JOURNAL OF APPLIED ENGINEERING SCIENCES VOL.

6(19), ISSUE 2/2016

ISSN: 2247-3769 / e-ISSN: 2284-7197 ART.NO. 207, pp. 33-40

A QGIS PLUGIN FOR OFFSHORE WAVE HINDCASTING BASED ON GEOGRAPHIC

TRANSPOSITION OF WAVE GAUGE DATA

F. Pasanisi a, *, C. Tebano a
a
ENEA, Italian National Agency for New Technologies, Energy and Sustainable Economic Development, Territorial and
Production Systems Sustainability Department, Models and Technologies for Mitigation of Human Impacts and Natural Hazards
Division, Seismic Engineering and Prevention of Natural Hazards Laboratory, Piazzale Enrico Fermi 1, 80055, Portici (NA), Italy

Received: 05.10.2016 / Accepted: 10.10.2016 / Revised: 25.11.2016 / Available online: 15.12.2016

DOI: 10.1515/jaes-2016-0015

KEY WORDS: sea waves, hindcasting, geographic transposition, QGIS

ABSTRACT:

The paper presents a first experimental version of the original QGIS plugin QWaveTransposition that numerically implements the
geographic transposition of wave gauge data method proposed by Contini and De Girolamo (1998) for offshore wave hindcasting.
The method allows one to transfer wave data measured at a given gauging station to a virtual station located offshore the area of
interest, by comparing the effective fetches at both stations. The QWaveTransposition plugin was implemented in Python
programming language, including the NumPy package for numerical computations. A graphical user interface was developed to
manage the input/output data and model parameters. The fetch geometry at real and virtual stations can be imported by selecting
appropriate vector layers from the QGIS map. An application to a sample site in southern Italy is presented for example purposes.

1. INTRODUCTION When both wind and wave data are available at a given project
site, the choice between the two approaches is a delicate task
The estimation of offshore wave climate is a key issue in and a universally accepted rule of thumb cannot be defined.
maritime and coastal engineering, as it is the first, fundamental Main factors to be considered are: (a) the data accuracy and
step of every project dealing, e.g., with navigation, harbours acquisition rate; (b) the duration of time series and the gauging
protection, design of offshore and coastal structures, coastal station efficiency; (c) the proximity of the station to the project
planning and remediation. site.

Basically, two main conceptual approaches to estimate wind- A brief overview of the above introduced families of methods
generated wave conditions can be distinguished. The first for wave prediction is given in the next sections.
comprises empirical methods and mathematical models in
which wave parameters are derived from wind data. When 1.1 Estimation of wave parameters from wind data
predictive methods are applied to current or predicted
meteorological data, the process is referred to as “forecasting”. A number of studies were performed since the 1940s, aimed at
Otherwise, when the above methods are applied to derive wave deriving empirical methods to predict wave conditions
information based on historical wind data, the process is considering the energy transfer from the wind to the sea surface
referred to as “hindcasting” (Arthur, 1950). The second (Sverdrup and Munk, 1947; Arthur, 1950; Bretschneider, 1965).
approach is to gain information from direct measurements of Among them, the SMB method (SPM, 1984), so called from the
wave parameters at gauging stations close to the study area. authors Sverdrup, Munk, and Bretschneider, is the most widely
used. In the SMB method spectral significant wave height and

* Corresponding author, e-mail: [email protected]

33

Unauthenticated
Download Date | 3/1/17 12:16 PM
JOURNAL OF APPLIED ENGINEERING SCIENCES VOL. 6(19), ISSUE 2/2016
ISSN: 2247-3769 / e-ISSN: 2284-7197 ART.NO. 207, pp. 33-40

peak period are predicted by means of empirical formulae as a

function of the following parameters:
− the fetch length (F), i.e. the linear extent of an ideal
region, measured along the wind direction, in which
the wind blows with constant speed and direction;
− the wind-stress factor (UA), i.e. an adjusted measure of
wind speed;
− the duration of the wind event (t).

The conditions of uniform fetch and constant wind speed and

direction are rarely satisfied in practice. However, for inland
waters or enclosed basins as bays, lakes or reservoirs, it can be
assumed that the wind conditions do not vary significantly over
the water body during a meteorological event. In this case the Figure 1. Relation between incident wind and wave direction
wave generation area is limited by the boundary of the water
body and the fetch extent at given point can be estimated In the above formula the angle θ between wind and wave
measuring the distance to the coast upwind. direction can theoretically vary in the range 0-180° and F(θ) is
the corresponding fetch length. In practical applications, the
The above assumption can be considered realistic also for the wind-wave angles in Equation (2) are discretized, e.g., at 1°
Mediterranean Sea, though an upper limit to fetch length is increments (Leenknecht et al., 1992).
normally imposed (typically, 500 km) to consider valid the
hypothesis of uniform wind conditions. Besides the above introduced empirical methods, a number of
mathematical models were proposed for the estimation of wind-
A number of corrective formulations were proposed to achieve generated wave parameters, based on different formulations of
a more realistic description of the wave generation process the wave growth and decay equations, as reported, e.g., in
(Saville, 1954; Seymour, 1977). In particular, the concept of Komen et al. (1994), Holthuijsen et al. (1989), Booij et al.
“effective fetch” was introduced to account for the effect of (1999), and Liu et al. (2002). Wind-wave modelling
fetch shape and directional spreading of wave energy. experienced a very rapid development and diffusion over the
last decades, similarly to other computer-based numerical
In a generalised formulation, for a given wind direction the techniques (Cavaleri et al., 2007; Rogers et al., 2014). The
effective fetch Feff can be defined as follows: literature on this topic is very extensive and its discussion is
beyond the aim of the present work. It is just noteworthy to
θw observe that numerical models generally require a
θ
∑θ F (θ ) ⋅ cos
=−
i
n +1
θi
(1)
computational effort that can be justified when significant
Feff = i w
θw
variations in meteorological conditions takes place, and
∑θ cos
frequent and accurate input data are available (Catini et al.,
n
θi
θ =−
2010; Liberti et al., 2013; Carillo et al., 2015; ENEA, 2016;
i w
ISPRA, 2016).
where: θi = angle from wind direction, typically varying 1.2 In situ measurement of wave parameters
either over a 90° arc (θw=45°) or a 180° arc (θw=90°);
F(θi) = straight line fetch length measured along the Visual observations of waves from coastal stations or ships are
direction θi (commonly referred to as “geographic the first historically reported methods for the direct estimation
fetch”); of wave parameters. However, they normally produce data with
n = exponent that accounts for the wind-generated a high degree of subjectivity and not compatible with those
wave energy directional spreading. Typical values are obtained from instrumental measures.
n=1 or n=2.
Since the 1960s, a number of different techniques and
Seymour (1977) also argued that due to wave energy directional instrumentations for in situ measurement of wave parameters
spreading, wave direction generally does not coincide with were proposed by scientists, practitioners, and manufacturers,
wind direction. Based on field data analysis, Donelan (1980) based on different theoretical formulations and operation
proposed to estimate the wave direction from the input wind principles. Available methods (WMO, 1998) comprise
direction by maximizing the resulting wave period that depends measurements: (a) from below the sea surface; (b) at the sea
on the fetch lengths and wind-wave angles (figure 1). Following surface; (c) from above the sea surface. A review can be found
Donelan’s approach, Smith (1991) proposed to predict the in Barstow et al. (2005) and DBCP (2016).
wind-generated wave direction by maximizing the function
given below: Direct measurements of wave parameters experienced a
significant growth during the last decades and monitoring
Φ (θ ) = (cosθ ) ⋅ F (θ )
0.44 0.28 programs by means of wave gauging networks were established
(2)
in most of developed countries.

34

Unauthenticated
Download Date | 3/1/17 12:16 PM
JOURNAL OF APPLIED ENGINEERING SCIENCES VOL. 6(19), ISSUE 2/2016
ISSN: 2247-3769 / e-ISSN: 2284-7197 ART.NO. 207, pp. 33-40

As an example, the first wave data collection stations along the 1

U.S. Pacific coast were deployed in the mid-1970s (Seymour g Hm ⎛ gF ⎞2

= 1.6 ⋅10 −3
⎜ 2 ⎟ (3)
and Sessions, 1976), giving rise to the monitoring network 2 ⎜U ⎟
UA ⎝ A ⎠
known as Coastal Data Information Program. The network 1
steadily grew during the last decades and currently covers the g Tm ⎛ g F ⎞3
= 2.857 ⋅ 10 −1 ⎜ 2⎟ (4)
entire U.S. coastline (CDIP, 2016).
UA ⎜U ⎟
⎝ A ⎠
In Italy, the Italian Data Buoy Network (RON, acronym of the
Italian words “Rete Ondametrica Nazionale”) was established Under the assumption that wind conditions are the same at real
in 1989. Initially composed of 8 directional buoys, the RON and virtual station, writing the Equations (3) and (4) at both
was successively upgraded and extended over the years, and sites, the following equations can be derived:
currently consists of 15 wave gauging stations (figure 2)
deployed at about 100 m depth along the Italian coastline 1
(H m )V ⎛F ⎞2
(Corsini et al., 2006; Bencivenga et al., 2012). = ⎜⎜ V ⎟⎟ (5)
( H m )R ⎝ FR ⎠
1
(Tm )V ⎛F ⎞3
= ⎜⎜ V ⎟⎟ (6)
(Tm )R ⎝ FR ⎠

The ratios at the right-hand side of Equations (5) and (6) are
referred to as “transposition coefficients” and allow one to
derive the wave height and period at the virtual station from the
data measured at real station, considering the different wave
exposures due to the different geographic positions of the sites.

The deviation between wind and wave direction is also to be

considered. Namely, for each recorded wave event with
direction (θwave)R the corresponding wind direction θwind (that is
the same for both real and virtual station) must be computed,
maximizing the function in Equation (2) for the real station
Figure 2. The Italian Data Buoy Network. Data from fetches. Then, considering the transposition coefficient for the
Bencivenga et al. (2012). Basemap: © OpenStreetMap direction θwind, the wave height and period at virtual station are
contributors, license CC-BY-SA computed using Equations (5) and (6). Finally, the wave
direction (θwave)V at virtual station is estimated by maximizing
the function in Equation (2) for the virtual station fetches.
2. GEOGRAPHIC TRANSPOSITION OF WAVE
GAUGE DATA
3. THE QWAVETRANSPOSITION PLUGIN
The “geographic transposition of wave gauge data” method,
originally formulated by Contini and De Girolamo (1998), is a The geographic transposition method is much appreciated by
widely used hindcasting procedure for the estimation of coastal scientists and engineers and is conceptually simple to
offshore wave parameters at a given project site (that can be use. However, in practical applications where a huge amount of
referred to as a “virtual station”) based on wave data measured wave data is to be processed, computer-based automated
at a “real” gauging station. The method is based on the procedures are needed. To this aim, the original software
following hypotheses: QWaveTransposition was developed as a plugin to QGIS
a) the wind speed and direction are the same at both real Geographic Information System (QGIS Development Team,
and virtual stations; 2016). The code was implemented in Python, including the
b) the extent of the wave generation region can be NumPy (2016) package for numerical calculation. The
described by the effective fetches; QWaveTransposition plugin can be run by clicking on the
c) the wind blows over the fetch long enough to assume special icon added to the QGIS toolbar (figure 3).
that wave conditions are independent of the wind duration
(fetch-limited conditions). The input/output data and model parameters can be managed
through a graphical user interface (GUI). The user form is
Under the above conditions, the spectral significant wave height composed of three sections (Fetch data, Wave data, and
Hm and peak period Tm in deep water can be predicted, using the Calculation), as described below.
SMB method, as a function of fetch F and wind-stress factor UA
by the following empirical formulae (SPM, 1984): 3.1 Fetch data

In this section the geographic fetches (input data) are specified

for the real and virtual wave station. Two options are available,

35

Unauthenticated
Download Date | 3/1/17 12:16 PM
JOURNAL OF APPLIED ENGINEERING SCIENCES VOL. 6(19), ISSUE 2/2016
ISSN: 2247-3769 / e-ISSN: 2284-7197 ART.NO. 207, pp. 33-40

that can be chosen checking one of the two boxes on the right
side of the form (figure 3):
− Load from map layers. If this option is checked, the
user selects from the map a line vector layer
containing the straight line fetches at real/virtual
station over the entire 360° compass, discretized at 1°
steps. The azimuth relative to north and the
corresponding fetch length are read from the attribute
table of the layer.
− Load from text files. If this option is checked, the
paths of external ASCII files containing the azimuth
and fetch length data are specified by the user. In this
case, it is not necessary that fetch layers are present in
the QGIS map.

Figure 4. Input wave data are imported from a text file specified
through the command button shown in the red rectangle

The input data are the wave conditions measured at the real
gauging station and are imported as an ASCII file Each line of
the text file represents a wave record, described by the
following data:
− ID number;
− wave direction;
− significant wave height;
− peak wave period.
The above information can be derived from available wave
gauging station data and normally not much effort is required
for data preparation.

Figure 3. The QWaveTransposition user interface and icon The model output is an ASCII file formatted in the same way as
(blue circle). If the option “load from map layers” is checked the input file, reporting transposed wave parameters for each
(red rectangle), appropriate layers containing the fetch data are event in the input wave series.
selected from map layers’ list
3.3 Calculation
Since the geographic transposition method is based on non-
dimensional equations, the input fetch lengths can be Once all the input/output files have been defined and model
indifferently expressed in any unit of measurements (e.g., parameters have been set, the “Run Wave Transposition”
meters, kilometres, or nautical miles), provided that the units command button can be clicked to run the model.
are the same for real and virtual station.
In the first steps of software execution the fetch data are
The value of n exponent in the Equation (1) for the effective processed. Namely, the following actions are performed for
fetch calculation is also specified by the user. The available both real and virtual station:
options (n=1 or n=2) can be selected checking one of the radio − Reading of geographic fetches azimuth and length;
buttons in the form. − Computation of effective fetches for each direction.
The summation in Equation (1) is extended over a
3.2 Wave data 180° arc centred on the current azimuth value.
− Computation of the wind-wave angles for each
In this section the user selects the paths for input and output direction, maximizing the function Φ in the Equation
wave data files (figure 4). (2) for different values of effective fetches computed
at previous step.
− Computation of the wave height and period
transposition coefficients for each direction, applying
Equations (5) and (6).

Once the previous calculations have been executed, the input

waves are processed to derive the transposed wave parameters

36

Unauthenticated
Download Date | 3/1/17 12:16 PM
JOURNAL OF APPLIED ENGINEERING SCIENCES VOL. 6(19), ISSUE 2/2016
ISSN: 2247-3769 / e-ISSN: 2284-7197 ART.NO. 207, pp. 33-40

at virtual station. Namely, for each wave record described by a

single row in the input file the following steps are performed:
− Reading of input wave direction, height and period at
real station;
− Computation of wind direction from wave direction,
considering the estimated wind-wave angles at the
real station;
− Computation of transposed wave height and period at
virtual station. The input wave parameters are
multiplied by the transposition coefficients
corresponding to the wind direction computed at the
previous step;
− Computation of transposed wave direction at virtual
station, considering the estimated wind-wave angles at
the virtual station;
− Writing of transposed wave parameters to the Figure 5. Geographic setting of the real and virtual station.
specified output file. Basemap: © OpenStreetMap contributors, license CC-BY-SA

It is to be observed that the estimation of wind directions at real Polar plots of geographic and effective fetches at real and
station based on measured wave directions and wind-wave virtual stations are reported, respectively, in figures 6 and 7.
angles reduces, in practice, the directional range of the incident Geographic fetches were limited to maximum 500 km. The
waves that can be processed. Namely, when fetch-based wind- value n=1 was chosen for the exponent in Equation (1) for the
wave angles are computed for each potential wind direction in effective fetches calculation.
the range 0-360°, the resulting range of potential wave
directions is generally narrower than 360°, depending on the The computed wind-wave angles and wave directions for each
values of the effective fetches that maximize the function Φ in wind direction at both stations are plotted in figures 8 and 9. It
Equation (2). can be observed that, as previously specified, the directional
range of the waves that can be processed is narrower than the
Thus, a wave record with direction outside the above computed entire 360° arc. In the present case, as illustrated in figure 9, the
theoretical range cannot be processed, as it cannot be associated geographic transposition can be applied only to waves
to any wind direction. In this case the model returns a zero comprised in the sector 89-352°.
values for wave height and period, and the record is flagged
assigning the 360° value to wave direction. Despite it can
appear unrealistic, the effects on the final results are generally
limited, since directions excluded from calculation are normally
characterized by short fetches and low values of wave height
and period are expected.

The above considerations will be better clarified in the next

section in which an example application is presented.

4. EXAMPLE APPLICATION

For example purposes, the QWaveTransposition software were

applied to estimate the incident wave climate offshore
Favignana Island, near the western coast of Sicily, in southern
Italy.

Input wave data were derived from the Mazara gauging station
of the Italian Data Buoy Network, located at latitude
37°31’05’’N, longitude 12°32’00’’E (Bencivenga et al., 2012).
A virtual wave station was placed off the southern coast of
Favignana island, at about 45 km from the Mazara station Figure 6. Polar plots of geographic fetches (solid lines) and
(figure 5). effective fetches (dashed lines) at the real station

37

Unauthenticated
Download Date | 3/1/17 12:16 PM
JOURNAL OF APPLIED ENGINEERING SCIENCES VOL. 6(19), ISSUE 2/2016
ISSN: 2247-3769 / e-ISSN: 2284-7197 ART.NO. 207, pp. 33-40

Figure 10. Transposition coefficients for wave parameters

Figure 7. Polar plots of geographic fetches (solid lines) and

effective fetches (dashed lines) at the virtual station

The computed transposition coefficients for each wind direction

are plotted in figure 10. The values are lower than 1 because, as
can be observed from figures 6 and 7, the effective fetches at
the virtual station are lower than those computed at the real
station. The figure 11 reports a rose plot of the mean wave Figure 11. Rose plot of mean wave climate derived from input
climate derived from input data. Consistently with the wave data at real gauging station
exposure of the site, two main directional sectors for prevailing The figure 12 illustrates the QWaveTransposition user interface
incident waves can be observed, comprising the east by south during the model run. A display label and a progress bar
and the west by north directions, respectively. indicate the current process, (e.g. fetch calculation, estimation
of wind-wave angles, performing transposition). The directional
wave range in which transposition can be performed is
displayed in the special labels above the progress bar.

Figure 8. Deviation angle between wind and wave directions

Figure 9. Wave directions for different wind directions, and Figure 12. A view of the user interface during the model run
indication of the directional range in which the geographic
transposition method is applicable

38

Unauthenticated
Download Date | 3/1/17 12:16 PM
JOURNAL OF APPLIED ENGINEERING SCIENCES VOL. 6(19), ISSUE 2/2016
ISSN: 2247-3769 / e-ISSN: 2284-7197 ART.NO. 207, pp. 33-40

The labels in the red rectangle (figure 12) display the limits of implementations of the algorithm and numerical procedures, as
the directional sector where the geographic transposition well as modifications to the user interface or input/output
method is applicable in the present case. format will also be possibly performed.

The final results are illustrated in figure 13, reporting a rose plot
of the mean wave climate at virtual station derived from 6. REFERENCES
transposed wave data.
Arthur, R.S., 1950. Wave Forecasting and Hindcasting. In:
Proceedings of 1st International Conference on Coastal
Engineering, Long Beach, CA, USA, pp. 82–87.

Barstow, S.F., Bidlot, J.-R.. Caires, S., Donelan, M.A.,

Drennan, W.M., Dupuis, H., Graber, H.C., et al., 2005.
Measuring and Analysing the Directional Spectrum of Ocean
Waves. COST Action 714, Brussels, Belgium.

Bencivenga, M., Nardone, G., Ruggiero, F., Calore, D., 2012.

The Italian Data Buoy Network (RON). Advances in Fluid
Mechanics IX, WIT Transactions on Engineering Sciences, 74,
pp. 321–332.

Booij, N., Ris, R.C., Holthuijsen, L.H., 1999. A Third-

Generation Wave Model for Coastal Regions. Journal of
Figure 13. Rose plot of mean wave climate derived from output Geophysical Research, 104(C4), pp. 7649–7666.
data transposed at virtual gauging station
Bretschneider, C.L., 1965. Generation of Waves by Wind. State
Comparing figures 11 and 13, some differences can be observed of the Art. NESCO Report SN-134-6, National Engineering
between real and virtual station that are consistent with Science Company, Washington, DC, USA.
different geographic positions and wave exposures of the two
sites. Carillo, A., Sannino, G., Lombardi, E., 2015. Wave Energy
Potential: a Forecasting System for the Mediterranean Basin.
In particular, a general reduction of wave heights at virtual Energia, Ambiente e Innovazione/Energy, Environment, and
station compared to the real station can be observed, as a Innovation, Special Issue II-2015, pp. 16–21.
consequence of the lower extensions of fetches. Moreover, a
Catini, F., Montagna, F., Franco, L., Bellotti, G., Corsini, S.,
less pronounced influence of the waves coming from the
Inghilesi, R., Orasi, A., 2010. Development of a High-
northern directions can be noticed, due to the sheltering effect
Resolution Nearshore Wave Forecasting/Hindcasting System
of Favignana and Marettimo islands north of the virtual station
for the Italian Coasts. In: Proceedings of 32nd International
(figure 5).
Conference on Coastal Engineering, Shanghai, China, pp. 409–
422.
Finally, the graph in figure 13 clearly shows the limitations to
the range of wave directions for which the transposition method Cavaleri, L., Alves, J.-H.G.M., Ardhuin, F., Babanin, A.,
is applicable. It can be observed, however, that the excluded Banner, M., Belibassakis, K., Benoit, M., et al., 2007. Wave
directions are characterized by the lowest values of fetch Modelling–The State of the Art. Progress in Oceanography, 75,
lengths. pp. 603–674.

CDIP, 2016. The Coastal Data Information Program.

5. CONCLUSIONS http://www.cdip.ucsd.edu (view at 12 Oct. 2016).

The original QGIS plugin QWaveTrasposition was developed Contini, P., De Girolamo, P., 1998. Impatto Morfologico di
to perform the geographic transposition of wave gauge data for Opere a Mare: Casi di Studio (Morphological Impact of
offshore wave hindcasting originally proposed by Contini and Structures at Sea: Case Studies). In: Atti VIII Convegno AIOM,
De Girolamo (1998). Associazione di Ingegneria Offshore e Marina /Proceedings of
8th Conference of the Offshore and Marine Engineering
QWaveTransposition is currently at a testing stage. However, a Association, Lerici, Italy, pp. 85–95.
general agreement was found between the model output of the
present example application and the results of a previous study Corsini, S., Inghilesi, R., Franco, L., Piscopia, R., 2006. Italian
(Pasanisi et al., 2015), in which the geographic transposition Wave Atlas. Agenzia per la Protezione dell’Ambiente e per i
was performed using a spreadsheet and different semi- Servizi Tecnici (APAT), Università degli studi di Roma Tre,
automated procedures. Roma, Italy.

Further testing applications to different sites are planned for DBCP, 2016. Data Buoy Cooperation Panel.
validation purposes. Based on the user experience, further http://www.jcommops.org/dbcp (view at 12 Oct. 2016).

39

Unauthenticated
Download Date | 3/1/17 12:16 PM
JOURNAL OF APPLIED ENGINEERING SCIENCES VOL. 6(19), ISSUE 2/2016
ISSN: 2247-3769 / e-ISSN: 2284-7197 ART.NO. 207, pp. 33-40

Donelan, M.A., 1980. Similarity Theory Applied to the Smith, J.M., 1991. Wind-Wave Generation on Restricted
Forecasting of Wave Heights, Periods, and Directions. In: Fetches. Miscellaneous Paper CERC-91-2, Waterways
Proceedings of the Canadian Coastal Conference, Burlington, Experiment Station, Coastal Engineering Research Center, U.S.
ON, Canada, pp. 47–61. Army Corps of Engineers, Vicksburg, MS, USA.

ENEA, 2016. Observations and modeling of climate change at SPM, 1984. Shore Protection Manual. Waterways Experiment
SSPT of ENEA. http://utmea.enea.it/labs/CLIM (view at 12 Station, Coastal Engineering Research Center, U.S. Army
Oct. 2016). Corps of Engineers, Vicksburg, MS, USA.

Holthuijsen, L.H., Booij, N., Herbers, T.H.C., 1989. A Sverdrup, H.U., Munk, W.H., 1947. Wind, Sea and Swell:
Prediction Model for Stationary, Short-Crested Waves in Theory of Relations for Forecasting. H. O. Pub. 601, U.S. Navy
Shallow Waters with Ambient Currents. Coastal Engineering, Department, Hydrographic Office, Washington, DC, USA.
13(1), pp. 23–54.
WMO, 1998. Guide to Wave Analysis and Forecasting, 2nd ed.,
ISPRA, 2016. Hydro-Meteo-Marine Forecasting System. Vol. WMO-No. 702, World Meteorological Organization,
http://www.isprambiente.gov.it/pre_meteo_eng/simm_eng.html Geneva, Switzerland.
(view at 12 Oct. 2016).

Komen, G.J., Cavaleri, L., Donelan, M., Hasselmann, K.,

Hasselmann, S., Janssen, P.A.E.M., 1994. Dynamics and
Modelling of Ocean Waves. Cambridge University Press, New
York, USA.

Leenknecht, D.A., Szuwalski, A., Sherlock, A.R., 1992.

Automated Coastal Engineering System Technical Reference.
Coastal Engineering Research Center, U.S. Army Corps of
Engineers, Vicksburg, MS, USA.

Liberti, L., Carillo, A., Sannino, G., 2013. Wave Energy

Resource Assessment in the Mediterranean, the Italian
Perspective. Renewable Energy, 50, pp. 938–949.

Liu, P.C., Schwab, D.J., Jensen, R.E., 2002. Has Wind–wave

Modeling Reached Its Limit? Ocean Engineering, 29, pp. 81–
98.

NumPy, 2016. http://www.numpy.org (view at 12 Oct. 2016).

Pasanisi, F., Tebano, C., Grauso, S., 2015. Coastal Morphology

and Dynamics of Two Beaches of Favignana. Energia,
Ambiente e Innovazione/Energy, Environment, and Innovation,
4/2015, pp. 48–59.

QGIS Development Team, 2016. QGIS Geographic Information

System. Open Source Geospatial Foundation Project.
http://www.qgis.org (view at 12 Oct. 2016).

Rogers, W.E., Dykes, J.D., Wittmann, P.A., 2014. US Navy

Global and Regional Wave Modeling. Oceanography, 27(3),
pp. 56–67.

Saville, T.J., 1954. The Effect of Fetch Width on Wave

Generation. Technical Memorandum 70, Beach Erosion Board,
U.S. Army Corps of Engineers, Washington, DC, USA, 1954.

Seymour, R.J., 1977. Estimating Wave Generation on

Restricted Fetches. Journal of the Waterway, Port, Coastal and
Ocean Division, 103, pp. 251–264.

Seymour, R.J., Sessions, M.H., 1976. Regional Network for

Coastal Engineering Data. In: Proceedings of 15th International
Conference on Coastal Engineering, Honolulu, HI, USA, pp.
60–71.

40

Unauthenticated
Download Date | 3/1/17 12:16 PM