See discussions, stats, and author profiles for this publication at: [600217]

See discussions, stats, and author profiles for this publication at:
http://www.researchgate.net/publication/244429132
Reanalysis of the Wave Conditions in
the Approaches to the Portuguese
Port of Sines
CHAPTER
· JANUARY 2005
DOI: 10.1201/[anonimizat].ch137
CITATIONS
10
READS
34
3 AUTHORS
, INCLUDING:
Liliana Celia Rusu
Universitatea Dunarea de Jos Gal
…
67

PUBLICATIONS

339

CITATIONS

SEE PROFILE
Carlos Guedes Soares
University of Lisbon
1,055

PUBLICATIONS

7,676

CITATIONS

SEE PROFILE
Available from: Carlos Guedes Soares
Retrieved on: 20 October 2015

Maritime Transportation and Exploitation of Ocean and
Coastal Resources – Guedes Soares, Garbatov & Fonseca (eds)
©2005 Taylor & Francis Group, London, ISBN 0 415 39036 2
11371INTRODUCTION
The economy of Portugal is in great part influenced by
investments and operations taking place in the coastalarea, and therefore real time predictions of the nearshorewave conditions are important for the safe design and
exploitation of marine structures and installations in
particular in the approaches to the ports. That is whythe characterization of the nearshore wave conditionsalong the Portuguese coast has been a subject that inthe last years has had increasing interest in coastalmanagement and environmental assessment studies.
One way to validate a wave operational system is to
apply it to past data, i.e. to use it to perform hindcasts.In this perspective the goal of the present work is tomake reanalysis studies based on numerical models forassessing the nearshore wave conditions in the coastalenvironment of Portugal continental in particular at theapproaches of the port of Sines. The methodology pro-posed herewith is based on the SWAN spectral model(Booij et al. 2004), applied on the nearshore area.SWAN, which is the acronym from Simulation WavesNearshore, is a phase averaging wave model designedto obtain realistic estimates of wave parameters incoastal areas, lakes and estuaries from given wind,bottom, and current conditions. The model is basedon the action balance equation (or energy balance inthe absence of currents) with sources and sinks.
The SWAN boundary conditions were provided
by WAM, which is a 3rd generation model (WAMDIGroup, 1988) implemented for the entire North Atlantic
basin.
The present work analyses one of the periods con-
sidered as being between the most energetic ones everregistered in the Portuguese nearshore. It includes theend of 2000 and the beginning of 2001. Thus in the clas-sification of first twenty five highest wave events reg-istered at the Sines buoy (located in the area of interest)in the last 25 years, three of these storms took place inthe three-month period between December 2000 tillthe end of February 2001 (Rusu et al. 2004).
The reanalysis wind field for the North Atlantic
basin was provided with a resolution of 0.5 degrees bythe project HIPOCAS “Hindcast of Dynamic Processesof the Ocean and Coastal Areas of Europe”. Within thisproject a hindcast analysis was made of 44 years windand wave conditions (between 1958 and 2001), in thecoastal European waters (Guedes Soares et al. 2002).
A further development of these reanalysis studies
would be to develop and calibrate in the Portuguese
nearshore an effective wave prediction system basedon numerical models. This system should cover anyarea and situation and would be focussed especially inapproaches to the important ports of Portugal.
2THE COMPUTATIONAL SCHEME
For assessing the wave conditions close to the
Portuguese coasts two state-of-the-art spectral waveReanalysis of the wave conditions on the approaches to the
Portuguese port of Sines
L. Rusu, P . Pilar & C. Guedes Soares
Unit of Marine Technology and Engineering, Technical University of Lisbon, Instituto Superior Técnico,
Av. Rovisco Pais, Lisboa, Portugal
ABSTRACT: A reanalysis is presented of the wave conditions on the approaches to the port of Sines in the
Portuguese continental coast. The main area of interest is south of Lisbon especially concentrated on the coastal
environment on the approaches to the port of Sines. The same scheme was used for investigating the wave con-ditions in other locations of the Portuguese nearshore both in the Northern and Southern coasts. The initial bound-ary conditions are provided by WAM simulations for the entire North Atlantic basin. After wards, three successivenested areas were implemented in SWAN. To validate the results, use was made of the data from a directionalbuoy located offshore the Sines port, at approximately at 100 meters water depth. The simulations were performedfrom the beginning of September 2000 to 10 March 2001, which includes one of the most energetic periods everregistered on this coast.

models were nested. These are WAM for generation
and SWAN for the coastal environment (used both forgeneration and wave transform in the nearshore). Theyare both third generation wave models and use the sameformulations for the source terms. SWAN containssome additional formulations specific for shallowwater. However, the numerical techniques implemented
are very different in the two models. Unlike WAM,SWAN uses an implicit upwind scheme which isunconditionally stable.
In order to increase gradually the resolution in the
models they were nested in themselves several times.Thus for the WAM simulations three domains wereused as presented in (Rusu et al. 2005).
The coupling between the two models was made
by nesting a large SWAN area into the WAM model
as presented in Figure 1. This area covers the entirewest Iberian coast and it is used as a general driver for
the coastal simulations. Medium and high resolutiondomains were successively nested into it in order tocover various areas in the Portuguese nearshore. Inthe present application the target area is centred onthe entrance to the port of Sines in the central part ofPortugal continental as illustrated in Figure 1.
The characteristics for the corresponding computa-
tional grids used in the SWAN simulations are given
in Table 1.SWAN simulations were performed using the non
stationary mode. Two numerical schemes available inthe model were used. Since this is a large area thegeneral driver chosen was the second-order upwindscheme with third-order diffusion, so-called S&L.For the other two areas the first-order scheme back-
ward space, backward time (BSBT) was used. This is
imposed into the model by the CFL condition ( /H9004t/H11088
/H9004x/c which is also known as the condition Courant-
Friedrichs-Levy).
As regards the physical processes a compromise
was made between the default parameterizations and
the most appropriate ones. For each phenomenon orprocess there are usually in SWAN more alternativesgiven by different formulations. Moreover for processessuch as bottom friction, whitecapping, etc. there aretuneable coefficients that can have values in a widescale. The default parameterization is usually the mostcommon one and also is certainly the simplest waywhen working with the wave models. However these
possibilities opened by the SWAN model of using dif-ferent formulations for physical processes as well as awide range of values for some coefficients the modelscan lead to considerably improvements in the resultsdespite the fact that they require additional work forcalibrating.
For the wave generation the Janssen formulation
with the growing coefficient activated was used, andalternatively the Cumulative Steepness Method for thewhitecapping dissipation was tested. The bottom fric-
tion used the JONSWAP formulation, while for thequadruplet wave-wave interactions the fully explicitcomputations of the nonlinear transfer with DIA (Dis-crete Interaction Appro ximation) per sweep were used.
In the medium and high resolution areas the dif-
fraction was activated, which in SWAN (version40.41) is introduced in a phase decoupled approach,in order to account for the diffraction effect of the twocapes located in those areas (Cabo Espichel and CaboSines). However at least for the present applicationthe diffraction effect is not very important. In the highresolution area the triad wave-wave interaction wasalso activated.
The computations were performed in the non sta-
tionary mode with a 20 min time step (for the case of
1138
Figure 1. Bathymetric map and geographical spaces of the
SWAN simulations.Table 1. Computational grids for the SWAN simulations.
ngx/H11003ngy
Grids /H9004x/H11003/H9004y (°) /H9004t(s) nf n /H9258/H11005 np
Large 0.05° /H110030.1° 1200 30 36 101 /H11003101
/H1100510201
Medium 0.02° /H110030.02° 1200 30 36 62 /H1100375
/H110054650
Small 0.005° /H110030.005° 1200 30 36 101 /H11003101
/H1100510201

the pure non stationary simulations). The number of
iterations was set from 1 (which is the default value),to 4. In this way the numerical accuracy reached untilthe model passes to another time step was increased.Every time when passing from one set of simulationsto the following one, hot files were used.
3FIELD DATA ANAL YSES
The simulations were performed in the period from
September 2000 until the middle of March 2001. Thistime interval is characterized by two periods: the firstthree months were normal from the energetic point ofview while the last period was highly energetic.
Figure 2 illustrates the results for the peak of the
storm at the beginning of December 2000 for theSWAN coarse area which covers the entire west IberianCoast. Significant wave heights over 9 meters and windvelocity greater than 20 m/s can be encountered at
that time in the area considered. The main wind direc-tion for this case is south west.
The black circle in Figures 2–3, 4, 5, 6 represents
the location where the largest significant wave heightswere identified.Figures 3–4 present the results for another impor-
tant storm registered at the beginning of January 2001in the same area. The maximum significant waveheights approach again the value of 9 meters while theswell component is about 8.5 meters as can be seen in
Figure 4. The wind pattern is this time from North West
(which is the most common in the Iberian nearshore),with wind velocities about 16 m/s.
In Figures 5–6 are given the wave vectors for the
medium and respectively high resolution areas for thesame storm.
Figure 6 presents also the location of the Sines
buoy. The coordinates of this buoy are ( /H110028.9289W
37.9211N) and it corresponds an average depth of 97meters to that position.
4DIRECT COMPARISONS AND
STATISTICAL RESULTS
Figures 7–8, 9, 10 present the direct comparisons for
the entire period analysed (1 September 2000 and 10March 2001) between the Sines buoy and the resultsof the SWAN simulations. These direct comparisons
1139
Figure 2. Significant wave height fields, wave and wind
vectors, 2000/12/07/h06.
Figure 3. Significant wave height fields, wave and wind
vectors, 2001/01/02/h00.

are made for the significant wave height, mean wave
period, peak wave period and mean wave direction.
A good correspondence has been encountered for
all parameters compared. From a climatologic pointof view the time interval considered can be divided into
two parts. The first three months which are commonfrom an energetic point of view and the second periodfrom December 2000 till 10th of March 2001 whichis highly energetic.
1140
Figure 4. Significant swell height fields and swell vectors,
2001/01/02/h00.
Figure 5. Medium resolution area for Lisbon-Sines
nearshore, bathymetric map and wave vectors, 2001/01/02/h00.
Figure 6. High resolution area for Sines approaches, bathy-
metric map, wave vectors and buoy location, 2001/01/02/h00.
Figure 7. Direct comparison SWAN – Sines Buoy, signifi-
cant wave height, 2000/09/01h00-2001/03/10/h00.
Figure 8. Direct comparison SWAN – Sines Buoy, mean
period, 2000/09/01h00-2001/03/10/h00.

In order to get the statistical results were computed
for the same parameters (the significant wave height,mean wave period, peak wave period and mean wavedirection) the average values for measurements (B
med)
and simulations (S med), as well as the bias, root mean
square error (RMSE), scatter index (SI) and Pearson’ sCorrelation Coefficient (r). The definitions of theseparameters are given in Rusu et al. 2005).
The values of these statistical parameters for the
entire period between 1 September 2000 to 10 March2001 are given in Table 2.
The response of the system based on numerical mod-
els might be different when passing from normal waveconditions to highly energetic conditions. It is actu-ally well known the tendency of most wave generationmodels of underevaluating the peak of the highest
storms. For this reason the statistical parameters (aver-age values, bias RMSE, scatter index and correlationcoefficient) were computed separately for normalwave conditions and for storm conditions. The limit
between the common wave climate and the highlyenergetic conditions was considered to be the signifi-cant wave height of 4.5 meters. Table 3 presents thestatistics computed for all cases when the significantwave height is lower than 4.5 meters while Table 4 are
presents the same statistics for the cases when the sig-nificant wave height is greater or equal to this limit.
Figures 11–12, 13 present the scatter plots of the
significant wave heights for the same three cases considered.
1141
Figure 9. Direct comparison SWAN – Sines Buoy, peak
period, 2000/09/01h00-2001/03/10/h00.
Figure 10. Direct comparison SWAN – Sines Buoy, mean
direction, 2000/09/01h00-2001/03/10/h00.
Table 2. Statistical results for the entire wave field in the
period 1 September 2000–10 March 2001.
n/H110051515 B med Smed Bias RMSE SI r
Hs (m) 2.40 2.59 /H110020.19 0.50 0.21 0.93
Tm (s) 7.18 8.00 /H110020.82 1.50 0.21 0.80
Tp (s) 12.02 12.61 /H110020.58 2.04 0.17 0.69
MDir (°) 296.7 290.1 6.57 13.16 0.04 0.76Table 3. Statistical results for normal energetic conditions
(Hs/H110214.5 m), in the period 1 September 2000–10 March 2001.
n/H110051401 B med Smed Bias RMSE SI r
Hs (m) 2.17 2.39 /H110020.21 0.48 0.22 0.92
Tm (s) 7.02 7.79 /H110020.78 1.46 0.21 0.78
Tp (s) 11.80 12.33 /H110020.53 2.03 0.17 0.64
MDir (°) 297.3 291.1 6.22 12.82 0.04 0.76
Table 4. Statistical results for high energetic conditions
(Hs/H11022/H11005 4.5 m), in the period 1 September 2000–10 March
2001.
n/H11005114 B med Smed Bias RMSE SI r
Hs (m) 5.1 5.05 0.05 0.68 0.13 0.63
Tm (s) 9.20 10.53 /H110021.33 1.83 0.20 0.76
Tp (s) 14.76 16.01 /H110021.26 2.10 0.14 0.75
MDir (°) 289.4 278.6 10.8 16.77 0.06 0.71
Figure 11. Hs scatter plot, the entire wave field 2000/09/
01h00-2001/03/10/h21.

5FINAL CONSIDERATIONS
Both the direct comparisons at the Sines buoy and the
statistics show a good correspondence for the resultsprovided by the numerical wave models.
In terms of mean wave direction the results are in
general good. This is probably due to the fact that theinfluence of the refraction is very strong in the coastalenvironment and the errors, induced by the windfields and the large scale model are being attenuated
in this way.
There are good correlation coefficients and RMS
errors for the significant wave heights and periods forthe entire period and also for the relatively calm con-ditions while the scatter indexes are also acceptable.For storm conditions the scatter index is very good,
the bias is positive and the RMSE is satisfactory.
In general SWAN overestimates the significant
wave heights and periods (negative bias), the ten-
dency being contrary only for the cases of the stormspeaks. For example in 2000.12.07h06 (Fig. 2) there isa consistent overestimation of Hs (6.01 m at the buoyand 6.53 m given by SWAN). In this case there was avery high wind velocity in the considered area (over
20 m/s). For the storm of 2001.01.02h00 a significantunderestimation of the significant wave height wasencountered (7.67 m at the buoy, while 6.54 m wasgiven by SWAN).
The results are influenced by the bathymetry, which
is not very accurate because it was interpolated fromdata obtained from the internet. Furthermore, the windfield with a resolution of 0.5° is very good for ocean
areas, but it cannot reflect well the local winds.
There are some ways to improve the present results.
The most obvious ones would be to use a more accuratebathymetry in the medium and high resolution areas,complemented by higher resolution wind fields.
ACKNOWLEDGMENT
The first author has been funded by Fundação para
a Ciência e Tecnologia (Portuguese Foundation forScience and Technology) under grant SFRH/BD/
13176/200 .
The wave data recorded by the buoy was kindly sup-
plied by the Administration of the Port of Sines.
REFERENCES
Booij, N., Haagsma, I.J.G., Holthuijsen, L.H., Kieftenburg,
A.T.M.M., Ris, R.C., van der Westhuysen, A.J. &Zijlema, M. 2004. User Manual for SWAN, Version 40.41,Delft University of Technology, Delft, the Netherlands,115p.
Guedes Soares, C., Weisse, R., Carretero, J.C. & Alvarez, E.
2002. A 40 years Hindcast of Wind, Sea Level and Wavesin European Waters, Proceedings of the 21st International
Conference on Offshore Mechanics and Arctic Engineering(OMAE’02) , ASME, Paper OMAE2002-SR28604.
Rusu, E., Ventura Soares, C., Pinto, J.P . & Silva, R. 2004.
Extreme Events and Wave Forecast in the IberianNearshore, Coastal Engineering 2004 , Jane McKee Smith,
World Scientific (Ed.), 727–739.
Rusu, L., Pilar, P . & Guedes Soares, C. 2005. Hindcasts of
the Wave Conditions in Approaches to Ports of the Northof Portugal, Fifth International Symposium on Ocean
Wave Measurement and Analysis – W AVES 2005 –Madrid,
Spain, 3rd–7th July 2005, CD edition, 9p.
WAMDI Group. 1988. The WAM model – a third generation
ocean wave prediction model, J. Phys. Oceanogr. , 18,
1775–1810.
1142
Figure 12. Hs scatter plot, normal energetic conditions
(Hs/H110214.5 m), 2000/09/01h00-2001/03/10/h21.
Figure 13. Hs scatter plot, high energetic conditions
(Hs/H11022/H11005 4.5 m), 2000/09/01h00-2001/03/10/h21.