Ocean Engineering 30 (2003) 739759

www.elsevier.com/locate/oceaneng

Modeling of coupled tidewavesurge process

in the Yellow Sea
B.H. Choi a,, H.M. Eum a, S.B. Woo b
a

Department of Civil and Environmental Engineering, Sungkyunkwan University, Suwon 440-746,

South Korea
b
School of Civil and Environmental Engineering, Cornell University, Ithaca, NY 14853, USA
Received 12 February 2002; accepted 18 April 2002

Abstract
A coupled wavetidesurge model has been developed in this study in order to investigate
the effect of the interactions among tides, storm surges, and wind waves. The coupled model
is based on the synchronous dynamic coupling of a third-generation wave model, WAM cycle
4, and the two-dimensional tidesurge model. The surface stress, which is generated by interactions between wind and wave, is calculated by using the WAM model directly based on an
analytical approximation of the results using the quasi-linear theory of wave generation. The
changes in bottom friction are created by the interactions between waves and currents and
calculated by using simplified bottom boundary layer model. In consequence, the combined
wavecurrent-induced bottom velocity and effective bottom drag coefficient were increased in
the shallow waters during the strong storm conditions.
2002 Elsevier Science Ltd. All rights reserved.
Keywords: Synchronously coupled model; Surge; Tide; Wave; Typhoon

1. Introduction
The Yellow and East China Seas form a complex oceanographic system with
considerable spatial and temporal variations in physical properties and behavior (Fig.

Corresponding author. Tel.: +82-31-290-7512; fax: +82-31-290-7546.

E-mail address: bhchoi@yurim.skku.ac.kr (B.H. Choi).

0029-8018/02/$ - see front matter 2002 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0029-8018(02)00064-1

740

B.H. Choi et al. / Ocean Engineering 30 (2003) 739759

1). The region is most remarkable for large tides along the west coast of Korea,
mid-east Chinese coast and for complexity of tidal phenomena along with seasonally
varying monsoon, typhoons, fresh water discharges from surrounding land masses
and the Kuroshio. During the severe storm conditions the interactions among tide,
wave, and storm surge processes are pronounced within the shallow sea basin. Previous studies (Zhang and Li, 1996; Li and Zhang, 1997; Moon, 2000; Ozer et al.,
2000) on the synchronous coupling of the wave model and the storm surge model
had not explored the wave-induced enhancement of bottom stress generated by the
tidal- and surge-induced currents.
In this study the dynamically coupled wavetidesurge model is developed based
on the theory of wave-dependent surface roughness for upper surface condition
(Janssen, 1991) and simplified bottom boundary layer model of Grant and Madsen
(1986) for bottom boundary condition. And then the developed model was used to
simulate a typhoon surge, which devastated Korean and Japanese coasts extensively
in 1959. Some of the results are presented and discussed.
2. Numerical model
2.1. Wave model
The third-generation wave model, WAM-cycle 4, is one of the most sophisticated
and validated models in the world (WAMDI Group, 1988) and integrates the basic

Fig. 1.

Model area for the study.

B.H. Choi et al. / Ocean Engineering 30 (2003) 739759

741

transport equation describing the evolution of a two-dimensional ocean wave spectrum without additional assumptions with respect to the spectral shape. The three
source functions describing the wind input, nonlinear transfer, and white-capping
dissipation are prescribed explicitly. In addition to the source functions, bottom dissipation and refraction terms are incorporated into the model of the finite-depth version. The evolution of a two-dimensional ocean wave spectrum F(f,q,f,l,t) with
respect to frequency f and direction q (measured clockwise relative to true north) as
a function of latitude f and longitude l on the spherical earth is governed by the
transport equation

F
(cosf)1 (f cosfF) (l F) (q F) S
t
f
l
q

(1)

where S is the net source function describing the change of energy of a propagating
wave group and
df
vR1cosq
f
dt

(2)

dl
vsinq(Rcosf)1
l
dt

(3)

dq
vsinqtanfR1
q
dt

(4)

represent the rates of change of the position and propagation direction of a wave
packet traveling along a great circular path. Here v g / 4f denotes the group velocity, g the acceleration of gravity, and R the radius of the earth. The right-hand side
of Eq. (1) represents the effects of wave generation and dissipation by wind, nonlinear resonant wave interactions, white capping, and wave energy dissipation due to
bottom friction. The linear theory of surface gravity waves with wave number (k)
and frequency (w) are interrelated in the dispersion relation as follows:

w s k U

(5)

s2 gk tanh kd

(6)

where w( 2fa) is the absolute radian frequency observed in a frame of reference

fixed to the bottom and s( 2fr)is the relative frequency observed in a frame of

reference that moves with the mean current velocity (U); fa and fr are the absolute

and relative frequencies, respectively, k the wave number vector with absolute value
of k, and g is the acceleration of gravity.
The model was calibrated with respect to the fetch-limited wave growth data. Only
two tuning parameters are introduced in the white-capping dissipation source function. The model is well-explained by Gu nther et al. (1992) and its scientific, basic,
and actual implementation are also described by Komen et al. (1994). Despite the
insufficient physical understanding of the processes with respect to the evolution and

742

B.H. Choi et al. / Ocean Engineering 30 (2003) 739759

dissipation, the results from the model prediction are very satisfactory in practical
point of view when some empirical relations of the processes are skillfully incorporated into the model.
2.2. Tide and surge model
The vertically integrated equations in spherical-coordinates for tides and storm
surges incorporate a quadratic law of bottom friction with the nonlinear advective
terms. The equations are given as

1 (dU) (dVcosf)
x
0

t
Rcosf x
f

(7)

U U
V
g x
1 P
U
2wsinfV

(Ucosf)

t
Rcosf x
Rcosff
Rcosx rRcosf x

(8)

kU(U2 V2)1/2 F(s)

d
rd
U U VV gx 1 P kV(U2 V2)1/2
V
2wsinfU

t
Rcosf x
R f
R f rR f
d

(9)

G(s)
rdL

where x and f are the east-longitude and north-latitude, respectively, t the time, x
the elevation of sea surface, h the undisturbed depth of water, d h x i.e. the
total depth of water, R the radius of the Earth, w the angular speed of the Earths
rotation, g the acceleration due to gravity, F(s),G(s) the components of wind stress
on sea surface to the east and the north, respectively, and P is the atmospheric
pressure at the sea surface. U and V are the components of the depth-mean current
given by

1
1
u(z)dz,V
v(z)dz
U
hx
hx
x

u(z) and v(z) are the components of current in the direction of increasing x and f,
respectively, at a depth of z below the undisturbed sea surface, and k is the coefficient
of quadratic bottom friction and 0.0025 for this study.
Eqs. (7)(9) are discretized in time and space with the use of a simple one-step
forward time difference, and a staggered spatial grid in which x, U, and V are calculated at different grid points. This grid scheme has been used previously in discretizing the hydro-dynamics equation (Choi, 1980). The nonlinear advective terms are
incorporated using the angled derivative method suggested by Robert and Weiss
(1967). The method centers the advective term in time domain, effectively damping
the high frequency waves generated by the nonlinearities. Thus, the difference

B.H. Choi et al. / Ocean Engineering 30 (2003) 739759

743

scheme is explicit and employs forward and backward differences for time and central difference for space. In order to solve Eqs. (7)(9), it is necessary to specify
both initial and boundary conditions. Along the coastline boundaries, the current
component normal to the boundary is set to zero and a radiation condition for the
open boundaries is used to prevent the artificial reflection of the disturbances generated in the model. The open boundary condition is specified in terms of a function
of position and time as follows:
c
T
M
T
qn qM
n qn (xx x )
h

(10)

M
the inputs of meteorological origin, and qTn and xT are
where c (gh)1 / 2, qM
n and x
the inputs of tidal origin. The tidal inputs of qTn and xT are expressed in standard
harmonic form, as follows:

xT

fiHicos(sit ui ViGi)

(11)

fiQicos(sit ui ViGi)

(12)

i1
n

qTn

i1

where Hi, Qi, and Gi for normal component of depth-mean currents of the ith constituent denote the harmonic constants, amplitudes, and phases of tidal elevation,
respectively, si the speed, n the number of constituent, Vi the phase of corresponding
equilibrium constituent in Greenwich at time t 0 (the start of a hindcast), and fi
and ui are the nodal factors allowing for the 18.6 year variation in amplitude and
phase of the constituents. The values of Vi, fi, and ui are computed at the start of
each hindcast suggested by Doodson (1928). The constants for the tide model are
given by the previous model experiment (Choi, 1980) and only M2, S2, K1, and O1
M
constituents are used. The surge input of qM
n and xn is assumed to be
M
(p apa) / rg
qM
n 0,x

(13)

where p a is 1013 mbar and pa is the atmospheric pressure at the sea surface.
With the numerical models covering the Northwest Pacific region, the water movements associated with real time tides and actual storms can be reproduced. Reasonably well-representative model simulations with the observed data in coastal areas
can lead to estimate the extreme condition of waves and currents for the design of
coastal defense and offshore structures. The model employed here was developed
originally at the Institute of Oceanographic Sciences (IOS) (Choi, 1980) to explore
general tidal dynamics of the Yellow Sea. This model was improved and extended
to cover the broad region from previous model (Choi, 1980) and to include additional
major tidal components (Choi, 1987, 1993). The model is two-dimensional
implementing nonlinear depth-averaged hydro-dynamical equations in spherical polar
coordinate. The model has the east and north components of the current on the grid
system of 1/12 latitude by 1/12 longitude utilizing Digital Bathymetric Data Base
5 min (DBDB5) with modified water depths in the Yellow Sea area. The model has

744

B.H. Choi et al. / Ocean Engineering 30 (2003) 739759

the 481 361 grid elements with the selected time step of 40 s. Open boundary
conditions of the tidal model for East Asian Marginal Seas have been selected for
this model with the model resolution of 1/6 (Choi and Ko, 1994), from which the
computed M2, S2, K1, O1, N2, K2, P1, and Q1 tidal charts are provided in Fig. 2.
Along the open boundary, eight major tidal constituents can be prescribed for real
time tide specification.
2.3. Surface wind stress and simplified bottom boundary layer model
Janssen (1991) concluded that the growth rate of the waves generated by wind
depends on a number of additional factors, such as the atmospheric density stratification, wind gustiness and wave age and gave a more realistic parameterization of
the interaction between wind and wave. Heaps (1983) had already identified the
need for a wave model to improve the specification of wind stress in surge models.
Mastenbroek et al. (1993) clearly showed the influence of a wave-dependent surface
drag coefficient on surge elevations. Even if these surge elevations can be reproduced
with an appropriate tuning of this parameter in conventional wind stress formulations
(e.g. the dimensionless constant a in the Charnock relation (Charnock, 1955), they
argue that a wave-dependent drag is to be preferred for storm surge modeling. Davies
and Lawrence (1994) noticed a significant change of the tidal amplitude and phase
in shallow near-coastal regions due to enhanced frictional effects associated with
wind-driven flow and wind wave turbulence.
The total surface stress consists of a turbulent and wave-dependent term as follows:
t tt tw

(14)

where tt is the turbulent stress parameterized with a mixing-length hypothesis and

it is explained as follows:
tt ra(z)2

U
z

(15)

where ra is the air density, the von Karman constant () is equal to 0.4, and U(z)
is the wind speed at height z.
tw is the wave-dependent stress, which is very large at the initial stages of wave
growth when young wind-sea prevails in the wave field. This occurs immediately
after an increase of the wind speed or change of the wind direction. Janssen (1991)
derived the wave-induced stress using the following wind profile in the presence
of waves
U (z)

z zez0
u
ln
k
ze

(16)

where u is the friction velocity, and z0 and ze are the roughness length in the absence
of waves and the effective roughness length (a function of the wave-induced
stress), respectively.

B.H. Choi et al. / Ocean Engineering 30 (2003) 739759

Fig. 2.

Computed co-tidal charts of the M2, S2, K1, O1, N2, K2, P1, and Q1 tides.

745

746

B.H. Choi et al. / Ocean Engineering 30 (2003) 739759

For the turbulent term of the stress the mixing length hypothesis is assumed. If
the wind profile described in Eq. (16) is differentiated, squared, and compared with
the expression of the turbulent stress described in Eq. (15) at z z0, the relationship
between z0 and ze is explained as
ze

z0
((1tw(z0)) / t)1/2

(17)

where Eq. (14) and the friction velocity u (t / ra)1 / 2 are used.
Since drag coefficient (CD) depends on the roughness length, t is written as
t CDU2(L)
where
CD

k
log(L / z0)

(18)

(19)

where L is the mean height above the wave. The wave-induced stress (tw) is given by

2p

t w rw

sSin(fr,q)dfdq

(20)

0 0

where s is the angular frequency, f the frequency, q the direction, rw the density of
water, and Sin is the wind input source function.
The wave-dependent stress tw is computed from Sin, which is based on an analytical approximation of the results obtained from quasi-linear theory of wave generation. Given the wind speed, U(L) at L, tw and the surface roughness are determined
from an iterative solution of Eqs. (17)(20). The ratio, t / tw, depends on the wave
age and indicates the strength of the coupling between winds and waves. For extreme
young wind-sea, the surface roughness can be enhanced by as much as a factor of 10.
Given the U(L), tw and the surface roughness are determined from iterative solution using the set of equations with the algorithm provided within WAM model.
There are general discrepancies between the present approach and conventional
method (e.g. Smith and Banke, 1975) as illustrated in Fig. 3.
From the wave and bottom roughness conditions considered in this study, it was
determined that the effect of the steady current on the oscillatory stress component
could be neglected because it only influences tw and not tc. The steady stress component (tc), on the other hand, is a strong function representing both waves and currents.
Therefore, the effective drag coefficient can be determined from an iterative process
followed by the method of Grant and Madsen (1986).
The procedure is as follows.
1. Assume CnR(stressenhancementfactor) 1.0. Determine t fru u with the friction factor (f 0.0025) and the air density, (r 1025kg / m3), where u is depthmean velocity. Current shear velocity (uc (t / r)1 / 2) and current velocity
(uc (uc / )ln(z / z0)) with the bottom roughness (z0 0.1cm) and reference

B.H. Choi et al. / Ocean Engineering 30 (2003) 739759

747

Fig. 3. Comparison surface wind stress using the WAM model (left) and Smith and Banke (1975) (right)
during winter monsoon in November 1983.

level (z 100cm) were selected for the simplicity. This covers biological as well
as most wave formed ripple cases.
2. Find a value of fw (combined wavecurrent friction factor) corresponding to CnR
by solving Eq. (21)

1
CnRub
log
log
1.65 0.24(4fw)
wz0
4fw
4fw
1

(21)

where radian frequency (w 2 / TS). TS is the wave period and ub is waveinduced orbital velocity.
3. Calculate uwm (enhanced wave shear velocity) from the following equation:
uwm twm / r CnRfw / 2ub
4. Calculate uc (current shear velocity) from the following equation:

(22)

748

B.H. Choi et al. / Ocean Engineering 30 (2003) 739759

uc

uc uc dcw
z
ln
ln
ucw z0
dcw

(23)

where ucw uwm (CnR)1 / 2, dcw 2l 2(ucw / w), and l is the scale of the
wave boundary layer thickness.
5. Calculate new CnR 1 from the following equation:
ucw uwm[1 2(uc / uwm)2cosf (uc / uwm)4]1/4 Cn+1
R uwm

(24)

where f denotes the angle between waves and current and is taken as collinear.
6. If (CnR 1CnR) / CnR 1 0.01, go to step 2. Otherwise, calculate the effective drag
coefficient, CD.
2.4. Simulation procedures
For the tide and surge simulation, the vertically integrated equations for the tides
and storm surges in spherical-coordinates were used (Choi, 1986) while WAM model
(WAMDI Group, 1988) was used for wave computations. The synchronous coupling
scheme of the two models to explore the mutual interactions of wavetidesurge
processes is shown in Fig. 4. Resolution of the two models is identical as 1/12 in
longitude and latitude covering inner NW Pacific including the East China Sea continental shelf and the East (Japan) Sea.
The following four steps are performed for synchronous simulation:

Fig. 4.

Schematic of synchronous coupling between WAM model and tidesurge model.

B.H. Choi et al. / Ocean Engineering 30 (2003) 739759

749

1. The SETINT sub-program of the coupled model generates the initial input data
of the wave model including surface wind forcing. Before the wave model is
coupled with the tidesurge model, the tidesurge model runs to confirm that tide
with the duration of three days propagates to the entire modeled region.
2. GWSBC sub-program is executed to obtain the mean value of the water surface
elevation at the center and currents at the edge center of the each grid from the
tidesurge model. And then these mean values obtained from the tidesurge model
are used for the input data of the wave model. Similarly, The GTSBC sub-program
is executed to obtain the mean value of the wave information (wave height, period,
and total surface stress) at four corners of the grid from the wave model. Similarly,
these mean values are used for the input data of the tidesurge model.
3. The wave model and the tidesurge model are performed with the constant time
step. Time steps of the tidesurge and wave model are 10 and 120 s, respectively.
To exchange the information of these two models, time step of 120 s can be used.
However, time step of 3600 s is selected for this study after making a series of
numerical tests.
4. In the coupled run, surface forcing for tide and surge model is prescribed by total
surface stress considering wave-dependent drag, which is computed by WAM and
atmospheric pressure gradient.

The steps from 2 to 4 are repeated to obtain the wave information and currents
at each time step.

3. Meteorological forcing input

For the simulation, Grid Point Values (GPVs) of sea level pressure and air and
sea surface temperature with 50 km intervals over the Northeast Asian seas were
digitized by Japan Weather Association (JWA). Those values were interpolated to
dense 1/12 grid resolution at one hourly intervals from six hourly dataset for the
coupled model. The overall marine wind fields are computed by adopting Planetary
Marine Boundary Layer model (Cardone, 1969) and then the typhoon model winds
by Rankin vortex model (Fujita, 1952) were inserted. Temporal interpolation along
the typhoon track was also performed. Fig. 5 shows the computed results of pressure
and wind vector fields for the typhoon Sarah (5914).
The track of the typhoon Sarah (5914) selected for the simulation is shown in
Fig. 6. Typhoon Sarah (5914) formed in September off the coast of Saipan and
tracked northwestward with a pressure of 905 hPa and winds up to 65 m/s over
northeastern Taiwan at a speed of about 25 km/h. It then moved north and northeastward with a pressure of 945 hPa passing south of Korean Peninsula. It resulted in
a death and missing toll of 849 with 2533 people injured.

750

B.H. Choi et al. / Ocean Engineering 30 (2003) 739759

Fig. 5.

Distribution of wind and pressure fields for the typhoon Sarah (5914).

Fig. 6.

Track of the typhoon Sarah (5914).

B.H. Choi et al. / Ocean Engineering 30 (2003) 739759

751

4. Model results
To investigate the coupling effects of tides, storm surges, and wind waves, the
simulation has been performed with two additional numerical experiments along with
synchronous coupling model runs. First, the independent wave model run was performed with only meteorological wind fields. Secondly, the wave model run was performed with meteorological wind fields and hourly data of the sea surface elevations
and depth-mean current velocity fields computed from tide and surge model in
loosely coupled manner.
The time series of the significant wave height (Hs), mean period (Ts), sea surface
elevation (tide surge), depth-mean current (u ), current shear velocity (uc), current
direction, near-bottom wave orbital velocity (ub), combined wavecurrent velocity
(ucw), and effective bottom drag coefficient (CD) from the synchronously coupled
model showing variations at off port of Pusan (PS) during typhoon Sarah as shown
in Fig. 7. This figure shows that the significant wave height exceeds about 7 m and
immediately increases to over 9 m and the mean wave period remains relatively
constant with 15 s for two-day long period of high winds. Similar to Fig. 7, Fig. 8
shows the wave and current parameters at Changjiang (CJ) river. The wave heights
grow up to 4.5 m with wave periods of 10 s and surge elevation is in the range
from 1.5 to 1.5 m. And the effective drag coefficient reaches up to about 0.03 or
more. The wave height at Mokpo (MK) does not exceed to 4.5 m and the wave

Fig. 7. Variation of wave and current parameters at off port of Pusan (PS) (depth 10.4m) during
typhoon Sarah (5914).

752

B.H. Choi et al. / Ocean Engineering 30 (2003) 739759

Fig. 8. Variation of wave and current parameters at Changjiang (CJ) (depth 4.5m) during typhoon
Sarah (5914).

period remains roughly in the range from 4.5 to 10 s as shown in Fig. 9. The nearbottom orbital velocity and the combined wavecurrent velocity reach 2.0 and 0.2
m/s, respectively.
Figs. 10 and 11 show the spatial distribution of the maximum significant wave
heights and mean period generated by the typhoon Sarah. The largest significant
wave heights of 79, 5 m or more, and 711 m occur along the track of typhoon
around the south coasts of Korea, the CJ river in China, and the west coasts of Japan,
respectively. The mean periods are 69, 13, and 13 s around the west coasts of
Korea, the Winzou in China, and the southwestern coasts of Kyushu in Japan. From
the change of the significant wave height between synchronous coupled model run
and independent wave model run brings the increment of 0.51.0 m around the CJ
river in China and 0.5 m around the southwestern coasts in Korea as shown in Fig.
12. However, it decreases up to 0.51.0 m along the track of the typhoon. Fig. 13
shows the change of the significant wave height between synchronous coupled model
run and loosely coupled model run. The change of the significant wave height has
a similar trend along the track of typhoon as shown in Fig. 12. The next figure shows
the change of the significant wave height between loosely coupled model run and
independent wave model (Fig. 14). This figure illustrates that the significant wave
height increases up to 1.01.5 and 0.51.0 m around the CJ river in China and the
southwestern coasts in Korea, respectively.

B.H. Choi et al. / Ocean Engineering 30 (2003) 739759

753

Fig. 9. Variation of wave and current parameters at Mokpo (MK) (depth 20.6m) during typhoon
Sarah (5914).

Fig. 10.

Distribution of maximum significant wave height (m) during typhoon Sarah in September 1959.

754

B.H. Choi et al. / Ocean Engineering 30 (2003) 739759

Fig. 11.

Distribution of maximum mean period (s) during typhoon Sarah in September 1959.

Fig. 12. Distribution of changes of significant wave height (m) between synchronous coupled model run
and loosely coupled model run during typhoon Sarah (5914).

The distributions of the computed maximum positive and negative surge elevation
and surge current for typhoon Sarah (5914) are shown in Figs. 15. As shown in
Fig. 15, the sea surface elevation increases up to 12.515.0, 0.70.8, and 0.9 m or
more around the southern coasts of Shandong in China, the Hangzhou Bay in Chinese
coast, and the western coasts in Korea, respectively. The computed maximum negative surge elevations are 0.5 to 1.25, 0.4 to 0.6, and 0.4 m or more for
the western and southern coasts in Korea, the Hangzhou Bay, and the Fujian coasts
in China, respectively, as shown in Fig. 16. Fig. 17 shows that the computed

B.H. Choi et al. / Ocean Engineering 30 (2003) 739759

755

Fig. 13. Distribution of changes of significant wave height (m) between synchronous coupled model run
and independent wave model run during typhoon Sarah (5914).

Fig. 14. Distribution of changes of significant wave height (m) between loosely coupled model run and
independent wave model run during typhoon Sarah (5914).

maximum surge currents of 0.5 m/s or exceeding this value and 0.6 m/s occur around
the southwestern part of Korea and the CJ river of China, respectively. The change
of the positive surge elevation between synchronous coupled model run and tide
surge model run is shown in Fig. 18. As shown in the figure, the sea surface elevation
of the entire modeled region increases. The elevations of 0.6 and 0.10.2 m are
increased around the CJ river in China and the southern coasts in Korea, respectively.
And the sea level has reduced to 0.1 m or less around the Fujian coasts in China. Fig.
19 shows the change of the maximum negative surge elevation between synchronous

756

B.H. Choi et al. / Ocean Engineering 30 (2003) 739759

Fig. 15.

Distribution of maximum positive surge height (cm) for typhoon Sarah (5914).

Fig. 16.

Distribution of maximum negative surge height (cm) for typhoon Sarah (5914).

coupled model run and tidesurge model run. The sea surface elevation decreased
to 0.4 m or more around the western side of the Yellow Sea. The maximum surge
current decreased up to 0.1 to 0.4 and 0.2 m/s around the East China Sea and
increased around the northern part of Yellow Sea, respectively, as shown in Fig. 20.

5. Conclusion
A synchronous coupled wavetidesurge model has been developed in this study
by considering the effect of surface waves entering the tide and surge simulation

B.H. Choi et al. / Ocean Engineering 30 (2003) 739759

Fig. 17.

757

Distribution of maximum surge current (cm/s) for typhoon Sarah (5914).

Fig. 18. Distribution of difference of maximum positive surge height (cm) between synchronous coupled
model run and tidesurge model run for typhoon Sarah (5914).

through an effective bottom drag coefficient calculated by using the theory of Grant
and Madsen (1986) and while the unsteadiness of surface elevation and currents due
to tide and surge enters the wave computation. In this study, a typhoon surge in
1959 was reasonably simulated based on the model developed by us. Further
improvement is being progressed to incorporate the stratified bottom conditions to
limit the local increase of effective bottom friction factor and replace the two-dimensional to three-dimensional tide and surge model.

758

B.H. Choi et al. / Ocean Engineering 30 (2003) 739759

Fig. 19. Distribution of difference of maximum negative surge height (cm) between synchronous coupled
model run and tidesurge model run for typhoon Sarah (5914).

Fig. 20. Distribution of difference of maximum surge current (cm/s) between synchronous coupled
model run and tidesurge model run for typhoon Sarah (5914).

Acknowledgements
The research was performed as a part of second phase (19982000) of Natural
Hazard Prevention Research, funded by Ministry of Science and Technology through
KISTEP(Korea Institute of Science and Technology Evaluation and Planning) and
Group for Natural Hazard Prevention Research.

B.H. Choi et al. / Ocean Engineering 30 (2003) 739759

759

References
Cardone, V.J., 1969. Specification of the wind distribution in the marine boundary layer for wave forecasting. Report GSL-TR 69-1, New York University school of Engineering and Science.
Charnock, H., 1955. Wind stress on a water surface. Quarterly Journal of the Royal Meteorological Society
81, 639640.
Choi, B.H., 1980. A tidal model of the Yellow Sea and the Eastern China Sea. Korea Ocean Research
and Development Institute, KORDI Report 80-02.
Choi, B.H., 1986. Surge hindcast in the East China Sea. Progress in Oceanography 17, 6781.
Choi, B.H., 1987. Development of tide and surge models of the Yellow Sea for coastal engineering
application. In: Proceedings of the Second Conference on Coastal and Port Engineering in Developing
Countries, Beijing, China., pp. 18801894.
Choi, B.H., 1993. Development of tidal models at Sungkyunkwan University. In: Proceedings of Workshop on Tidal and Oil Spill Modeling, Pohang Institute of Science and Technology., pp. 81101.
Choi, B.H., Ko, J.S., 1994. Modeling of tides in the East Asian marginal seas. Journal of Korean Society
of Coastal and Ocean Engineers 6 (1), 94108.
Davies, A.M., Lawrence, J., 1994. Examining the influence of wind and wave turbulence on tidal currents,
using a three-dimensional hydrodynamic model including wavecurrent interaction. Journal of Physical
Oceanography 24, 24412460.
Doodson, A.T., 1928. The analysis of the tides and surges round North and East Britain. Philosophical
Transactions of the Royal Society of London A263, 155.
Fujita, T., 1952. Pressure distribution within typhoon. Geophysics Magazine 23, 437451.
Grant, W.D., Madsen, O.S., 1986. The continental shelf bottom boundary layer. Annual Review Fluid
Mechanics 18, 265305.
Gu nther, H., Hasselmann, S., Janssen, P.A.E.M., 1992. The WAM model Cycle 4, Report No 4, Hamburg.
Heaps, N.S., 1983. Storm surges 19671982. Geophysical Journal of the Royal Astronomical Society 74,
331376.
Janssen, P.A.E.M., 1991. Quasi-linear theory of wind wave generation applied to wave forecasting. Journal
of Physical Oceanography 21, 16311642.
Komen, G.J., Cavaleri, L., Donelan, M., Hasselmann, K., Hasselmann, S., Janssen, P.A.E.M., 1994. In:
Dynamics and Modelling of Ocean Waves. Cambridge University Press, Cambridge, 532.
Li, Y.S., Zhang, M.Y., 1997. Dynamic coupling of wave and surge models by EulerianLagrangian
method. Journal of Waterway, Port, Coastal and Ocean Engineering 123 (1), 17.
Mastenbroek, C., Burgers, G., Janssen, P.A.E.M., 1993. The dynamical coupling of a wave model and a
storm surge model through the atmospheric boundary layer. Journal of Physical Oceanography 23,
18561866.
Moon, I.J., 2000. Development of a coupled ocean wave-circulation model and its applications to numerical experiments for wind waves, storm surges and ocean circulation of the Yellow and East China
Seas. PhD Thesis, Seoul National University, Korea.
Ozer, J., et al. 2000. A coupling module for tides, surges and waves. Coastal Engineering 41, 95124.
Robert, K.V., Weiss, N.O., 1967. Convective difference schemes. Mathematics of Computation 20,
272299.
Smith, S.D., Banke, E.G., 1975. Variation of the sea surface drag coefficient with wind speed. Quarterly
Journal of the Royal Meteorological Society 101, 665673.
WAMDI Group 1988. The WAM modela third generation ocean wave prediction model. Journal of
Physical Oceanography 18, 17751810.
Zhang, M.Y., Li, Y.S., 1996. The synchronous coupling of a third-generation wave model and a twodimensional storm surge model. Ocean Engineering 23 (6), 533543.