See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/228708998

Application of numerical models to evaluate oil spills propagation in the

coastal environment of the Black Sea

Article  in  Journal of Environmental Engineering and Landscape Management · December 2010

DOI: 10.3846/jeelm.2010.33

CITATIONS READS

54 421

1 author:

Liliana Rusu
Universitatea Dunarea de Jos Galati
200 PUBLICATIONS   2,276 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

ROmanian MArine Renewable solutions (ROMAR) View project

CLIMEWAR - CLimate change IMpact Evaluation on future WAve conditions at Regional scale for the Black and Mediterranean seas marine system View project

All content following this page was uploaded by Liliana Rusu on 30 May 2014.

The user has requested enhancement of the downloaded file.

 JOURNAL OF ENVIRONMENTAL ENGINEERING
AND LANDSCAPE MANAGEMENT
2010
18(4): 288–295

APPLICATION OF NUMERICAL MODELS TO EVALUATE OIL SPILLS

PROPAGATION IN THE COASTAL ENVIRONMENT OF THE BLACK SEA

Liliana Rusu
Faculty of Mechanical Engineering, University Dunarea de Jos of Galati,
Domneasca 111, 800 008 Galati, Romania
E-mail: [email protected]
Submitted 28 May 2009; accepted 20 Oct. 2009

Abstract. The last decades continuously increasing of the economical activities in the coastal environment of the Black
Sea is obviously leading to the enhancement of the pollution risks due to accidental oil spillages. Starting from the fact
that most accidents were generated by an inadequate forecast of the wave conditions, the aim of the present work is to de-
velop a methodology based on spectral phase–averaging wave models able to predict the wave propagation in the coastal
environment. The wave induced currents that may be a key factor in driving the pollution are also assessed. This implies
both the Stokes drift and the wave induced nearshore currents. The surface streaming effect due to the molecular viscosity
was also accounted for. In the nearshore, close to the surf zone, the pollution is usually spread along the coast due to the
longshore currents. In this connection, the results of a simple but effective model system called ISSM are also presented.
As an alternative simulations with the SHORECIRC model system are also performed. Finally, as a case study, the propa-
gation of the pollution towards the Romanian coast generated by a hypothetic accident at the Gloria drilling platform was
assessed.
Keywords: environmental alert, oil spill, waves, wave induced currents, numerical models, Black Sea.

1. Introduction enhanced by the fact that the wave climate is often char-
acterized by sudden storms with relatively high intensity.
The economical development is in a great percentage The most recent accident in the Black Sea was in
influenced by the proximity of the sea. Nevertheless this November 2007 when the sea faced its most serious eco-
increasing of the economical activities in the nearshore logical threat for years after a fierce storm sank five
implies also a higher risk in accidents with negative con- ships, including an oil tanker and bulk carriers laden with
sequences on the coastal environment especially due to sulphur. Fuel barges were also washed ashore by the
the oil spills discharged accidentally. An oil spill is the heavy seas and more than 20 sailors were swept from
release of a liquid petroleum hydrocarbon into the envi- their vessels.
ronment due to human activity, and is a form of pollution. A better knowledge of the wave conditions is deci-
This term refers here to marine oil spills, where oil is sively contributing to the decrease of the accident risks in
released into the sea or coastal waters. the sea. On the other hand, in the case of accidents it is
According to the existent statistics the north and essential to be able to assess operationally the oil spills
west Iberian coasts of the Atlantic Ocean are included propagation. Taking into account the above considera-
between the nearshores with the highest risks in the world tions the present work is aiming towards the implementa-
in major accidents having as a result oil spillages. The tion of a system based on spectral wave models that will
accident of the oil carrier Prestige offshore of Vigo, Spain allow the wave prediction close to the Romanian Black
in November 2002, have had a huge ecological and eco- Sea coast and would help in case of environmental alerts
nomical impact, but this was only the last of a long series in assessing the oil spills propagation and coastal impact.
of accidents and incidents that happened in that area in
the last two decades. 2. The wave model
The existent experience shows that the risk of acci-
dents is seriously increased by an inappropriate Numerical models are used nowadays more and more to
knowledge of the evolution of the environmental condi- provide forecast products for the oceanographic data.
tions (especially wave and wind) in the areas with high This usually requires first large scale wave models as
maritime traffic. WAM (WAMDI Group 1988) or WW3 (Tolman 1999)
The Black Sea became in the last years an important for the generation area. They are third generation wave
corridor for the energy transfer from east to west. This is models based on the integration of the spectral action
increasing the nearshore and offshore activities in the balance equation in all the fifth dimensions (time, geo-
area, but induces also a higher risk as regards the acci- graphical space, directional and frequency spaces). In this
dental oil spillages. The risks in the Black Sea are usually equation the effects of wave generation, dissipation and

288 Journal of Environmental Engineering and Landscape Management ISSN 1648–6897 print / ISSN 1822-4199 online
http:/www.jeelm.vgtu.lt/en doi:10.3846/jeelm.2010.33
Journal of Environmental Engineering and Landscape Management, 2010, 18(4): 288–295 289

nonlinear wave-wave interactions are introduced by the action in geographical space (with propagation velocities
source terms. For the transformation scale, where the cx and cy in x – and y – space, respectively). The fourth
physical processes are more complexes, based on the term represents shifting of the relative frequency due to
same action balance equation, were designed some other variations in depths and currents (with propagation veloc-
phase averaged models like SWAN (Booij et al. 1999) or ity c in  – space). The fifth term represents depth-
STWAVE (Smith et al. 2001). Their physics is more induced and current-induced refraction (with propagation
elaborated to account better for the processes specific to velocity c in  – space). The expressions for these prop-
the coastal environment as refraction, shoaling, breaking agation speeds are taken from linear wave theory. At the
or triad wave-wave interactions. However, in relationship right hand side of the action balance equation is the
with SWAN has to be underlined that in the last versions source term S  S (, ) , expressed as a function of ener-
it goes far beyond the condition of a model only for trans-
gy density representing the effects of generation, dissipa-
formation scales its range being extended for almost oce-
tion and nonlinear wave-wave interactions.
anic scales. This allows the possibility of implementing The wave growth due to wind is described as the
in the Black Sea of a system for wave prediction entirely sum of the linear and the exponential growth term of a
based on the SWAN (acronym from Simulation of the
wave component:
WAves Nearshore) model by performing simulations at
various scales. This would imply first the generation area Sin ,   A  BE, , (2)
that includes the entire basin of the Black Sea and will in which A describes the linear growth and BE the expo-
provide the boundary conditions for the transformation nential growth. Two optional expressions for the coeffi-
area that corresponds to the Romanian nearshore. Higher cient B are used in the model. The first is taken from an
resolution domains can be subsequently nested inside this early version of the WAM model, known as WAM Cycle 3.
area. The second expression is due to Janssen (1991) and it is
In SWAN the following wave propagation processes based on the quasi-linear wind-wave theory. A brief sum-
are implemented: propagation through geographic space, mary of the formulations that are used for the various
refraction due to bottom and current variations, shoaling source terms in SWAN is given next.
due to bottom and current variations, blocking and reflec- The dissipation term of wave energy is represented by
tions by opposing currents and transmission through or the summation of three different contributions: whitecap-
blockage by sub-grid obstacles. On the other hand, the ping Sds, w (, ) , bottom friction Sds,b (, ) and depth-
effects of wave generation and dissipation accounted are:
redistribution of wave energy over the spectrum by non- induced breaking Sds,br (, ) . Whitecapping is primarily
linear wave-wave interactions (quadruplets and triads), controlled by the steepness of the waves. In presently oper-
generation by wind, dissipation by whitecapping, dissipa- ating generation wave models, the whitecapping formula-
tion by depth-induced wave breaking and dissipation by tions are based on the pulse-based model of Hasselmann
bottom friction. (1974), as adapted by the WAMDI Group (1988):
The waves are described with the two-dimensional
wave action density spectrum, even when nonlinear phe- ~ k E (, ),
Sds, w (, )   (3)
~
nomena dominate (e.g., in the surf zone). The rationale k
for using the spectrum in such highly nonlinear condi- where  is a steepness dependent coefficient, k is wave
tions is that, even in such conditions it seems possible to
number and  ~ and k~ denote a mean frequency and a
predict with reasonable accuracy this spectral distribution
mean wave number, respectively. Depth-induced dissipa-
of the second order moment of the waves (although it
tion may be caused by bottom friction, by bottom motion,
may not be sufficient to fully describe the waves statisti-
by percolation or by back scattering on bottom irregulari-
cally). The spectrum that is considered here is the action
ties. For continental shelf seas with sandy bottoms, the
density spectrum N(σ, θ) since in the presence of currents
dominant mechanism appears to be bottom friction which
action density is conserved whereas energy density is not.
can generally be represented as:
The independent variables are the relative frequency 
(as observed in a frame of reference moving with the 2
Sds,b (, )  Cbottom E (, ), (4)
action propagation velocity) and the wave direction  (the g 2 sinh 2 (kd )
direction normal to the wave crest of each spectral com-
ponent). The action density is equal to the energy density in which Cbottom is a bottom friction coefficient.
divided by the relative frequency: N (σ,θ)  E(σ,θ) / σ. The expression that is used in SWAN for estimating
The evolution of the wave spectrum is described by the depth-induced wave breaking is:
the spectral action balance equation which for Cartesian D
Sds,br (, )   tot E (, ), (5)
coordinates is (e.g., Holthuijsen 2007): Etot
     S in which Etot is the total wave energy and Dtot is the rate of
N  cx N  c y N  c N  c N  . (1)
t x y    dissipation of the total energy due to wave breaking. The
The first term in the left-hand side of this equation value of Dtot depends critically on the breaking parameter
represents the local rate of change of action density in   H max / d (in which Hmax is the maximum possible
time, the second and third term represent propagation of individual wave height in the local water depth d).
290 L. Rusu. Application of numerical models to evaluate oil spills propagation in the coastal environment of the Black Sea

In deep water quadruplet wave-wave interactions gencies (Rusu et al. 2005). This means not only to predict
dominate the evolution of the spectrum. They transfer the wave climate in the area affected by the calamity, but
wave energy from the spectral peak both to lower fre- also to assess the wave induced currents that play an im-
quencies, moving the peak frequency to lower values, and portant role in driving the pollution.
to higher frequencies, where the energy is dissipated by A model for a quick estimation of the wind and
whitecapping. Hasselmann (1962) derived the transfer wave contributions to the propagation of the oil on the
rate to, and from, a spectral component arising from in- offshore sea surface was implemented and connected to
teractions with sets of three other spectral components. the SWAN model. This is based on the relationship:
The resulting source term S nl 4, takes the form of an dx d 
 0.035U wind  U Stokes , (7)
integral over the phase space of interacting quadruplets dt
called the Boltzmann integral. This source term is evalu- 
ated for infinite water depth and then adjusted for finite where: x is the vector defining the centre of mass of the
depth by an empirical scaling (Herterich and Hasselmann oil and the term 0.035Uwind accounts the wind-driven
1980; Hasselmann, S. and Hasselmann, K. 1981). currents. The direction of oil spill motion induced by the
A full computation of the Boltzmann integral ex- wind is at a non-zero angle to the direction of the wind as
pressing this type of non linear interactions is extremely a result of the Ekman effect.
time consuming and not convenient in any operational The formula proposed by Samuels et al. (1982) is
wave model. To reduce the computational demands in- considered to compute the deflection angle. According to
volved in implementing this source term in SWAN, as in this, the wind-driven current is deflected with an angle θ
WAM, a discrete interaction approximation (DIA), (Has- to the right (clockwise) from the wind direction (for the
selmann et al. 1985) is applied, in which only two quad- Northern Hemisphere).
ruplets are summed. Both of them have the same set of  U3 
frequencies:   25 exp  108 wind , (8)
 g 
1  1.25, 2  0.75, 3  4  . (6)  
In both quadruplets, the third and fourth wave vec- where Uwind is the wind speed, g is the gravitational ac-
tors coincide, while the others lie at angles 1  11.5 celeration and υ the kinematic viscosity of the seawater.
The effects of wind induced surface were modelled by the
and 2  33.6 to this common direction, for the first inclusion of a logarithmic velocity profile. In analogy
quadruplet, or at 1  11.5 and 2  33.6 for the se- with flow structure within a bottom boundary layer it was
cond. This DIA has been found to be quite successful in assumed that the surface layer of thickness zo moves at a
describing the essential features of a developing wave velocity (typically 3.5% of the wind speed) which decays
spectrum using the fact that the interactions between with the depth, Elliott (1986), according to:
closely neighbouring wave numbers reproduce the prin-
 logz zo  
cipal features of the nonlinear transfer. Four different U w z   0.035U wind 1  , (9)
 logzc zo  
numerical procedures for DIA can be activated into the 
SWAN model: semi-implicit computation of the nonline-
ar transfer with DIA per sweep, fully explicit computa- where zc is the depth at which the velocity is zero. It has
tion of the nonlinear transfer with DIA per sweep, fully been assumed that zc scales on the wavelength of the
explicit computation of the nonlinear transfer with DIA surface waves, i.e. zc   L .
per iteration and fully explicit computation of the nonlinear In terms of wave spectrum, the Stokes drift can be
transfer with DIA per iteration, but neighbouring interac- estimated with the relationship bellow:
tions are interpolated in a piecewise constant manner.  2
1 2
ke 2kz E ,  dd,
  0
The integration of the action balance equation has us  (10)
been implemented in SWAN with finite difference
1
schemes in all five dimensions (time, geographic space
and spectral space). Time is discretized with a simple where k is the wave number and z the water particle posi-
constant time step t for the simultaneous integration of tion relatively to the still water level. Nevertheless, to
the propagation and the source terms. Geographic space integrate the wave spectrum outside the wave model in all
is discretized with a rectangular grid with constant resolu- the grid points is computationally not convenient because
tions x and y in x and y-direction respectively. The of the great amount of output data requested, (the wave
spectrum in the model is discretized with a constant di- spectra files in all the points of the computational grid).
rectional resolution and a constant relative frequency From this reason the Stokes drift was estimated using the
mass transport relationship from the second order theory
resolution   (logarithmic frequency distribution).
(Lakshmi and Clayson 2000):

 H  C cosh 4z  d  / L ,
2
3. Wave-induced currents driving oil spills
u ( z)      (11)
The Prestige accident close to the Iberian coast, in No-  L  2 sinh 2 2d / L 
vember 2002, showed the importance of the wave models where C is the phase velocity defined using the standard
in providing the support for such environmental emer- definitions (CERC 1984).
Journal of Environmental Engineering and Landscape Management, 2010, 18(4): 288–295 291

Thus in the case of shallow water waves (when many scales, forced by several processes. Schematically,
d < L/20, where d is the depth and L the wavelength) the the total nearshore current u can be expressed as a super-
waves are considered non dispersive and consequently position of these interrelated components:
the phase velocity and the group velocity (Cg) are equal: u  uw  ut  ua  uo  ui , (19)
C  Cg  gd . (12) where uw is the steady current driven by breaking waves,
ut is the tidal current (this is not the case of the Black Sea
In deep water (d > L/2): where this component can be neglected), ua is the wind-
C  2Cg  gL / 2 . (13) driven current, and uo and ui are the oscillatory flows due
Finally for the intermediate water (L/2 ≤ d ≤ L/20), we to wind waves and infragravity waves.
have: Currents generated by the breaking of obliquely in-
cident wind waves generally dominate in and near the
C  gL / 2 tanh( 2d / L) and surf zone on open coasts. Strong local winds can also
1  4d / L  (14) drive significant nearshore currents. Wave- and wind-
Cg  C  1   , driven currents are important in the transport and disper-
2  sinh(4d / L) 
sal of sediment and pollutants in the nearshore. These
evaluated using the second order theory. For the deep currents also transport sediments mobilized by waves.
water waves the expression (11) becomes:
2 4. Case study: Accident scenario at the Gloria drilling
 H   4z 
u z      C  exp . (15) platform
 L   L 
In this section the propagation of the oil spill towards the
Since the goal was to estimate an average value for coast due a hypothetic accident at the Gloria drilling plat-
the Stokes drift at the surface it was considered z = 0 and form will be evaluated using the above presented meth-
the root mean square wave height (Hrms) was used ins- odology. This platform is operating at about 50 meters
tead of H in the relationships (11) and (15). Hrms was depth in the western part of the Black Sea at the location
deduced from the significant wave height considering the defined by the coordinates (29o34'E, 44o31'N). The envi-
standard Rayleigh distribution (Hs = 1.416 Hrms). ronmental conditions considered for this scenario were
Hence, the values found for the mass transport velo- the real conditions corresponding to the time frame
city in the case of an irrotational motion lead to a slow 2002/03/11/h18, which was a typical storm that affected
drift in direction of the wave motion (on the order of ka2, the western side of the Black Sea basin and represents the
where k is the wave number and a the wave amplitude). zero moment of the accident. The results of the SWAN
However, when the water surface is contaminated by an simulations and the model focusing towards the Romani-
inextensible slick a surface streaming relative to the fluid an coast for the conditions corresponding to the time
bellow will be induced due to molecular viscosity. Philips frame 2002/03/11/h18 are presented in Fig. 1. The NCEP
(1977) showed that in deep water the oil slick streams reanalysis wind with a 1.875º spatial resolution was used
ahead the fluid particles at the surface by a velocity for the wave model simulations.
excess of: The SWAN model implementation in the Black Sea
2 basin is described in Guedes Soares and Rusu (2005) and
3  H  2  2d  Rusu et al. (2006). For validating the results of the wave
us     C coth  . (16)
4 L   L  prediction system in the western side of the sea, data
In deep water the expression for this excess of velocity registered by a wave staff located on the Gloria drilling
due to the surface streaming becomes: unit were considered. The comparisons performed by
2 Guedes Soares and Rusu (2005) and Rusu et al. (2006)
3  H  using ‘in situ’ measured data show that the system for
us    C . (17)
4 L  wave prediction, SWAN based, that was implemented in
It means that a patch of slick move faster than the the Black Sea basin provides in general reliable results.
surrounding clean surface, the total velocity being the The model configuration was different at each level of
sum of (17) and the Lagrangian drift in deep water (15), simulations but as regards the generation area the empha-
or seven-fourths the classical result of Stokes. Conse- sis was put on the whitecapping which is still considered
quently the wave induced drift driving the pollution in the the weak link in deep water wave modelling. The models
sea that should be added to the others components of the for wave generation of Komen et al. (1984) coupled with
surface currents, should be evaluated with the relation- the pulse-based model of Hasselmann (1974) for white-
ship bellow: capping were found more effective for the specific condi-
2 tions of the Black Sea. For the higher resolution domains
7  H  physical processes accounting for the finite depths effects
U Stokes  us  us     C. (18)
4 L  (as bottom friction and triad wave-wave interactions)
When the pollution arrives close to the coast the were also introduced in the model simulations.
nearshore currents are the most important factor that The west coast is usually the most energetic part of
drives the spreading of the pollution along the coasts. the Black Sea and this makes in this area the probability
These nearshore currents are composed of motions at of accidents having as a result oil spillages considerably
292 L. Rusu. Application of numerical models to evaluate oil spills propagation in the coastal environment of the Black Sea

At this point, a brief discussion will be made in rela-

tionship with the nearshore, system focusing, in general,
and the typical storm case presented in Fig. 1, in particu-
lar. For the global and coastal areas the nonstationary
mode (NS) is in general required and a 20 minutes time
step seems to be effective, when the number of iterations
is increased to four.
However, the numerical schemes are rather differ-
ent, while the global area uses the second order scheme
S&L (Stelling and Leendertse 1992) characteristic for
large areas in non stationary mode. For the case of the
coastal areas the first order BSBT scheme (backward
space backward time) was applied.
The importance of the coastal driver (level II) is il-
lustrated in Fig. 1 where can be seen that, due to the local
effects, for the case studied, an increase of almost one
meter in terms of maximum significant wave height was
encountered when passing from the generation area to the
nearshore transformation domain. In the case of the se-
cond and third levels, that are still using spherical coordi-
nates, the connection with the level before is made using
the standard nesting procedure. When passing to the last
level, defined in Cartesian coordinates, the 2D spectra,
generated in some corresponding points by simulations at
the previous level (level III) are used as variable bounda-
ry conditions. There are two reasons for passing in the
Fig. 1. Model focusing towards the western coast, significant case of the last level from spherical to Cartesian coordi-
wave height fields and wave vectors, case study corresponding nates. One is that in SWAN model some processes as
to the time frame 2002/03/11/h18, typical storm diffraction or wave induced set up are working properly
only in Cartesian coordinates. The second is that, in this
higher. Such a high energetic case is presented in Fig. 1 way, a simpler link can be made with surf models
and refers to the storm corresponding to the time frame (SHORECIC for example) that work usually in a Carte-
2002/03/11/h18. Nevertheless, has to be mentioned that sian reference system. Moreover, the option to rotate the
this is a typical strong storm and not an extreme event. In computational grids, which is some time useful in model
Fig. 1, the significant wave height scalar fields and the simulations close to the coast, is valid only for such Car-
wave vectors are presented for various levels of resolu- tesian coordinates.
tion in a downscaling process towards the west coast. The configuration of the SWAN model is in general
As illustrated in Fig. 1, the system focalization as- rather different from one level to another, while very
sumed four different levels of spatial resolution. Level I important at the levels I and II of the system, deep water
correspond to the generation area, level II for the coastal processes are no longer relevant for the subsequent levels
transformation, level III for local focusing and level IV III and IV. Starting with level III of the system, triad non
for high resolution simulations. While for the first three linear interactions and bottom friction may become im-
levels the simulations were performed with spherical portant whereas at the last level, beside diffraction, the
coordinates the last level uses Cartesian coordinates. For depth induced breaking effect might be also relevant.
the case presented in Fig. 1 the characteristics of the cor- Fig. 2 illustrates the main factors that influence the
responding computational grids are given in Table 1. oil spill propagation for 3 different time frames.
In the upper side of this Fig. 1 the wave and wind
Table 1. Characteristics of the computational grids used in the vectors are represented having in background the scalar
case study considered for the system focusing towards significant wave height fields, while in the lower side the
the western coast Stokes drift with the bathymetric map represented in
background. For each time moment, the maximum values
Grids Δx×Δy Δt min nf nθ ngx × ngy of the significant wave height and wind velocity are indi-
cated as well as the average values of the Stokes drift.
Global 0.08º × 0.08º 20 (NS) 30 24 176 × 76
The propagation of the oil spill was estimated con-
Coastal 0.02º × 0.02º 20 (NS) 30 36 141 × 141 sidering the relationships 10, 11, 12 and 21. Consequent-
ly according to the conditions illustrated in Fig. 2 will
Local 0.005º × 0.005º 20 (NS) 30 36 161 × 141 result that the oil spill will approach the coast after ap-
Cartesian 50 m × 50 m 180 (S) 25 36 96 × 107 proximately 30 hours in the coastal environment close to
the city of Mangalia.
Journal of Environmental Engineering and Landscape Management, 2010, 18(4): 288–295 293

Fig. 2. Wave, wind and Stokes drift driving the oil spill towards the Romanian coast: a) time frame 2002/03/11/h18 (peak of the
storm considered the zero moment of the hypothetic accident); b) time frame 2002/03/12/h06; c) time frame 2002/03/13/h00
(predicted coastal impact of the oil spill)

In the nearshore the pollution is usually spread along The positions of the two lines considered are illus-
the coast due to the longshore currents. A simple, but trated in Fig. 3a while the cross shore variations of the
very effective, model to estimate these currents, is SURF longshore currents (and of some other parameters like
(Mettlach et al. 2002), known also as the Navy Standard significant and maximum wave height, wave direction
Surf Model (NSSM). This is a parametric 1D model that and bathymetry) are illustrated in Fig. 3b. Velocities of
estimates the wave induced longshore currents by solving almost 1m/s can be encountered for the nearshore cur-
only the longshore momentum balance equation. Hence, rents without accounting for the surface streaming effect
such a model can estimate only the longshore component that may additionally increase the oil spill propagation as
of the nearshore currents, while a 3D model can estimate discussed in the previous section.
also the cross shore component and the vertical variation Since in the upper side the current is negative and in
of the currents. The justification for the existence of the the lower side is positive the cross-shore currents might
1D models is that longshore currents are most relevant in be also relevant and a simulation with the SHORECIRC
the coastal applications. model system was also performed. SHORECIRC (Svend-
Fig. 3 presents the ISSM estimations for the long- sen et al. 2002) is a quasi–3D model that uses REFDIF
shore currents along two lines corresponding to the time (Kirby and Dalrymple 1994) as wave driver, and com-
frame 2002/03/13/h00 when it is supposed that the oil bines a numerical solution for the depth-integrated 2D
spill is approaching the coast. The ISSM is a user friendly horizontal momentum balance equations with an analyti-
model system recently developed for a quick evaluation cal solution for the 3D current profiles.
of the nearshore waves and currents. The ISSM stands for The limitations of the model exist, but the basic cir-
the Interface for SWAN and SURF models (Rusu et al. culation equations are solved accurately by considering
2008), and is a MATLAB GUI in the foreground which the non- linear feedback between wave-generated cur-
directs the integration of the 2D SWAN model with the rents and waves that generate them. SHORECIRC is a
1D SURF model in the background. time-domain model and computationally resource de-
294 L. Rusu. Application of numerical models to evaluate oil spills propagation in the coastal environment of the Black Sea

manding. The entire current system, as given by he Coupling the wave prediction system with a circula-
SHORECIRC simulation, is illustrated in Fig. 3a from tion model would contribute to the development of a joint
where results that the oil spill would concentrate south of prediction system that may be use in an iterative way in
Mangalia close to the Bulgarian border and the pollution order to account better for the interactions between waves
may be spread again offshore due to the rip current that is and currents and to identify all the components that may
formed in that region. drive the oil spills in the Black Sea basin.

References
Booij, N.; Ris, R. C.; Holthuijsen, L. H. 1999. A third-
generation wave model for coastal regions. Part I. Model
description and validation, Journal of Geophysical Re-
search 104(C4): 7649–7666. doi:10.1029/98JC02622
Coastal Engineer Research Center (CERC) 1984. Shore Protec-
tion Manual, Army Corps of Engrs. Washington D.C.:
Govt. Printing Office.
Elliott, A. J. 1986. Share Diffusion and spread of oil in the
surface layers of the North Sea, Deutsche hydrographis-
che Zeitschrift 39: 113–137. doi:10.1007/BF02408134
Guedes Soares, C.; Rusu, E. 2005. SWAN Hindcast in the
Black Sea [CD], in Fifth International Symposium on
Ocean Wave Measurement and Analysis (WAVES2005),
3–7 July 2005. Madrid.
Hasselmann, K. 1962. On the non-linear energy transfer in a
gravity–wave spectrum. Part 1. General theory, Journal of
Fluid Mechanics 12: 481–500.
doi:10.1017/S0022112062000373
Hasselmann, K. 1974. On the spectral dissipation of ocean
waves due to white-capping, Boundary-Layer Meteorolo-
gy 6(1–2): 107–127.
Hasselmann, S.; Hasselmann, K. 1981. A symmetrical method
of computing the nonlinear transfer in a gravity-wave
spectrum, Hamburger Geophys. Einzelschriften, Reihe A:
Wiss. Abhand, 52, 138.
Hasselmann, S.; Hasselmann, K.; Allender, J. H.; Barnett, T. P.
1985. Computations and parameterizations of the nonline-
ar energy transfer in a gravity wave spectrum. Part II. Pa-
rameterizations of the nonlinear transfer for application in
wave models, Journal of Physical Oceanography 15(11):
Fig. 3. Estimation of the nearshore currents driving the pollu- 1378–1391. doi:10.1175/1520-0485(1985)015<
tion along the coast (time frame 2002/03/13/h00): a) 2D repre- 1378:CAPOTN>2.0.CO;2
sentation of the wave conditions and of the wave induced Herterich, K.; Hasselmann, K. 1980. A similarity relation for
currents in the nearshore (resulting from SWAN and the nonlinear energy transfer in a finite-depth gravity-
SHORECIRC simulations); b) Cross shore variations of the wave spectrum, Journal of Fluid Mechanics 97: 215–224.
longshore currents (resulting from ISSM simulations) doi:10.1017/S0022112080002522
Holthuijsen, H. 2007. Waves in oceanic and coastal waters.
5. Conclusions Cambridge: Cambridge University Press. 387 p.
doi:10.1017/CBO9780511618536
The main objective of the present work is to illustrate the Janssen, P. A. E. M. 1991. Quasi-linear theory of wind-wave
utility of numerical models in preventing the accidents by generation applied to wave forecasting, Journal of Physi-
a better knowledge of the environmental conditions in the cal Oceanography 21: 1631–1642. doi:10.1175/152 0-
nearshore. In the same time, for the case when an acci- 0485(1991)021<1631:QLTOWW>2.0.CO;2
dent is produced the main factors that are driving the Kirby, J. T.; Dalrymple, R. A. 1994. Combined Refraction/
pollution were analysed and a hypothetic case study was Diffraction Model – REF/DIF version 2.5. Documentation
considered. and User's Manual, Centre for Applied Coastal Research,
University of Delaware, Newark, DE 19716 CACR Re-
The present SWAN model implementation in the port No. 94–22. 171 p.
Black Sea brings the advantage of the model flexibility in
Komen, G. J.; Hasselmann, S. and Hasselmann, K. 1984. On the
tuning parameters and physical processes for a better existence of a fully developed wind sea spectrum, J. Phys.
calibration of the model in a specific site. From this point Oceanogr. 14: 1271–1285.
of view SWAN seems to be more appropriate for such Lakshmi, H. K.; Clayson, C. A. 2000. Small Scale Processes in
regional domains than WAM or WW3 that are considered Geophysical Fluid Flows. Academic Press – International
traditionally as the state of the art generation wave mod- Geophysics Series, Volume 67, San Diego, California. 888 p.
els for oceanic scales.
 Journal of Environmental Engineering and Landscape Management, 2010, 18(4): 288–295 295

Mettlach, T. R.; Earle, M. D.; Hsu, Y. L. 2002. Software Design Smith, J. M.; Sherlock, A. R.; Resio, D. T. 2001. STWAVE:
Document for the Navy Standard Surf Model Version 3.2, Steady-State Spectral Wave Model, user’s guide for
Naval Research Laboratory memorandum report NRL/ STWAVE Version 3.0 ERDC/CHL SR–01–01, U.S. Army
MR/7320–02–8289, Stennis Space Center, Mississippi. 187 p. Engineer Research and Development Center, Vicksburg,
Phillips, O. M. 1977. The Dynamics of the Upper Ocean. Cam- MS.
bridge University Press, Cambridge. 336 p. Stelling, G. S.; Leendertse, J. J. 1992. Approximation of
Rusu, E.; Ventura, Soares, C. 2005. Post Prestige Developments convective processes by cyclic AOI methods, in 2nd
for the Wave Modeling Techniques in the Coastal Envi- International Conference on Estuarine and Coastal
ronment of Portugal [CD], in Fifth International Sympo- Modeling, ASCE, 13–15 November, 1991. Tampa,
sium on Ocean Wave Measurement and Analysis Florida, 771–782.
(WAVES2005), 3–7 July, 2005, Madrid. Svendsen, I. A.; Haas, K.; Zhao, Q. 2002. Quasi-3D Nearshore
Rusu, E.; Rusu, L.; Guedes Soares, C. 2006. Assessing of Ex- Circulation Model SHORECIRC version 2.0. Center for
treme Wave Conditions in the Black Sea with Numerical Applied Coastal Research, University of Delaware, New-
Models, in 9th International Workshop on Wave ark, DE 19716, USA.
Hindcasting and Forecasting, 25–29 September, 2006. Tolman, H. L. 1999. User manual and system documentation of
Victoria, Canada. Available from Internet: <http:// WAVEWATCH III version 1.18, NOAA / NWS / NCEP /
www.waveworkshop.org/9thWaves/Papers/Rusu.pdf>. OMB technical note 166, 110 p.
Rusu, E.; Conley, D. C.; Coelho, E. F. 2008. A Hybrid Frame- WAMDI Group. 1988. The WAM model – a third generation
work for Predicting Waves and Longshore Currents, ocean wave prediction model, Journal of Physical Ocean-
Journal of Marine Systems 69: 59–73. ography 18: 1775–181. doi:10.1175/1520-0485(19
doi:10.1016/j.jmarsys.2007.02.009 88)018<1775:TWMTGO>2.0.CO;2
Samuels, W. B.; Huang, N. E.; Amstutz, D. E. 1982. An Oil
Spill Trajectory Analysis Model with a Variable Wind
Deflection Angle, Ocean Engineering 9: 347–360.
doi:10.1016/0029-8018(82)90028-2

SKAITINIŲ MODELIŲ TAIKYMAS VERTINANT IŠSILIEJUSIOS NAFTOS PASKLIDIMĄ JUODOSIOS

JŪROS PRIEKRANTĖS ZONOJE
L. Rusu
Santrauka
Per pastaruosius dešimtmečius plėtojant pramonę Juodosios jūros priekrantės zonoje, dėl atsitiktinių naftos išsiliejimų pa-
didėjo taršos rizika. Viena iš priežasčių, dėl kurių įvyksta tokios avarijos, yra netikslios bangų būklės prognozės. Darbo
tikslas – remiantis spektrinių fazės vidurkių apskaičiavimų modeliais, sukurti metodologiją, kuri leistų numatyti bangų
sklidimą priekrantės zonoje. Taip pat įvertinamos bangos sukeltos srovės. Jos gali būti vienas iš pagrindinių veiksnių, tu-
rinčių įtakos teršalų sklidimui jūroje. Į skaičiavimus įtraukiama Stokso tėkmė bei bangos sukeltos pakrantės srovės,
išreiškiamas paviršiaus srovių efektas pagal molekulinę klampą. Dėl ilgųjų pakrantės srovių prie kranto iš arčiau pavir-
šiaus zonos esančio šaltinio tarša dažniausiai sklinda palei pakrantę, todėl rezultatai pristatomi taikant modeliavimo siste-
mas ISSM ir SHORECIRC. Modeliuojant remtasi studija, kurioje vertinamas taršos sklidimas ties Rumunijos pakrante dėl
spėjamai įvykusios avarijos Glorijos gręžinio platformoje.
Reikšminiai žodžiai: ekologinis įspėjimas, naftos išsiliejimai, bangos, bangos sukeltos srovės, skaitiniai modeliai, Juodoji
jūra.

ИСПОЛЬЗОВАНИЕ ЧИСЛЕННЫХ МОДЕЛЕЙ ДЛЯ ОЦЕНКИ РАСПРОСТРАНЕНИЯ РАЗЛИВА

НЕФТИ В ПРИБРЕЖНОЙ ЗОНЕ ЧЁРНОГО МОРЯ
Л. Русу
Резюме
Постоянно увеличивающаяся в последние десятилетия экономическая активность в прибрежной зоне Чѐрного
моря ведѐт к бóльшему риску загрязнения из-за случайных разливов нефти. В связи с тем, что большая часть слу-
чаев разлива была связана с неадекватным прогнозом волнения, целью настоящей работы было разработать мето-
дологию, основанную на спектральных моделях волнения с усреднением фазы, способную предсказывать
распространение волн в прибрежной зоне. Было оценено также течение, вызванное волнением, которое может
стать решающим фактором в распространении загрязнения. Оно включает в себя стоксово дрейфовое течение и
прибрежное течение. Учтено также поверхностное течение, вызываемое ветром через касательные напряжения. В
качестве примера рассмотрено распространение в направлении румынского берега загрязнения, вызванного гипо-
тетической аварией на буровой платформе Глория.
Ключевые слова: экологическое предупреждение, разливы нефти, волны, течения, вызываемые волной,
численные модели, Черное море.

Liliana RUSU. Dr, Assist. Prof. (since 2004), University Dunarea de Jos of Galati, Faculty of Mechanical Engineering,
Department of Applied Mechanics. Publications: author of 1 textbook and over 30 publications. Research interests: wave
propagation and coastal transformation, wave-current interactions, oil spills propagation.

View publication stats