels decrease from south to north.Xieet al.(2018) claimed that typhoons making straight landfall and typhoons active in offshore areas brought relatively high surge levels in the offshore area in Jiangsu Province,especially in Jianggang and Liyashan.Another study concluded that typhoons making straight landfall caused the surge level to rise southward and then decrease (Wanget al.,2015).In conclusion,these differences in the spatial distribution of surge level can be explained partially by different types and tracks of typhoons striking Jiangsu Province.

In Sections 3.3 and 3.4,numerical experiments of hypothetical typhoons were conducted to analyze the spatial distribution of the maximum surge levels in the Jiangsu coastal region.Fig.11 displays that the extreme values of storm surges in coastal areas changed with the variations of typhoon tracks,and the spatial distribution of the probable maximum surge levels in coastal areas also varies.This result may support the hypothesis that for typhoons active in offshore areas,the variations of typhoon tracks lead to differences in the spatial distribution of the maximum surge levels along the Jiangsu coast.

For all the numerical experiments of typhoons,the surge level in the radial sand ridge area reached a maximum of 2.91 m,and that in the middle straight coastline area and Haizhou Bay area reached up to a peak of 2.08 and 2.37 m,respectively.Consequently,for typhoons active in offshore areas,the radial sand ridge area was most likely threatened by massive storm surges.

One of the most noticeable phenomena was the extreme value of surge level induced by the typhoon along path 2,which occurred in the tidal flat of the Rudong coast,i.e.,the southern part of the radial sand ridge area.Zhao and Gao (2015) indicated that the tidal range is directly proportional to the water level change rate.A semiclosed siltmud tidal flat was formed at the Rudong coast due to the masking effect of shore-attached tidal ridges.The Rudong coastal area belongs to the regular semi-diurnal tide,and the maximum tidal range reaches 8 m.Shore-attached tidal ridges and tidal flats intermittently rise above sea level near the Rudong coast.When the tidal waves entered the Rudong coast,the shallow water tide proportion gradually increased due to the shallow water depth,and the nonlinear effect of tidal waves was remarkable;thus,the storm surge in the Rudong coastal area was severe.This finding provided specific reference suggestions for the prevention of coastal disasters in the offshore area of Jiangsu Province.

5 Conclusions

To reproduce the evolution of storm surges and waves,a coupled tide-surge-wave model in the SYS was developed by using ECMWF reanalysis data,the Holland parameter model,MIKE 21 HD,and MIKE 21 SW.This model simulated storm surges caused by extreme weather events,which could be categorized into four classes:a tropical cyclone,a cold wave,an extratropical cyclone coupled with a cold wave,and a tropical cyclone coupled with a cold wave.The modeled tides were validated with tidal level records at the Liyashan Station,and the simulated significant wave heights were verified with the GDR products from Jason-2 satellites.Model results of the five extreme weather events showed that the coupled model could capture the storm surge process.Compared with the uncoupled model results,the coupled model could increase the simulation accuracy of storm surges by up to 3.8 cm.

The numerical experiment results of typhoons with different intensities showed that the wind speed of Typhoon A2 was 1.2 times that of Typhoon Chan-hom;thus,the maximum surge level caused by Typhoon A2 was larger than that of Typhoon Chan-hom.However,when the wind speed increased to 1.5 times,the radius to the maximum wind speed decreased to the extent that the impact on the surge levels in the Jiangsu coastal region was weakened.

On the basis of Typhoon Chan-hom,eight hypothetical typhoons were designed to investigate the variations of the potential maximum surge level in the Jiangsu coastal region.Given the characteristic of path 2,the typhoon along path 2 had a major effect on the wind speed of the Rudong coast,and the tidal flats and sand ridges on the Rudong coast were advantageous to the generation of storm surges.The typhoon along path 2 imposed the greatest threat to Jiangsu Province,and the probable extreme value of surge level reached up to 2.91 m.

In addition,for typhoons active in offshore areas,the variations of typhoon track led to changes in the spatial distribution of the maximum surge level in the Jiangsu coastal area.The extreme value of surge levels induced by designed typhoons reached 2.91,2.08,and 2.37 m in the radial sand ridge area,the middle straight coastline area,and the Haizhou Bay area,respectively.In terms of typhoons active in offshore areas,the radial sand ridge area was more susceptible to the threat of storm surges compared with the Haizhou Bay area and the middle straight coastline area.This finding provides a valuable reference for the prevention and protection of storm surges along the Jiangsu coast.

Acknowledgements

This work was funded by the Fundamental Research Funds for the Central Universities (No.B210202031),the National Natural Science Foundation of China (No.4160 6042),and the Marine Renewable Energy Foundation,State Oceanic Administration,China (No.GHME2017YY01).