ArticlePDF Available

Abstract and Figures

High-frequency (HF) coastal radars measure current velocity at the ocean surface with a 30–100 km range and 1–3 km resolution, every 0.25–1 h. HF radars are well suited to many applications, such as search and rescue (SaR), oil-spill mitigation and ecosystem management. Here we present a first organized core of 12 HF radars installed in five sites in four countries (Greece, Italy, France and Spain) within the European MED project, the Tracking Oil Spill and Coastal Awareness (TOSCA) network. Dedicated experiments tested radar capabilities to estimate transport driven by currents, which is the key feature for all the above applications. Experiments involved the deployment of drifters, i.e., floating buoys, acting as proxies for substances passively advected by currents. Using HF radars the search range is reduced by a factor of 1.6 to 5.3 after 24 h. The paper also underlines the importance of sharing common tools for HF radar data processing and the need to mitigate radio frequency interference. The effort can be regarded as an initial step toward the creation of a Mediterranean or European HF radar network, crucial for any European integrated ocean observing system (IOOS).
Content may be subject to copyright.
Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=tjoo20
Download by: [Consiglio Nazionale delle Ricerche] Date: 30 November 2015, At: 02:09
Journal of Operational Oceanography
ISSN: 1755-876X (Print) 1755-8778 (Online) Journal homepage: http://www.tandfonline.com/loi/tjoo20
Toward an integrated HF radar network in the
Mediterranean Sea to improve search and
rescue and oil spill response: the TOSCA project
experience
L. Bellomo, A. Griffa, S. Cosoli, P. Falco, R. Gerin, I. Iermano, A. Kalampokis,
Z. Kokkini, A. Lana, M.G. Magaldi, I. Mamoutos, C. Mantovani, J. Marmain, E.
Potiris, J.M. Sayol, Y. Barbin, M. Berta, M. Borghini, A. Bussani, L. Corgnati,
Q. Dagneaux, J. Gaggelli, P. Guterman, D. Mallarino, A. Mazzoldi, A. Molcard,
A. Orfila, P.-M. Poulain, C. Quentin, J. Tintoré, M. Uttieri, A. Vetrano, E.
Zambianchi & V. Zervakis
To cite this article: L. Bellomo, A. Griffa, S. Cosoli, P. Falco, R. Gerin, I. Iermano, A. Kalampokis,
Z. Kokkini, A. Lana, M.G. Magaldi, I. Mamoutos, C. Mantovani, J. Marmain, E. Potiris, J.M. Sayol,
Y. Barbin, M. Berta, M. Borghini, A. Bussani, L. Corgnati, Q. Dagneaux, J. Gaggelli, P. Guterman,
D. Mallarino, A. Mazzoldi, A. Molcard, A. Orfila, P.-M. Poulain, C. Quentin, J. Tintoré, M. Uttieri,
A. Vetrano, E. Zambianchi & V. Zervakis (2015): Toward an integrated HF radar network in the
Mediterranean Sea to improve search and rescue and oil spill response: the TOSCA project
experience, Journal of Operational Oceanography, DOI: 10.1080/1755876X.2015.1087184
To link to this article: http://dx.doi.org/10.1080/1755876X.2015.1087184
Published online: 25 Nov 2015. Submit your article to this journal
View related articles View Crossmark data
Toward an integrated HF radar network in the Mediterranean Sea to improve search and rescue
and oil spill response: the TOSCA project experience
L. Bellomo
a,b
*, A. Griffa
c,d
, S. Cosoli
e
, P. Falco
f
, R. Gerin
e
, I. Iermano
f
, A. Kalampokis
f,g
, Z. Kokkini
g
, A. Lana
h
,M.
G. Magaldi
c,i
, I. Mamoutos
g
, C. Mantovani
c
, J. Marmain
a,b
, E. Potiris
g
, J.M. Sayol
h
, Y. Barbin
a,b
, M. Berta
c
, M. Borghini
c
,
A. Bussani
e
, L. Corgnati
c
, Q. Dagneaux
a,b,j
, J. Gaggelli
a,b
, P. Guterman
k
, D. Mallarino
l
, A. Mazzoldi
e
, A. Molcard
a,b
,
A. Orla
h
, P.-M. Poulain
e
, C. Quentin
m
, J. Tintoré
h
, M. Uttieri
f
, A. Vetrano
c
, E. Zambianchi
f
and V. Zervakis
g
a
University of Toulon, National Center for Scientic Research (CNRS), Research and Development Institute (IRD), Mediterranean Institute
of Oceanography (MIO), La Garde, France;
b
Aix-Marseille University, CNRS, IRD, MIO, Marseille, France;
c
Institute of Marine Sciences,
National Council of Italy (CNR-ISMAR), La Spezia, Italy;
d
Rosenstiel School of Marine and Atmospheric Science (RSMAS), Ocean
Sciences (OCE) Department, University of Miami, FL, USA;
e
Istituto Nazionale di Oceanograa e Geosica Sperimentale (OGS), Trieste,
Italy;
f
Dipartimento di Scienze e Tecnologie (DiST), Università degli Studi di Napoli Parthenopeand Consorzio Nazionale
Interuniversitario per le Scienze del Mare (CoNISMa), Napoli, Italy;
g
Department of Marine Sciences, University of the Aegean, Mytilene,
Greece;
h
Mediterranean Institute for Advanced Studies (IMEDEA), Esporles, Balearic Islands, Spain;
i
Earth and Planetary Sciences,
Johns Hopkins University, Baltimore, MD, USA;
j
Collecte Localisation Satellites (CLS), Toulouse, France;
k
Technical Division of the
National Institute for Earth Sciences and Astronomy (INSU), La Seyne-sur-mer, France;
l
Aix-Marseille University, CNRS, IRD, OSU
Pythéas, Marseille, France;
m
Aix-Marseille University, CNRS, University of Toulon, IRD, MIO, Marseille, France
High-frequency (HF) coastal radars measure current velocity at the ocean surface with a 30100 km range and 13km
resolution, every 0.251 h. HF radars are well suited to many applications, such as search and rescue (SaR), oil-spill
mitigation and ecosystem management. Here we present a rst organized core of 12 HF radars installed in ve sites in
four countries (Greece, Italy, France and Spain) within the European MED project, the Tracking Oil Spill and Coastal
Awareness (TOSCA) network. Dedicated experiments tested radar capabilities to estimate transport driven by currents,
which is the key feature for all the above applications. Experiments involved the deployment of drifters, i.e., oating
buoys, acting as proxies for substances passively advected by currents. Using HF radars the search range is reduced by a
factor of 1.6 to 5.3 after 24 h. The paper also underlines the importance of sharing common tools for HF radar data
processing and the need to mitigate radio frequency interference. The effort can be regarded as an initial step toward the
creation of a Mediterranean or European HF radar network, crucial for any European integrated ocean observing system
(IOOS).
Introduction
In the last two decades, the importance of integrated ocean
observing systems (IOOSs) has been widely recognized,
not only for scientic purposes but also in order to
support societal needs such as the management of coastal
and marine environments, mitigation of accidents at sea
and planning for climate changes (Siddorn et al. 2007;
Weisberg et al. 2009; Tintoré 2013; Sayol et al. 2014).
IOOSs are typically composed of monitoring instruments
providing observations, numerical models assimilating
data and software infrastructures that guarantee real-time
access to the products. Various types of monitoring instru-
ments are used, with different sensors and platforms,
ranging from remote satellites to in situ instruments on
xed or moving locations (Glenn et al. 2000; Seim et al.
2009; Schiller & Brassington 2011).
In particular, high-frequency (HF) radars (Stewart &
Joy 1974; Barrick & Evans 1977) are crucial in coastal
areas for those applications related to transport by ocean
currents, such as monitoring and predicting the spread of
pollutants and biological quantities (Zelenke et al. 2009)
and search and rescue (SaR) operations (Barrick et al.
2012; Breivik et al. 2012). Lagrangian instruments and
HF radars are in many aspects complementary. Among
the former, drifters (Davis 1985) follow the current with
good approximation, providing direct information on hori-
zontal transport with small errors, typically within 13cm/s
for current velocity (Poulain et al. 2002). Their drawback is
that their coverage in a given region is often limited and
dependent on dedicated releases. HF radars, on the other
hand, provide autonomously continuous information in
terms of two-dimensional surface velocity maps with a
typical range of 30100 km from the coast, a spatial resol-
ution of 13 km, and an integration time of 0.251h
(Paduan & Washburn 2013). Although the combination
of these values is unique among oceanographic instruments
© 2015 Institute of Marine Engineering, Science & Technology
*Corresponding author. Email: bellomo@univ-tln.fr
Journal of Operational Oceanography, 2015
http://dx.doi.org/10.1080/1755876X.2015.1087184
Downloaded by [Consiglio Nazionale delle Ricerche] at 02:09 30 November 2015
and well suited to the aforementioned applications, the
unresolved spatial variability of velocity elds at subgrid
scales may still be signicant, so that the joint use of drif-
ters and HF radars appears promising and advantageous
(Berta et al. 2014).
Integrated HF radar observatories providing real-time
information with unied quality control have been operating
in the United States (US) as part of the US-IOOS (Harlan
et al. 2010;http://www.ioos.noaa.gov/hfradar/)andinAustra-
lia through the dedicated Australian Coastal Ocean Radar
Network (ACORN) (Heron et al. 2008;ACORN), and they
provide important support for agencies in charge of SaR
and pollution mitigation (Harlan et al. 2011). Following the
same approach, the First Ocean Radar Conference for Asia
(ORCA) has recently conducted a census of all Asian HF
radar installations (Fujii et al. 2013).
In Europe, a large number of individual HF radars is
present, providing relevant scientic and technological
advancements as well as practical applications (Gurgel
et al. 2011; Uttieri et al. 2011;Cosolietal.2012; Guihou
et al. 2013; Sentchev et al. 2013; Berta et al. 2014;
Marmain et al. 2014;Orla et al. 2014). Nevertheless,
although some countries have started to devote a signicant
effort toward the implementation of national HF radar net-
works (Quentin et al. 2013; Carrara et al. 2014), a unied
HF coastal radar network has not yet been implemented
nor have the parameters observed and derived from HF
radars been part of the main European IOOS projects such
as Copernicus and MyOcean 2 (Pinardi & Coppini 2010).
This issue has recently been addressed in the framework of
EuroGOOS (EuroGOOS) through an ongoing initiative
aimed at providing an inventory of existing HF radar
systems and starting to organize a coordinated European
HF radar group (EuroGOOS Conference 2014). The coordi-
nation will most likely be organized in terms of a distributed
system built on EuroGOOS regional alliances (ROOSs), pro-
viding a research and operational framework to develop and
deliver observations and products, similarly to what has
already been done for satellite and in situ data.
In this paper, we present the rst organized core of an
HF radar network in the Mediterranean Sea, set up in the
framework of the European MED project, the Tracking
Oil Spill and Coastal Awareness (TOSCA) network,
20102013 (TOSCA). TOSCA has focused on the identi-
cation of good practices based on science knowledge for
the mitigation of accidents at sea, such as oil spills or
SaR incidents, and one of its main achievements has been
to experimentally test observational systems based on
coastal HF radars and drifters, coupled with numerical
models.
The HF radar network consists of ve installation sites
with coverage within the main Mediterranean sub-basins,
i.e., the Aegean, Adriatic, Tyrrhenian, Ligurian and Balea-
ric Seas (Figure 1). Although each site was operated inde-
pendently from the others during TOSCA, common
practices have been established and dedicated experiments
have been carried out in the ve locations using a similar
methodology, involving the combined use of HF radar
and drifters. The experiments had a number of goals,
including (1) the validation and comparison of drifter and
HF radar measurements, (2) the investigation of optimal
blending techniques of the two datasets (Berta et al.
2014), (3) the validation of numerical circulation models
(Cosoli et al. 2013), (4) the estimation of statistical dis-
persion (Berta et al. 2014) and ow properties (Bellomo
et al. 2013) and (5) the assessment of the water- and oil-fol-
lowing capabilities of different types of surface drifters.
Here, we present the results of the TOSCA HF radar
network and focus on the rst objective. Comparisons are
performed in two distinct ways. First, the so-called radial
velocities of the HF radars, that is, the projection of
current velocities along the line of sight of each radar
station, are compared with the velocities measured by drif-
ters projected along the same direction. As radial velocities
are the ones actually sensed by HF radars, this can be
regarded as the rawest possible comparison, so that its
interest lies in the direct validation of HF radar measure-
ments and in the evaluation of their precision at distinct
locations and/or dates. Then, drifter trajectories are com-
pared with the synthetic ones obtained by numerically
advecting point-like particles within the HF radar vectorial
current eld, built in turn by aggregating two or more over-
lapping radial velocity elds. This approach is well suited
to practical applications such as oil-spill monitoring, SaR
and connectivity, where trajectories are needed to estimate
transport by surface currents. Therefore, since drifters can
be considered as proxies for passive particles advected by
currents, the comparison of trajectories provides a straight-
forward measure of the benets that HF radars would offer
in the aforementioned applications.
From a geographical and dynamical point of view the
ve sites are very different, as are the HF radar technologies
and signal processing techniques that have been used; thus,
we do not expect the results to necessarily be similar. Rather,
the purpose of the comparison is to investigate on variability,
practices and possible existing problems and solutions. In
this light, the results should provide an important scientic
basis for the ongoing process of building an integrated HF
radar system in the Mediterranean regional alliance.
The TOSCA sites, experimental set-ups and methods
are introduced, followed by the results. Finally a brief dis-
cussion is presented.
Data and methods
The TOSCA HF radar sites
HF radar systems transmit signals typically in the range of
342 MHz and rely on the backscattering of electromag-
netic waves from resonant sea surface waves. Radial
2L. Bellomo et al.
Downloaded by [Consiglio Nazionale delle Ricerche] at 02:09 30 November 2015
current velocities are retrieved over a distance/azimuth grid
from the analysis of the Bragg peaks in the Doppler spec-
trum of the back-scattered signal (Crombie 1955; Stewart
& Joy 1974; Barrick 1978). Radial velocities are combined
geometrically from at least two HF radar stations to
produce vectorial maps of surface current (Lipa &
Barrick 1983; Kohut et al. 2012). Such recombination
comes with geometric errors, indicated as geometric
dilution of precision (GDOP), which arise from deviations
from the perpendicular of the relative angles between radial
components (Chapman et al. 1997).
The locations of the TOSCA HF radar sites are shown
in Figure 1, together with examples of instantaneous vector
velocity elds from the HF radars during the TOSCA
experiments. These elds are unaveraged snapshots
chosen for their representativeness of coverage and local
circulation during the experiments. All the sites are
environmentally relevant and located in sensitive areas
affected by intense ship trafc and/or the presence of oil
pipelines. Their oceanographic characteristics, though, are
very different. Two of them are situated in gulfs, Trieste
and Naples, located in the Adriatic and Tyrrhenian Seas,
respectively. They are characterized by an average cyclonic
circulation and by a very high variability in time and space
induced by wind forcing and submesoscale dynamics
(Uttieri et al. 2011; Cosoli et al. 2013; Cianelli et al.
2015). The Ligurian Sea site is located in Toulon and
covers the Northern Current (Millot 1999) that ows cyclo-
nically along the coast with signicant mesoscale variabil-
ity (Bellomo 2013; Guihou et al. 2013). The Aegean site,
situated on the East coast of Lemnos, is characterized by
a prominent westward surface ow from the Dardanelles,
modulated by a seasonal cycle (Zervakis & Georgopoulos
2002). At both the Ligurian and the Aegean sites, the
Figure 1. Location of the TOSCA HF radar installation sites with a snapshot of current velocity eld measured during the TOSCA exper-
iments.
Note: Locations of HF radars are indicated using the following symbols: = CODAR SeaSondes; = quasi-monostatic WERAs; × and
+= WERA transmitters and receivers in bistatic conguration, respectively.
Journal of Operational Oceanography 3
Downloaded by [Consiglio Nazionale delle Ricerche] at 02:09 30 November 2015
patterns observed during the experiment are indeed well
representative of the climatological ow (Figure 1).
Finally, the Balearic site covers the Ibiza Channel, which
is a critical point for the exchange of oldMediterranean
and newAtlantic waters and is marked by a high mesos-
cale variability impacted by atmospheric forcing (Pinot
et al. 2002; Sayol et al. 2013).
The characteristics of the HF radar installations are
summarized in Table 1. Three of them Trieste, Naples
and Balearic use CODAR SeaSonde systems (Barrick
2008), whereas WERA technology (Gurgel et al. 1999)is
employed for the Toulon and Lemnos sites . While the Sea-
Sonde systems all consist of one emitting and three receiv-
ing, all co-localized antennas (Figure 2(a)), WERA is based
on antenna arrays namely, in Lemnos the radar stations
comprise one transmitting and one receiving square
arrays with four antennas each (Figure 2(b)), whereas in
Toulon either one or two transmitting and eight receiving
antennas are used. The minimum separation distance
between transmitting and receiving arrays is set to 300 m.
The radar operating frequency and bandwidth are chosen
so as to obtain higher spatial resolution with reduced cover-
age in the Gulfs of Naples and Trieste, which are dominated
by small-scale dynamics and have dimensions of 30 km,
and less resolved but wider elds in the other sites. Current
velocity maps are produced every 2060 min after aver-
aging signals over a similar duration.
The measured antenna pattern (Kohut & Glenn 2003)is
used to compute radial velocities for all the sites except for
Lemnos and Trieste, where the measurements have not yet
been performed. Therefore, an additional error source has
to be accounted for in these cases. Nonetheless, in Lemnos
the ideal patterns are not expected to be greatly different
from the actual ones, since both radars occupy empty areas
without urbanization and metallic environments.
In all sites, direction-nding algorithms are used for the
azimuthal analysis. An adapted version of the MUltiple
SIgnal Classication (MUSIC) algorithm (Schmidt 1989;
Lipa et al. 2006; Abramovich et al. 2009) is employed in
all the SeaSonde systems as well as in the WERA system
in Toulon, whereas the least squares single-source
method (Lipa & Barrick 1983) is used in the Lemnos
WERA system. Vectorial velocity computations from
radials are all performed over a regular grid using local
interpolation schemes based on least squares minimization
(Lipa & Barrick 1983). Although the grid step size should
scale with the radar range resolution, its value is rather arbi-
trary due to the coarser resolution of the radial grids as the
distance from the radar increases. Values ranging from 13
km are commonly accepted in the HF radar community
(Paduan & Washburn 2013) and have been retained in
the TOSCA sites. Finally, the GDOP is accounted for by
discarding grid points with values >2.5.
Radial and vectorial velocities are quality controlled
using thresholding methods based on rst- and second-
order nite differences in time and space (Kovacevic
et al. 2004). Besides, the signal-to-noise ratio (SNR) of
individual Doppler lines and the energy of the MUSIC
algorithm sources are also taken into account in Trieste
and Toulon, respectively.
Finally, data gaps are partially lled through linear
interpolation in space and time. However, for the Lemnos
and Balearic sites the vectorial current elds are interp-
olated through the Data INterpolating Empirical Orthog-
onal Functions (DINEOF) method (Beckers & Rixen
2003) that produces a uniform and constant coverage area
based on an EOF analysis of the gappy eld. This will be
discussed further in the next sections.
Drifter deployments
In the framework of TOSCA, a set of experiments with
similar methodologies have been performed within the
HF radar observation areas from winter 2011 to autumn
2012 (Figure 3). Although two experiments each were con-
ducted in Toulon, Lemnos and Trieste, given the qualitat-
ively similar results only one has been selected for this
work (Table 1).
Various types of drifters have been launched during the
experiments, to investigate surface currents and wind
effects. These include CODE-type drifters (Davis 1985;
Poulain 1999), designed to follow currents from surface to
a depth of 1 m, and oil-spill drifters(Zervakis et al.
2009), conceived to follow pollutants at the sea surface.
Table 1. Characteristics of the TOSCA sites in terms of HF radar installations and drifter deployments.
Site name Installation set-up
Operating frequency
(MHz) Range resolution (km) Experiment date (2012) Drifters (res. time, days)
Trieste 3 SeaSonde 25.0 1.0 April 37 CODE (2)
Naples 3 SeaSonde 25.0 1.0 August 22 CODE (10)
Toulon 2 WERA 16.1 3.0 August 23 CODE (1)
Lemnos 2 WERA 13.5 3.0 October 8 CODE (1)
Balearic 2 SeaSonde 13.5 1.7 October 4 oil spill (1)
Note: The residence time indicated for each site corresponds to the average number of days spent by drifters within the HF radar coverage during the
experiments.
4L. Bellomo et al.
Downloaded by [Consiglio Nazionale delle Ricerche] at 02:09 30 November 2015
Here we mostly concentrate on CODE drifters, since they are
more suitable for HF radar comparison, as discussed in the
following section. They are drogued in the rst metre
below the surface and are designed to minimize slippage
due to the direct action of wind and waves, whose impact
has been quantied to 13cm/s(Poulainetal. 2002).
CODE drifters have been launched and used in all the
sites, except for the Balearic one. There, only oil-spill drif-
ters were available, but some of them have been ballasted
with an additional 2 kg in order to reduce windage. Since
only the latter have been retained in this study and since
they had a low residence time within the radar coverage
area (Table 1), the number of drifter points used for com-
parison is very limited. Therefore, any statistical inference
presented in the results has to be considered with care for
this site.
In some cases drifters were caught and relaunched
during the experiments in order to maintain coverage of
the HF radar area. All drifters were equipped with GPS recei-
vers with an accuracy of approximately 510 m. Drifter pos-
itions were edited to remove outliers and spikes (Poulain
et al. 2004;Gerin&Bussani2011) and interpolated at
uniform 1 h intervals (Hansen & Poulain 1996), and vel-
ocities along trajectories were computed from the positions
by central nite differences. Considering the aforementioned
GPS accuracy for a duration comparable to the HF radar
integration time, the drifter velocity accuracy is expected
to be below 1 cm/s.
The experiment period, the number and kind of drifters
used and the average time spent by drifters in the HF radar
coverage are shown in Table 1 for all sites. The atmospheric
conditions during the experiments were rather calm (wind
Figure 2. Examples of HF radar systems belonging to the TOSCA network: (a) a CODAR SeaSonde installation in Formentera (Balearic
site; the TX/RX monopole is installed above a closed dome containing two crossed-loop magnetic RX antennas); and (b) a WERA transmit-
ting array made up of four monopoles in Plaka (Lemnos site). Source: (a) L. Bellomo, (b) Z. Kokkini.
Figure 3. Timeline of the TOSCA experiments.
Journal of Operational Oceanography 5
Downloaded by [Consiglio Nazionale delle Ricerche] at 02:09 30 November 2015
velocities smaller than 7 m/s) in Lemnos, Trieste and
Naples, with a typical summer breeze regime in the case
of Naples. In Toulon, conditions were also calm, except
for a north-westerly wind episode with velocities up to
15 m/s that lasted for 1.5 days. Finally, the Balearic exper-
iment was performed under very energetic atmospheric
conditions, with intense winds (velocities larger than 20
m/s) and high waves (signicant heights larger than 3 m)
during two days.
Considerations on HF radar and drifter velocity
comparison
When comparing HF radar and drifter-based velocities, it is
important to keep in mind the differences between the
two types of instrument and sampling methods. While
drifters feel the velocity at a scale corresponding to their
physical horizontal and vertical size, of the order of 1 m
for the CODE type used here, HF radar velocities are quan-
tities averaged over very different vertical and horizontal
scales.
In the vertical, the HF radar velocity is the exponentially-
weighted average of the upper ocean velocity prole, and it
depends on the vertical shear of the horizontal current and on
the HF radar frequency (Stewart & Joy 1974;Ha1979;
Shrira et al. 2001;Ivoninetal.2004). In the case of a
linear shear, in particular, the measurement corresponds to
an effective depth in the range between 50 cm for a
working frequency of 25 MHz and 90 cm for 13.5 MHz.
Consequently, the comparison with CODE-type drif-
ters, which provide the vertical average of the velocity in
the upper 1 m, is expected to be appropriate, unless very
strong shears are present in the upper water column that
can affect the two averages in a different way.
In the horizontal, on the other hand, there is clearly a
mismatch of scales. Indeed, since the HF radar velocity is
an average over the two-dimensional grid cell, the size of
which is of the order of 13 km for the TOSCA sites
(Table 1), submesoscale variability associated with, for
example, high horizontal shears is hard to resolve, as the
Rossby deformation radius in the regions considered is
10 km (Grilli & Pinardi 1998). Therefore, the comparison
between HF radar and drifters can only be considered sat-
isfactory when it falls within the range of expected variabil-
ity within the horizontal grid (Ohlmann et al. 2007). For our
sites, specic estimates of such variability, which depend
on both ow characteristics and grid size, are not available.
As a general guidance, results from the literature suggest
that discrepancies of the order of 515 cm/s can be con-
sidered acceptable and within the bounds of expected varia-
bility (Paduan & Rosenfeld 1996; Chapman et al. 1997;
Shay, Lee et al. 1998, Shay, Lentz et al. 1998; Essen
et al. 2000; Shay et al. 2001,2007; Emery et al. 2004;
Kaplan et al. 2005; Paduan et al. 2006; Ohlmann et al.
2007; Molcard et al. 2009; Rypina et al. 2014).
Results
HF radar and drifter velocities have been compared for all
the sites, considering drifters within the HF radar coverage.
Such comparisons are performed both on radial velocities
and on the trajectories computed from vectorial velocities,
and they are presented separately in the following.
Comparison between radial velocities
Radial velocities from HF radars (uR
r) and those computed
from drifter data (uR
d) are compared at the same time and
locations. Drifter data are resampled on the uniform radar
time grid, and for each drifter position the radar velocities
corresponding to the nearest cells are bilinearly interpolated
to obtain the radar velocity estimate. The difference between
the two estimates of radial velocities, DuR=uR
ruR
d,is
then calculated, and its statistics are quantied in terms of
bias, bD=DuR, and root mean square (RMS),
rms2
D= (DuR)2, where overbars stand for averages over
all drifter positions and times. For comparison, the RMS
value of the drifter velocity is also computed as
rms2
u= (uR
d)2.
Notice that this methodology differs slightly with
respect to that of, for example, Ohlmann et al. (2007),
Molcard et al. (2009) and Rypina et al. (2014), as in
those cases HF radar velocities are compared with the
average of all drifter velocities within each radar grid
cell. To verify our choice, sensitivity tests performed for
the Naples and Toulon sites, where the two approaches
are compared, have shown that a slightly better matching
is obtained with the methodology proposed here.
Results are summarized in Table 2 in terms of rmsD,u
and bDaveraged among all the radar stations of each site.
Furthermore, an example of drifter trajectory map with
the associated DuRis shown in Figure 4 for one of the
stations belonging to the Naples system. For the three
sites of Trieste, Naples and Toulon, rmsDis within 510
cm/s, i.e., well within the range of values expected from
the literature, and it is smaller than the rmsuof the drifters
by a factor of 3.1 for Toulon, 2 for Naples and 1.3 for
Trieste. Biases range from 0.40.5 cm/s for Naples and
Table 2. Results of the comparison between radial velocities.
Site name rmsu(cm/s) rmsD(cm/s) bD(cm/s)
Trieste 12.9 9.6 2.1
Naples 9.6 4.7 0.4
Toulon 15.5 5.0 0.5
Lemnos 24.0 ––
Balearic 22.0 16.0 5.5
Note: Values are computed for each site as averages over all the HF radar
stations; rmsuis computed from radial drifter velocities, whereas rmsD
and bDare computed from differences between HF radar and drifter
radial velocities.
6L. Bellomo et al.
Downloaded by [Consiglio Nazionale delle Ricerche] at 02:09 30 November 2015
Toulon to 2.1 cm/s for Trieste. The relatively higher rmsD
and bDfor Trieste might be due to at least two reasons.
The rst is the use of the measured antenna pattern for
only one out of the three stations, conrmed by the fact
that for this one station, rmsDimproves from 19.2 cm/s to
9.6 cm/s when computed with ideal and measured
pattern, respectively. On the other hand, bDis hardly
changed by the measured pattern, corroborating the
second possible cause of error, that is, the presence of a
low-salinity surface layer of about 1 m at the time of the
experiment, originating at the Isonzo river outow and
likely characterized by signicant vertical shear (Malacic
et al. 2006). As mentioned in the previous section, the
shear could indeed affect in a different way the effective
depth of HF radar and CODE drifters, increasing the differ-
ences between the two velocity estimates.
The values of rmsDand bDfor the Balearic site are
higher, reaching 16 and 5.5 cm/s, respectively. As demon-
strated in the next section, this is probably related to the oil-
spill drifters used for the comparison, which are not CODE
design and are consequently more directly inuenced by
the wind, which was very rough during the experiment.
Such atmospheric conditions might well explain the high
rmsuvalue, conrming the fact that the ow is more ener-
getic than at the other sites (Table 2).
Finally, at the Lemnos site very strong radio-frequency
interference (RFI) occurred in the frequency spectrum (at
the time of the experiment, voices speaking an eastern
Asiatic language could clearly be heard with an amateur
AM radio receiver within the HF radar band!), character-
ized by a daily cycle with strong perturbations during the
daytime and not during the night. The effect of the RFI is
both reduced coverage, clearly visible in Figure 5, and
the presence of outliers in the computed velocity eld.
Because of this, and considering the relatively small
number of drifters, the comparison results are not signi-
cant and are therefore not reported in Table 2.
Comparison between drifter trajectories and synthetic
HF radar-based trajectories
For each drifter trajectory within the HF radar coverage,
synthetic trajectories have been computed using the vectorial
velocity eld from the HF radar. The numerical advection
was performed using a fourth-order Runge-Kutta integration
scheme. Synthetic trajectories were rst initialized at
the same launch time and location as the drifters, and then
re-initialized at constant time-intervals at the corresponding
drifter positions. The comparison between the synthetic
trajectories xr(t)and drifter trajectories xd(t)is quantied
computing their mean separation distance,
D(t)=k||xr(t)−xd(t)||l, as a function of time tfrom
deployment, where the mean is computed averaging over
all the trajectories at time tand || · || is the Euclidean norm.
The mean absolute distance,D0(t)=k||xd(t)−xd(0)||l,tra-
velled by the drifters is also computed. Indeed, in a number of
previous papers (e.g., Ullman et al. 2006; Molcard et al.
2009) it has been regarded as an estimate of the persistency
error, i.e., of the error that would be made in SaR situations
assuming that the target is not moving and there is no avail-
able information on ocean currents. When D(t)is smaller
than D0(t), this is an indication that the use of HF radar
data allows a reduction in the error of the estimated target
position and the associated search range. As a consequence,
we dene the search range reduction factor (SRRF) as
the ratio between D(t)and D0(t)at the end of the advection
time.
The results are shown in Figures 610 in terms of D(t)
and D0(t)for each site up to t= 24 h, which corresponds to
a typical correlation time scale in coastal regions (Ozgok-
men et al. 2000; Bauer et al. 2002), except for Lemnos
where due to insufcient data within the HF radar coverage
the value of twas reduced to 21 h. Furthermore, in order to
assess their variability, the standard deviations
s
D(t)and
s
D0(t)associated to D(t)and D0(t), respectively, are also
shown as shaded areas around the mean values. Table 3
Figure 4. Drifter trajectories for the Portici radar (square black
dot) of the Naples site.
Note: The colour associated with the positions of drifters is the
difference DuRbetween the drifter and the radar radial velocities.
Figure 5. HF radar radial coverage as a function of time for the
Plaka radar of the Lemnos site during the days of the experiment
(very similar results are obtained for the other radar of the site).
Note: The coverage is computed as the number of valid radial vel-
ocities normalized by its largest value. The gray-shaded areas cor-
respond to daytime (between sunrise and sunset). It can be clearly
observed how the coverage drops dramatically during daytime,
with almost zero coverage spikes due to extremely strong RFI.
Journal of Operational Oceanography 7
Downloaded by [Consiglio Nazionale delle Ricerche] at 02:09 30 November 2015
summarizes the nal dispersion values together with the
SRRF.
In the Gulfs of Trieste and Naples (Figures 6 and 7,
respectively), the drifter travel distance D0(t)reaches a
plateau after 1214 h, suggesting that the effects of bound-
aries and re-circulations become dominant after that time,
corresponding to scales of 5km.D(t)on the other hand
increases almost linearly while being smaller than D0(t)
at all times, especially for Naples. In the Toulon site
(Figure 8), D0(t)keeps increasing and reaches approxi-
mately 10 km after 24 h, while D(t)increases much more
slowly, only reaching 2 km. In addition,
s
D0(t)is the smal-
lest among all sites and practically does not increase with
time, as is the case in Trieste and Naples. This is probably
due to a smaller space-time variability of the current elds
in Toulon, whose covered area is mostly occupied by a
boundary current.
For all these sites, the difference between D(t)and
D0(t)indicates a reduction in the position error and in the
search range using HF radar data. Such a reduction is
especially evident for the Toulon site, where the SRRF is
5.3 after 24 h. For the Gulfs, the attening of D0(t)
makes the comparison with D(t)meaningful only in the
rst 1214 h, leading to an SRRF of 2.2 for Naples and
1.6 for Trieste.
For the Balearic site (Table 3), D0(t)increases almost
linearly in time, reaching almost 25 km after 24 h, in agree-
ment with the fact that the drifters moved under energetic
conditions with rmsu= 22 cm/s (Table 2). D(t)is
Figure 6. Comparison between drifter trajectories for the Trieste
site, mean absolute distance D0(t)covered by the drifters and
mean separation distance D(t)between drifter and radar-based tra-
jectories.
Note: The shaded regions correspond to the condence intervals at
± one standard deviation. The vertical dashed line at 12 h indicates
the time after which recirculation effects in the Gulf become
dominant.
Figure 7. Comparison between drifter trajectories for the Naples
site, mean absolute distance D0(t)covered by the drifters and
mean separation distance D(t)between drifter and radar-based tra-
jectories.
Note: The shaded regions correspond to the condence intervals at
± one standard deviation. The vertical dashed line at 14 h indicates
the time after which recirculation effects in the Gulf become
dominant.
Figure 8. Comparison between drifter trajectories for the Toulon
site, mean absolute distance D0(t)covered by the drifters and
mean separation distance D(t)between drifter and radar-based tra-
jectories.
Note: The shaded regions correspond to the condence intervals at
± one standard deviation.
Figure 9. Comparison between drifter trajectories for the Balea-
ric site, mean absolute distance D0(t)covered by the drifters and
mean separation distance D(t)between drifter and radar-based tra-
jectories.
Note: The shaded regions correspond to the condence intervals at
± one standard deviation.
8L. Bellomo et al.
Downloaded by [Consiglio Nazionale delle Ricerche] at 02:09 30 November 2015
signicantly lower, reaching approximately 12 km, giving
an SRRF equal to 2. Yet, the very limited number of
drifter points makes the statistics not completely reliable,
as also evidenced by unusually large values of
s
D(t)so
this result must be treated with caution. An interesting
point is to verify whether or not there is a signicant
wind inuence on the drifter motion, as suggested in pre-
vious sections, and keeping in mind that the drifters
deployed in this site are of the oil-spill type. A test was per-
formed correcting the trajectories generated through the HF
radar data by adding 2% of the wind from the HIgh Resol-
ution Local Area Model (HIRLAM) model (5 km and 3 h
spatial and temporal resolution, respectively) (Unden et al.
2002). The results show a signicant reduction of D(t),
which reaches only 6 km at 24 h, conrming that wind cor-
rection indeed reduces the error. An example of particle tra-
jectories highlighting the impact of the wind correction is
shown in Figure 11.
As already mentioned, the Lemnos site is characterized
by signicant daytime gaps due to RFI that are lled by
DINEOF. The results in Figure 10, cut at 21 h since no
drifter stayed in the covered area for more than this
duration, show that D0(t)increases up to almost 18 km
while D(t)maintains smaller values, reaching approxi-
mately 10 km and resulting in an SRRF equal to 1.7.
This result suggests that even in the presence of signicant
errors or gaps in radial velocities, trajectory estimates can
lead to an improvement in particle position using the appro-
priate velocity reconstruction techniques, provided the vel-
ocity eld is sufciently persistent in space and time to
allow for data gap lling. This is indeed the case in the
Lemnos site during the experiment, when the ow
showed the characteristic seasonal pattern owing from
the Dardanelles to the Aegean, with small variability
between daytime and night-time.
Summary and concluding remarks
In this paper, the HF radar network operated during the
TOSCA project is presented and results from the compari-
son between HF radar and drifter velocities during dedi-
cated experiments are shown. In the three sites of Trieste,
Naples and Toulon, where the comparison was performed
using more than 20 CODE drifters drogued in the rst
metre under the sea surface, with average residence time
within the radar coverage area of 2, 10 and 5 days, respect-
ively, qualitatively similar results are found. In all cases, the
RMS value of the radial velocity difference between HF
radar and drifters (rmsD) falls in the range of 510 cm/s,
which is well within what is considered acceptable in the
literature given the expected variability at the HF radar
subgrid level. The average difference D(t)between drifter
trajectories and synthetic ones computed from HF radar
velocities is smaller than the mean distance travelled by
drifters D0(t)by a factor ranging from more than 5 for
Figure 10. Comparison between drifter trajectories for the
Lemnos site, mean absolute distance D0(t)covered by the drifters
and mean separation distance D(t)between drifter and radar-based
trajectories.
Note: The shaded regions correspond to the condence intervals at
± one standard deviation. The comparison is only shown up to 22
h since no drifter remained in the area covered by the HF radar for
a longer time.
Table 3. Results of the comparison between drifter trajectories.
Site
name
Advection time
(h)
D0(t)
(km) D(t)(km) SRRF
Trieste 12 5.2 3.3 1.6
Naples 14 3.3 1.5 2.2
Toulon 24 9.6 1.8 5.3
Lemnos 21 17.9 10.3 1.7
Balearic 24 24.4 12.2 (6.1) 2.0 (4.0)
Note: For the Balearic site, values within parentheses are obtained with a
2% wind correction.
Figure 11. Trajectory of one among the four drifters of the
Balearic site.
Note: The red dot indicates the deployment point; the two black
squares in the Ibiza and Formentera islands on the right represent
the location of radar stations.
Journal of Operational Oceanography 9
Downloaded by [Consiglio Nazionale delle Ricerche] at 02:09 30 November 2015
Toulon to 1.6 for Trieste. Referring to SaR or oil-spill appli-
cations using HF radar, we have interpreted these numbers
as SRRFs to stress the reduction on the position error with
respect to the persistency error quantied by D0(t). The
quantitative differences between the sites could be due to
several factors, including vertical shear in the surface
layer that was likely to be more pronounced in Trieste at
the time of the experiment. The Balearic site has a higher
rmsDof 15 cm/s, but it should be considered that during
the experiment, because of very strong winds (>20 m/s),
the ow was the most energetic plus the drifters used
for comparison are not CODE and are more directly
affected by winds. Nevertheless, the SRRF amounts to 2,
and the improvement raises to a factor 4 when numerical
trajectories are computed adding a direct wind inuence
in terms of correction. The Lemnos results are strongly
inuenced by the RFI that affects the radial velocities
during the daytime. Even in this case, though, when the
vector velocity eld is reconstructed using the DINEOF
technique, which exploits EOF information to ll the
gaps and smooth the eld, the trajectory results show a
reduced position error with an SRRF of 1.7. Notice
nally that the lowest SRRFs (Table 3) were obtained for
Trieste and Lemnos, where ideal antenna patterns were
used at least partially in the case of the former. Therefore,
a larger SRRF might be expected even in these cases, pro-
vided that measured antenna patterns were properly
employed.
In summary, despite the differences in oceanographic
sites, installations and data analysis, the results consistently
show that the use of HF radar data enables a decrease in the
position errors and search range by a factor varying
between 5.3 and 1.6. At the same time, the results point
to a number of important issues that should be considered
in future studies and applications.
First of all, the comparison between drifters and HF
radar appears to be inuenced by the details of vertical
shear and air-sea interaction, implying that transport in
the surface layer is very complex and can vary signicantly
with depth. Indeed, the nature of the response of the ocean
in the rst metre of water is still largely unknown, and these
results contribute to indicating the implications of this lack
of knowledge. While HF radar elds provide important
information on current velocity, they need to be comple-
mented by wind knowledge, current shear estimates and
buoyancy and drags of oating bodies in order to be used
in the most effective way in practical applications such as
SaR or pollution mitigation.
The other important issue highlighted by the TOSCA
experiments is the problem of radio interference and its
impact on HF radar performance. The presence of RFI is
indeed a well-known phenomenon in the HF band below
30 MHz (American Radio Relay League 2013), caused
by the remote propagation of electromagnetic waves
thanks to the excitation of the ionosphere by sun radiation,
which in turn justies its daily cycle. RFI has been
observed particularly in 13 and 16 MHz systems, including
Lemnos, Toulon, Balearic and the Calypso system installed
between Malta and Sicily (A. Drago, personal communi-
cation). As shown for the Lemnos case, this problem can
be alleviated a posteriori using dedicated signal processing
methods able to reconstruct the velocity eld by smoothing
and lling the gaps, but this is likely to be acceptable only
for ows with relatively low variability. On the other hand,
only a few exploratory works have tackled the problem of
suppressing RFI from the measured signals before comput-
ing velocity maps (Jun et al. 2004; Gurgel & Barbin 2008;
Chen et al. 2010; Wang & Wyatt 2011). Furthermore, given
the diverse nature of all possible interference (pure carrier
or modulated signal, transient or continuous, etc.), which
cannot be handled with a universal method, an effort at
the international HF radar community level should be put
forth in developing and sharing signal processing tech-
niques for RFI mitigation.
In parallel, the presence of RFI should be avoided by a
global coordination plan for the occupation of the fre-
quency spectrum. Such a plan can only marginally rely
on national or border policies, as electromagnetic waves
in the HF frequency band can easily propagate up to dis-
tances of thousands of kilometres thanks to the conducting
properties of the ionosphere (American Radio Relay
League 2013). In this respect, only very recently at the
World Radiocommunication Conference held in Geneva,
Switzerland in 2012 (WRC-12), the International Telecom-
munication Union (ITU) has ofcially allocated frequency
bands for the use of HF radars (ITU Radio Regulations
2012). Yet this allocation is not even exclusive or
primary, especially in Europe, and today such bands are
far from being emptied out by other ofcial and non-ofcial
radio services. Even more importantly, each of these bands
can host only one or two HF radars without frequency over-
lapping. Coordination is therefore mandatory, even among
HF radar manufacturers and users, in order to employ time
or modulation division duplexing or possibly orthogonal
codes to have many systems sharing the same frequency
band without interfering with each other.
Finally, the TOSCA project has allowed a useful
exchange of different HF radar practices, implementations
and data analyses. This has provided an interesting survey
and has led to general results in terms of comparison and
applications. On the other hand, the many differences
between sites did not allow a systematic comparison of
specic issues, e.g., quality control (QC) and RFI suppres-
sion. In order to achieve this, a specic experiment should
be chosen and the results treated and compared with differ-
ent methodologies (Kohut et al. 2012). An effort in this
direction is presently being carried out, and the TOSCA
results are available in a common format within a web-
based GIS (TOSCA GIS) and proposed as a common
open benchmark for testing.
10 L. Bellomo et al.
Downloaded by [Consiglio Nazionale delle Ricerche] at 02:09 30 November 2015
The TOSCA network presented here includes Mediter-
ranean installations that were active at the beginning of the
project in 2010. Today, other installations and HF radar
projects are active or in preparation (e.g., the aforemen-
tioned Calypso system or the SIstema di COntrollo
MARino (SICOMAR) installation in the Tuscan Archipe-
lago), and there is a clear need to include them and
proceed toward building an effective network for real-
time HF radar data with common standards in data
format, QC and signal processing.
Coherently with one of the main goals of the TOSCA
project, we conclude suggesting that the installation of HF
radars and the establishment of pollution mitigation facili-
ties equipped with CODE and oil-spill drifters should be a
requirement of obtaining permission to build new offshore
mining facilities. This claim is justied by the present
context of developments in the exploration and prospective
exploitation of deep-sea oil resources in the Mediterranean.
In this framework, besides conrming the effectiveness of
these instruments and methodologies, the study presented
herein documents the existence of a rst core of a European
HF radar alliance, and stresses the need for a unied network
that would be mandatory to obtain, among others, the afore-
mentioned objective.
Acknowledgements
The Italian partners acknowledge support from the agship
national project RITMARE and the Rete Nazionale Mareograca
for providing wind data. The Spanish partners acknowledge
support from SOCIB, HF radar, modelling, IT and engineering
facilities.
Disclosure statement
No potential conict of interest was reported by the authors.
Funding
This work was supported by the Tracking Oil Spills and Coastal
Awareness (TOSCA) Network project [grant number 2G-MED09-
425], co-nanced by the European Regional Development Fund in
the framework of the MED programme.
References
Abramovich Y, Johnson BA, Mestre X. 2009. Classical and
modern DOA estimation, chap. DOA estimation in the
small-sample threshold region, 219287. AP-Elsevier.
ACORN [Internet]. Available from http://imos.org.au/acorn.html.
American Radio Relay League (Ed). 2013. The 2014 ARRL
Handbook for Radio Communications, chap. Propagation of
Radio Signals.
Barrick DE. 1978. HF radio oceanography a review. Bound
Layer Meteorol. 13:2343.
Barrick DE. 2008. 30 years of CMTC and CODAR. Proc. IEEE/
OES/CMTC Ninth Working Conf. Curr. Meas. Technol.,
Charleston, SC (USA), 131136.
Barrick DE, Evans MW. 1977. Ocean surface currents mapped by
Radar. Science. 198:138144.
Barrick D, Fernandez V, Ferrer MI, Whelan C, Breivik Ø. 2012. A
short-term predictive system for surface currents from a
rapidly deployed coastal HF radar network. Ocean Dyn.
62:725740.
Bauer S, Swemson MS, Griffa A. 2002. Eddy-mean ow
decomposition and eddy diffusivity estimates in the tropical
Pacic Ocean. 2: results. J Geophys Res: Oceans. 107:
C103154.
Beckers JM, Rixen M. 2003. EOF calculations and data lling
from incomplete oceanographic datasets J Atmos Ocean
Technol. 20:18391856.
Bellomo L, et al. 2013. Observational evidence of mesoscale
variability of the northern current (north-western mediterra-
nean sea): a combined study via gliders, hf radar, surface drif-
ters, and vessel data. EGU General Assembly, Wien (Austria),
Vol. 15, EGU2013-5469-2.
Berta M, Bellomo L, Magaldi MG, Griffa A, Molcard A, Marmain
J, Borghini M, Taillandier V. 2014. Estimating Lagrangian
transport blending drifters with HF radar data and models:
results from the TOSCA experiment in the Ligurian Current
(North Western Mediterranean Sea). Prog Oceanog 128:1529.
Breivik Ø, Allen AA, Maisondieu C, Olagnon M. 2012. Advances
in search and rescue at sea. Ocean Dyn. 63:8388.
Carrara P, et al. 2014. The RITMARE coastal radar network and
applications to monitor marine transport infrastructures. EGU
General Assembly, Wien (Austria), Vol. 16, EGU2014-14 149-1.
Chapman RD, Shay LK, Graber HC, Edson JB, Karachintsev A,
Trump CL, Ross DB. 1997. On the accuracy of HF radar
surface current measurements: intercomparisons with ship-
based sensors. J Geophys Res. 102(18):737748.
Chen G, Zhao Z, Zhu G, Huang Y, Li T. 2010. HF radio-frequency
interference mitigation. IEEE Geoscience and Remote
Sensing Letters. 7:479482.
Cianelli D, Iermano I, Mozzillo P, Uttieri M, Zambardino G,
Buonocore B, Falco P, Zambianchi E. 2015. Inshore/offshore
water exchange in the Gulf of Naples. J Mar Sys. 145:3752.
Cosoli S, Gacic M, Mazzoldi A. 2012. Surface current variability
and wind inuence in the northeastern Adriatic Sea as
observed from high-frequency (HF) radar measurements.
Cont Shelf Res. 33:113. doi:10.1016/j.csr.2011.11.008
Cosoli S, Licer M, Vodopivec M, Malacic V. 2013. Surface circu-
lation in the Gulf of Trieste (northern Adriatic Sea) from
radar, model, and ADCP comparisons. J Geophys Res:
Oceans. 118:61836200.
Crombie DD. 1955. Doppler spectrum of Sea Echo at 13.56 Mc./s.
Nature. 175:681682.
Davis RE. 1985. Drifter observations of coastal surface currents
during CODE: the method and descriptive view. J Geophys
Res. 90:47414755.
Emery BM, Washburn L, Harlan JA. 2004. Evaluating radial
current measurements from CODAR high-frequency radars
with moored current meters. J Atmos Ocean Technol.
21:12591271.
Essen HH, Gurgel KW, Schlick T. 2000. On the accuracy of
current measurements by means of HF radar. IEEE
J Oceanic Eng. 25:472480.
EuroGOOS [Internet]. Available from http://eurogoos.eu/about-
eurogoos/overview/.
EuroGOOS conference 2014 [Internet]. Available from http://
eurogoos2014.hidrograco.pt/eurogoos-conference.php.
Fujii S, et al. 2013. An overview of developments and appli-
cations of oceanographic radar networks in Asia and
Oceania Countries. Ocean Sci J. 48:6997.
Journal of Operational Oceanography 11
Downloaded by [Consiglio Nazionale delle Ricerche] at 02:09 30 November 2015
Gerin R, Bussani A. 2011. Nuova procedura di editing automatico
dei dati drifter impiegata su oceano per MyOcean e prodotti
web in near-real time e delay mode. Tech Rep OGS 2011/
55 OGA 20 SIRE, OGS, Trieste (Italy).
Glenn SM, Dickey TD, Parker B, Boicourt W. 2000. Long term
real-time coastal ocean observation networks.
Oceanography. 13:2434.
Grilli F, Pinardi N. 1998. The computation of Rossby radii of
deformation for the MErditerranean Sea. MTP News. 6(4).
Guihou K, Marmain J, Ourmieres Y, Molcard A, Zakardjian B,
Forget P. 2013. A case study of the mesoscale dynamics in
the North-Western Mediterranean Sea: a combined data-
model approach. Ocean Dyn. 63:793808.
Gurgel K-W, Antonischki G, Essen H-H, Schlick T. 1999. Wellen
Radar (WERA): a new ground-wave HF radar for ocean
remote sensing. Coastal Engineering. 37:219234.
Gurgel K-W, Barbin Y. 2008. Suppressing radio frequency inter-
ference in HF radars. Sea Technology. 49:3942.
Gurgel K-W, Schlick T, Voulgaris G, Seemann J, Ziemer F. 2011.
HF radar observations in the German Bight: Measurements
and quality control. Current, Waves and Turbulence
Measurements (CWTM), 2011 IEEE/OES 10th, Monterey,
CA (USA), 5156.
Ha E-C. 1979. Remote sensing of ocean surface current and
current shear by HF backscatter radar. Tech Rep D415-1,
Stanford University.
Hansen DV, Poulain P-M. 1996. Quality control and interpola-
tions of WOCE-TOGA drifter data. J Atmos Ocean
Technol. 13:900909.
Harlan J, et al. 2011. National IOOS high frequency radar search
and rescue project. OCEANS 2011, Waikoloa, HI (USA).
Harlan J, Terrill E, Hazard L, Keen C, Barrick D, Whelan C,
Howden S, Kohut J. 2010. The integrated ocean observing
system high-frequency radar network: status and local,
regional, and national applications. Mar Technol Soc J.
44:122132.
Heron ML, Prytz A, Searson S. 2008. The Australian Coastal
Ocean Radar Network (ACORN). Current Measurement
Technology, 2008. CMTC 2008. IEEE/OES 9th Working
Conference on, Charlston, SC (USA).
Ivonin DV, Broche P, Devenon J-L, Shrira VI. 2004. Validation of
HF radar probing of the vertical shear of surface currents by
acoustic Doppler current proler measurements. J Geophys
Res. 109:C04003.
Jun Y, Biyang W, Shicai W. 2004. Method to suppress radio-fre-
quency interference in HF radars. Electronics Letters.
40:145146.
Kaplan DK, Largier J, Botsford LW. 2005. HF radar observations
of surface circulation off Bodega Bay (northern California,
USA). J Geophys Res. 110:C10020.
Kohut J, Roarty H, Randall-Goodwin E, Glenn S, Lichtenwalner
CS. 2012. Evaluation of two algorithms for a network of
coastal HF radars in the Mid-Atlantic Bight. Ocean Dyn.
62:953968.
Kohut JT, Glenn SM. 2003. Improving HF radar surface current
measurements with measured antenna beam patterns. J
Atmos Ocean Technol. 20:13031316.
Kovacevic V, Gacic M, Mosquera IM, Mazzoldi A, Marinetti S.
2004. HF radar observations in the northern adriatic: surface
current eld in front of the Venetian Lagoon. J Mar Sys.
51:95122.
Lipa BJ, Barrick DE. 1983. Least-squares methods for the extrac-
tion of surface currents from CODAR Crossed-loop data:
application at ARSLOE. IEEE J Oceanic Eng. OE-8:226
253.
Lipa B, Nyden B, Ullman DS, Terrill E. 2006. SeaSonde radial
velocities: derivation and internal consistency. IEEE
J Oceanic Eng. 31:850861.
Malacic V, Celio M, Cermelj B, Bussani A, Comici C. 2006.
Interannual evolution of seasonal thermohaline properties in
the Gulf of Trieste (northern Adriatic) 19912003. J
Geophys Res. 111:C08009.
Marmain J, Molcard A, Forget P, Barth A, Ourmières Y. 2014.
Assimilation of HF radar surface currents to optimize
forcing in the North Western Mediterranean sea. Nonlin.
Processes Geophys. 21:659675.
Millot C. 1999. Circulation in the western mediterranean sea. J
Mar Sys. 20:423442.
Molcard A, Poulain P, Forget P, Griffa A, Barbin Y, Gaggelli J,
Maistre JD, Rixen M. 2009. Comparison between vhf radar
observations and data from drifter clusters in the gulf of la
spezia (mediterranean sea). J Mar Sys. 78:S79S89.
Ohlmann C, White P, Washburn L, Terrill E, Emery B, Otero M.
2007. Interpretation of coastal hf radarderived surface cur-
rents with high-resolution drifter data. J Atmos Ocean
Technol. 24:666680.
Orla A, Molcard A, Sayol J, Marmain J, Bellomo L, Quentin C,
Barbin Y. 2014. Empirical forecasting of HF-Radar velocity
using Genetic Algorithms. IEEE Trans Geosci Remote
Sens. 53:28752886.
Ozgokmen TM, Griffa A, Piterbarg LI, Mariano AJ. 2000. On the
predictability of Lagrangian trajectories in the ocean. J Atmos
Ocean Tech. 17:366383.
Paduan JD, Kim KC, Cook MS, Chavez FP. 2006. Calibration and
validation of direction-nding high-frequency radar ocean
surface current observations. IEEE J Oceanic Eng. 31:862
875.
Paduan JD, Rosenfeld LK. 1996. Remotely sensed surface cur-
rents in Monterey Bay from shore based HF radar (Coastal
Ocean Dynamics Application Radar). J Geophys Res.
101:2066920686.
Paduan JD, Washburn L. 2013. High-frequency radar obser-
vations of ocean surface currents. Annu Rev Mar Sci.
5:115136.
Pinardi N, Coppini G. 2010. Operational oceanography in the
Mediterranean Sea: the second stage of development. Ocean
Sci. 6:263267.
Pinot J-M, Lopez-Jurado JL, Riera M. 2002. The CANALES
experiment (19961998). Interannual, seasonal, and mesos-
cale variability of the circulation in the Balearic Channels.
Prog Oceanog. 55:335370.
Poulain P-M. 1999. Drifter observations of surface circulation in
the Adriatic Sea between December 1994 and March 1996. J
Mar Sys. 20:231253.
Poulain P-M, Barbanti R, Cecco R, Fayes C, Mauri E, Ursella L,
Zanasca P. 2004. Mediterranean surface drifter database: 2
June 1986 to 11 November 1999. Tech. Rep. 75/2004/
OGA/31, OGS, Trieste (Italy). CD-ROM and online dataset
http://nettuno.ogs.trieste.it/drifter/database_med/.
Poulain P-M, Ursella L, Brunetti F. 2002. Direct measurements of
water-following characteristics of CODE surface drifters.
Extended Abstracts, LAPCOD Meeting, Key Largo, FL
(USA).
Quentin C, et al. 2013. HF radar in French Mediterranean Sea: an
element of MOOSE Mediterranean Ocean Observing System
on Environment. Ocean & Coastal Observation: Sensors and
observing systems, numerical models & information, Nice
(France).
Radio Regulations. Resolutions and Recommendations. 2012.
International Telecommunications Union (ITU). 319320.
12 L. Bellomo et al.
Downloaded by [Consiglio Nazionale delle Ricerche] at 02:09 30 November 2015
Rypina II, Kirincich AR, Limeburner R, Udovydchenkov IA.
2014. Eulerian and Lagrangian correspondence of high-fre-
quency radar and surface drifter data: effects of radar resol-
ution and ow components. J Atmos Ocean Technol.
31:945966.
Sayol JM, Orla A, Simarro G, Conti D, Renault L, Molcard A.
2014. A Lagrangian model for tracking surface spills
and SaR operations in the ocean. Environ Model Softw.
52:7482.
Sayol JM, Orla A, Simarro G, López C, Renault L, Galán A,
Conti D. 2013. Sea surface transport in the Western
Mediterranean Sea: a lagrangian perspective. J Geophys
Res: Oceans. 118:63716384.
Schiller A and Brassington GB, editor. 2011. Operational ocean-
ography in the 21st century. Dordrecht: Springer.
Schmidt RO. 1989. Multiple emitter location and signal parameter
estimation. IEEE Trans Antennas Propag. 34:276280.
Seim HE, Fletcher M, Mooers CNK, Nelson JR, Weisberg RH.
2009. Towards a regional coastal ocean observing system:
an initial design for the Southeast Coastal Ocean Observing
Regional Association. J Mar Sys. 77:261277.
Sentchev A, Forget P, Barbin Y, Yaremchuk M. 2013. Surface cir-
culation in the Iroise Sea (W. Brittany) from high resolution
HF radar mapping. J Mar Sys. 109110:S153S168.
Shay LK, Cook TM, Hallock Z, Haus BK, Graber HC, Martinez J.
2001. The Strength of the m2 Tide at the Chesapeake Bay
Mouth. J Phys Oceanogr. 31:427449.
Shay LK, Lee TN, Williams EJ, Graber H, Rooth CGH. 1998.
Effects of low-frequency current variability on submesoscale
near-inertial vortices. J Geophys Res. 103:1869118714.
Shay LK, Lentz SJ, Graber HC, Haus BK. 1998. Current structure
variations detected by high-frequency radar and vector-measur-
ing current meters. J Atmos Ocean Technol. 15:237256.
Shay LK, Martinez-Pedraja J, Cook TM, Haus BK, Weisberg RH.
2007. High-frequency radar mapping of surface currents
using WERA. J Atmos Ocean Technol. 24:485503.
Shrira VI, Ivonin DV, Broche P, de Maistre JC. 2001. On remote
sensing of vertical shear of ocean surface currents by means
of a single-frequency vhf radar. Geophys Res Lett.
28:39553958.
Siddorn J, Allen J, Blackford J. 2007. Modelling the hydrodyn-
amics and ecosystem of the North-West European continental
shelf for operational oceanography. J Mar Sys. 65:417429.
Stewart RH, Joy JW. 1974. HF radio measurements of surface cur-
rents. Deep-Sea Res. 21:10391049.
Tintoré J, Coauthors. 2013. SOCIB: The Balearic Islands Coastal
Ocean observing and forecasting system responding to
science, technology and society needs. Mar Technol Soc J.
47:101117.
TOSCA [Internet]. Available from http://tosca-med.eu/.
TOSCA GIS [Internet]. Available from http://toscagis.univ-tln.fr/.
Ullman DS, ODonnell J, Kohut J, Fake T, Allen A. 2006.
Trajectory prediction using HF radar surface currents:
Monte Carlo simulations of prediction uncertainties. J
Geophys Res. 111:C12005.
Unden P, et al. 2002. HIRLAM-5 Scientic Documentation.
US-IOOS HF Radar [Internet]. Available from http://www.ioos.
noaa.gov/hfradar/.
Uttieri M, Cianelli D, Nardelli BB, Buonocore B, Falco P, Colella
S, Zambianchi E. 2011. Multiplatform observation of the
surface circulation in the Gulf of Naples (Southern
Tyrrhenian Sea). Ocean Dyn. 61:779796.
Wang W, Wyatt LR. 2011. Radio frequency interference cancella-
tion for sea-state remote sensing by high-frequency radar. IET
Radar, Sonar & Navigation. 5:405415.
Weisberg RH, Barth A, Alvera-AzcarateA, Zgeng L. 2009. A coor-
dinated coastal ocean observing and modeling system for the
West Florida continental shelf. Harmful Algae. 8:585597.
Zelenke B, et al. 2009. Evaluating connectivity between marine
protected areas using CODAR high-frequency radar.
OCEANS 2009, MTS/IEEE Biloxi Marine Technology
for Our Future: Global and Local Challenges, Biloxi, MS
(USA), 110.
Zervakis V, Georgopoulos D. 2002. Hydrology and circulation in
the North Aegean (eastern mediterranean) throughout 1997
and 1998. Mediterranean Marine Science. 3(1):519.
Zervakis V, Ktistakis M, Nikolaides G, Moschopoulos K, Tragou
E, Vousdoukasa M, Leloudas V, Kozyrakis G. 2009. A new
drifter for oil-spill tracking. Proceedings 9th Hellenic
Symposium in Oceanography and Fisheries, Vol. I, 535539.
Journal of Operational Oceanography 13
Downloaded by [Consiglio Nazionale delle Ricerche] at 02:09 30 November 2015
... The credibility of HFR-derived current data has been extensively proven in numerous coastal areas of the Mediterranean Sea by adopting Eulerian or Lagrangian approaches. Previous investigations included direct comparisons against independent in situ sensors like PCMs, moored ADCPs, drifters, ship-based sensors, or similar (Cosoli et al., 2010;Berta et al., 2014;Lorente et al., 2014Lorente et al., , 2015Lorente et al., , 2021Corgnati et al., 2019a;Lana et al., 2016;Kalampokis et al., 2016;Capodici et al., 2019;Guérin et al., 2021, Molcard et al., 2009Bellomo et al., 2015). When the HFR footprint overlooks a moored instrument within its spatial coverage ( Fig. 4a), an accuracy assessment of HFR surface currents can be performed with radial or total vectors. ...
... In terms of Lagrangian assessment, it is worth mentioning that the Tracking Oil Spill and Coastal Awareness (TOSCA) project experience (Bellomo et al., 2015) constituted one of the first coordinated initiatives at the Mediterranean level to test the precision of a core of 12 HFRs and identify a set of good practices for pollution mitigation. Among other valuable goals, the five-country TOSCA experiment aimed at comparing HFR-derived measurements against the trajectories provided by 20 Coastal Ocean Dynamics Experiment (CODE) drifters (Davis, 1985), which were drogued in the upper 1 m of the oceanic layer and acted as proxies for substances passively advected by currents. ...
... As an overall summary of the validation works, RMSE and CORR values have been typically reported to fall in the ranges 5-20 cm s −1 and 0.32-0.92, respectively (Cosoli et al., 2010;Berta et al., 2014;Lorente et al., 2014Lorente et al., , 2015Lorente et al., , 2021Corgnati et al., 2019a;Lana et al., 2016;Kalampokis et al., 2016;Capodici et al., 2019;Guérin et al., 2021;Molcard et al., 2009;Bellomo et al., 2015). Relative HFR velocity errors can vary widely depending on the characteristics of the site, the radar transmission frequency, the sensor type, and location within the sampled domain, as well as the data processing scheme used (Rypina et al., 2014;Kirincich et al., 2012). ...
Article
Full-text available
Due to the semi-enclosed nature of the Mediterranean Sea, natural disasters and anthropogenic activities impose stronger pressures on its coastal ecosystems than in any other sea of the world. With the aim of responding adequately to science priorities and societal challenges, littoral waters must be effectively monitored with high-frequency radar (HFR) systems. This land-based remote sensing technology can provide, in near-real time, fine-resolution maps of the surface circulation over broad coastal areas, along with reliable directional wave and wind information. The main goal of this work is to showcase the current status of the Mediterranean HFR network and the future roadmap for orchestrated actions. Ongoing collaborative efforts and recent progress of this regional alliance are not only described but also connected with other European initiatives and global frameworks, highlighting the advantages of this cost-effective instrument for the multi-parameter monitoring of the sea state. Coordinated endeavors between HFR operators from different multi-disciplinary institutions are mandatory to reach a mature stage at both national and regional levels, striving to do the following: (i) harmonize deployment and maintenance practices; (ii) standardize data, metadata, and quality control procedures; (iii) centralize data management, visualization, and access platforms; and (iv) develop practical applications of societal benefit that can be used for strategic planning and informed decision-making in the Mediterranean marine environment. Such fit-for-purpose applications can serve for search and rescue operations, safe vessel navigation, tracking of marine pollutants, the monitoring of extreme events, the investigation of transport processes, and the connectivity between offshore waters and coastal ecosystems. Finally, future prospects within the Mediterranean framework are discussed along with a wealth of socioeconomic, technical, and scientific challenges to be faced during the implementation of this integrated HFR regional network.
... The credibility of HFR-derived current data has been extensively proven in numerous coastal areas of the Mediterranean Sea by adopting Eulerian or Lagrangian approaches. Previous investigations included direct comparisons against independent in situ sensors like PCMs, moored ADCPs, drifters, ship-based sensors, or similar (Cosoli et al., 2010;Berta et al., 2014;Lorente et al., 2014Lorente et al., , 2015Lorente et al., , 2021Corgnati et al., 2019a;Lana et al., 2016;Kalampokis et al., 2016;Capodici et al., 2019;Guérin et al., 2021, Molcard et al., 2009Bellomo et al., 2015). When the HFR footprint overlooks a moored instrument within its spatial coverage ( Fig. 4a), an accuracy assessment of HFR surface currents can be performed with radial or total vectors. ...
... In terms of Lagrangian assessment, it is worth mentioning that the Tracking Oil Spill and Coastal Awareness (TOSCA) project experience (Bellomo et al., 2015) constituted one of the first coordinated initiatives at the Mediterranean level to test the precision of a core of 12 HFRs and identify a set of good practices for pollution mitigation. Among other valuable goals, the five-country TOSCA experiment aimed at comparing HFR-derived measurements against the trajectories provided by 20 Coastal Ocean Dynamics Experiment (CODE) drifters (Davis, 1985), which were drogued in the upper 1 m of the oceanic layer and acted as proxies for substances passively advected by currents. ...
... As an overall summary of the validation works, RMSE and CORR values have been typically reported to fall in the ranges 5-20 cm s −1 and 0.32-0.92, respectively (Cosoli et al., 2010;Berta et al., 2014;Lorente et al., 2014Lorente et al., , 2015Lorente et al., , 2021Corgnati et al., 2019a;Lana et al., 2016;Kalampokis et al., 2016;Capodici et al., 2019;Guérin et al., 2021;Molcard et al., 2009;Bellomo et al., 2015). Relative HFR velocity errors can vary widely depending on the characteristics of the site, the radar transmission frequency, the sensor type, and location within the sampled domain, as well as the data processing scheme used (Rypina et al., 2014;Kirincich et al., 2012). ...
Article
Full-text available
Due to the semi-enclosed nature of the Mediterranean Sea, natural disasters and anthropogenic activities impose stronger pressures on its coastal ecosystems than in any other sea of the world. With the aim of responding adequately to science priorities and societal challenges, littoral waters must be effectively monitored with high-frequency radar (HFR) systems. This land-based remote sensing technology can provide, in near-real time, fine-resolution maps of the surface circulation over broad coastal areas, along with reliable directional wave and wind information. The main goal of this work is to showcase the current status of the Mediterranean HFR network and the future roadmap for orchestrated actions. Ongoing collaborative efforts and recent progress of this regional alliance are not only described but also connected with other European initiatives and global frameworks, highlighting the advantages of this cost-effective instrument for the multi-parameter monitoring of the sea state. Coordinated endeavors between HFR operators from different multi-disciplinary institutions are mandatory to reach a mature stage at both national and regional levels, striving to do the following: (i) harmonize deployment and maintenance practices; (ii) standardize data, metadata, and quality control procedures; (iii) centralize data management, visualization, and access platforms; and (iv) develop practical applications of societal benefit that can be used for strategic planning and informed decision-making in the Mediterranean marine environment. Such fit-for-purpose applications can serve for search and rescue operations, safe vessel navigation, tracking of marine pollutants, the monitoring of extreme events, the investigation of transport processes, and the connectivity between offshore waters and coastal ecosystems. Finally, future prospects within the Mediterranean framework are discussed along with a wealth of socioeconomic, technical, and scientific challenges to be faced during the implementation of this integrated HFR regional network.
... The credibility of HFR-derived current data has been extensively proved in numerous coastal areas of the Mediterranean Sea 355 by adopting Eulerian or Lagrangian approaches. Previous investigations included direct comparisons against independent in situ sensors like PCMs, moored ADCP's, drifters, ship-based sensors or similar (Cosoli et al., 2010;Berta et al., 2014;Lorente et al., 2014Lorente et al., , 2015Lorente et al., and 2021Corgnati et al., 2019a;Lana et al., 2016;Kalampokis et al., 2016;Capodici et al., 2019;Guérin et al., 2021, Molcard et al., 2009Bellomo et al., 2015). ...
... In terms of lagrangian assessment, it is worth mentioning that the Tracking Oil Spill and Coastal Awareness (TOSCA) 390 project experience (Bellomo et al., 2015) constituted one of the first coordinated initiatives at Mediterranean level to test the precision of a core of 12 HFRs and identify a set of good practices for pollution mitigation. Among other valuable goals, the 5-country TOSCA experiment aimed at comparing HFR-derived measurements against the trajectories provided by 20 CODE-type drifters (Davis, 1985), which were drogued in the first upper meter of the oceanic layer and acted as proxies for substances passively advected by currents. ...
... Notwithstanding, this allocation is not exclusive, especially in the Mediterranean Sea, and such bands are nowadays used by other official and non-official radio services. As previously pointed out by Bellomo et al. (2015) in the frame of TOSCA project, acquiring a frequency 1000 allocation that allows HFR as a primary user constitutes a key objective for the Mediterranean community in order to mitigate the presence of radio frequency interferences that significantly impact on HFR performance. With ITU regulations becoming increasingly adopted around the world, more and more HFR stations have to share limited, fixed frequency bands. ...
Preprint
Full-text available
Due to the semi-enclosed nature of the Mediterranean Sea, natural disasters and anthropogenic activities impose stronger pressures on its coastal ecosystems than in any other sea of the world. With the aim of responding adequately to science priorities and societal challenges, littoral waters must be effectively monitored with High-Frequency radar (HFR) systems. This land-based remote sensing technology can provide, in near real-time, fine-resolution maps of the surface circulation over broad coastal areas, along with reliable directional wave and wind information. The main goal of this work is to showcase the current status of the Mediterranean HFR network and the future roadmap for orchestrated actions. Ongoing collaborative efforts and recent progress of this regional alliance are not only described but also connected with other European initiatives and global frameworks, highlighting the advantages of this cost-effective instrument for the multi-parameter monitoring of the sea state. Coordinated endeavours between HFR operators from different multi-disciplinary institutions are mandatory to reach a mature stage at both national and regional levels, striving to: i) harmonize deployment and maintenance practices; ii) standardize data, metadata and quality control procedures; iii) centralize data management, visualization and access platforms; iv) develop practical applications of societal benefit, that can be used for strategic planning and informed decision-making in the Mediterranean marine environment. Such fit-for-purpose applications can serve for search and rescue operations, safe vessel navigation, tracking of marine pollutants, the monitoring of extreme events or the investigation of transport processes and the connectivity between offshore waters and coastal ecosystems. Finally, future prospects within the Mediterranean framework are discussed along with a wealth of socio-economic, technical and scientific challenges to be faced during the implementation of this integrated HFR regional network.
... In recent years, surface currents have been applied for civil purposes. Measured surface current data have been used to detect tsunami arrivals [4][5][6][7] and oil spill trajectories [8][9][10][11]. ...
Article
Full-text available
Prior studies have highlighted the importance of calibrating receiver antennas in target direction-of-arrival (DOA) estimation and surface current measurement for high-frequency (HF) radar systems. It is worth noting that the calibration contributes to the performance of both shore-based HF radar and platform-mounted HF radar. Compared with shore-based HF radar, the influence of six-degrees-of-freedom (six-DOF) platform motion should be considered in the calibration of platform-mounted HF radar. This paper initially describes a calibration scheme that receives phasedarray antennas for an anchored platform-mounted HF radar incorporating a model of free rotation, which is called yaw rotation and dominates the six-DOF platform motion in this study. In the presence of yaw rotation, the amplitude and phase of the source calibration signal from the other shore-based radar sites reveal the directional sensitivity of the receiver phased-array antennas. The calibration of receiver phased-array antennas is composed of channel calibration (linking cables and receiver hardware calibration) and antenna pattern calibration. The antenna pattern at each bearing can be represented by the Fourier series. The estimation of channel calibration and antenna pattern calibration depends on an overdetermined HF radar system consisting of observed values and theoretical constraints, so the least-squares fits of the channel calibration coefficients and antenna pattern calibration coefficients are obtained. The experimental results show that the target DOA estimation and surface current measurement can be improved if the phased-array platform-mounted HF radar system is calibrated.
... The use of HF radars in oceanography was introduced in the 1970s [7,8], but the technology matured and was widespread to cover many U.S. and European coastal areas in the 1990s and 2000s [9]. One of the first HF radars installed in the Eastern Mediterranean was the "Dardanos" system, which was installed initially in 2009 as a collaboration between two HIMIOFoTS partners, the University of the Aegean and HCMR [10]. ...
Article
Full-text available
Research infrastructures have been established throughout Europe in order to create robust organizations that will facilitate and enhance research and innovation processes and will advance society with innovative products and services. The Hellenic Integrated Marine Observing, Forecasting and Technology System (component of HIMIOFoTS RI) has been implemented in the framework of the National Roadmap for Research Infrastructures to form a large-scale infrastructure for the marine environment in Greece. It links together ocean observing and forecasting systems, coastal zone monitoring and management practices, as well as ocean engineering testing facilities. The overarching framework of the system supports the coordination of five organizations with expertise in the field of marine science and technology, the central management of research activities, and the common development of services and products. It comprises facilities and resources while it provides open access to research communities (academia, industry) to support the scientific advancements and innovation in their fields. The Hellenic Marine Observing, Forecasting and Technology System was further enhanced during its implementation through significant upgrades and developments in order to extend its observing capacity and the forecasting and technological abilities, while advancing the provided services and products.
Chapter
This chapter aims to describe and analyse different types of sea monitoring networks and their potential and characteristics. The life history of operational instruments involves different phases, such as planning according to the recorded parameter, selection and installation of sensors, operation, calibration, maintenance and training activities. These factors are illustrated and discussed. Moreover, the quality of the measurements data is considered of paramount importance because it must guarantee the efficacy of monitoring, design, environmental studies and civil protection activities along the coast. Main procedures of data quality control and dissemination, both in real and delayed time, are summarized, and reference is made to international guidelines. Still today, a series of meteo-marine observations are currently spread in a large number of public and private institutions, making difficult the realistic and updated knowledge of what is available and how to obtain part of those data for climatic studies, numeric model calibration, planning and management of coastal zone. The need of an integrated national sea state monitoring system is demonstrated as the way to be undertaken to reach a correct management and safety of both coastal and open sea areas; this system must ever take into account a series of technological and scientific progress which has broadened the spectrum of potentially available tools to the scientific community and institutional stakeholders.
Chapter
Every year, vast quantities of plastic debris arrive at the ocean surface. Nevertheless, our understanding of plastic movements is largely incomplete and many of the processes involved with the horizontal and vertical displacement of plastics in the ocean are still basically unknown. In this chapter we review the dynamics associated with the transport of plastics and other pollutants at oceanic fronts. Fronts had been historically defined as simple barriers to exchange, but here we show that the role of these structures in influencing the transport of plastics is more complex. The tools used to investigate the occurrence of frontal structures at various spatial scales are reviewed in detail, with a particular focus on their potential applications to the study of plastic pollution. Three selected case studies are presented to better describe the role of fronts in favoring or preventing plastic exchanges: the large-scale Antarctic Circumpolar Current, a Mediterranean mesoscale front, and the submesoscale fronts in the Gulf of Mexico. Lastly, some aspects related to the vertical subduction of plastic particles at oceanic fronts are discussed as one of the most promising frontiers for future research. The accumulation of floating debris at the sea surface is mainly affected by the horizontal components of frontal dynamics. At the same time, vertical components can be relevant for the export of neutrally buoyant particles from the surface into the deep sea. Based on these evidences, we propose that submesoscale processes can provide a fast and efficient route of plastic transport within the mixed layer, while mesoscale instabilities and associated vertical velocities might be the dominant mechanism to penetrate the deeper ocean on slower but broader scales. We conclude that given the ubiquitous presence of fronts in the world’s ocean, their contribution to the global plastic cycle is probably not negligible and the role of these processes in vertically displacing neutrally buoyant microplastics should be investigated in more detail.
Article
We report on the installation and first results of one compact oceanographic radar in the region of Nice for a long-term observation of the coastal surface currents in the North-West Mediterranean Sea. We describe the specific processing and calibration techniques that were developed at the laboratory to produce high-quality radial surface current maps. In particular, we propose an original self-calibration technique of the antenna patterns, which is based on the sole analysis of the database and does not require any shipborne transponder or other external transmitters. The relevance of the self-calibration technique and the accuracy of inverted surface currents have been assessed with the launch of 40 drifters that remained under the radar coverage for about 10 days.
Article
Coastal regions are vulnerable areas with often high population density, as well as tourism and maritime activities that may have negative impact on the environment. From a physical point of view, coastal areas may be characterized by high gradient topography and irregular coastline shapes resulting in complex dynamic systems. The monitoring of coastal circulation becomes necessary to support coastal management and to understand the high variability of the dynamics. The simultaneous use of comprehensive observational systems and numerical models may compensate the drawback of each method used separately. The Toulon coastal area is under investigation in this paper by means of HF RADAR and ADCP observations coupled with nested models. The integration of the different data sets allows the monitoring of the coastal ocean continuum from regional oceans and shelf areas. Summer and winter 2018 data are analyzed to depict the seasonal variability of the regional circulation mainly characterized by the geostrophic Northern boundary Current, the wind-driven bay circulation and the connectivity between the bay, the surrounding Marine Protected Area and the open sea.
Chapter
Full-text available
This chapter introduces the expected likelihood approach as a mechanism that allows assessment of the quality of estimates without resorting to asymptotic or clairvoyant analysis. To address an important low-sample support regime, the expected likelihood approach was expanded into the sample data circumstances where the number of samples T does not exceed the dimension M of each snapshot. Under sampled likelihood ratios that satisfy these requirements were then derived via two different mechanisms: projection of the covariance matrix model onto the subspace spanned by the sample covariance matrix, and a maximum-entropy extension of the central band of the rank-deficient sample covariance matrix. The expected likelihood detection/estimation methodology was applied over an important class of “under sampled” scenarios, where the number of independent Gaussian samples T does not exceed the number of antenna array elements M, by applying likelihood ratios formulated for operation with singular sample covariance matrices. Specific application of the introduced methodology dealt with the well-known “threshold effect” in MUSIC, where the lack of SNR and/or sample support T below certain threshold values causes MUSIC to generate severely erroneous DOA estimates (outliers) with high probability.
Article
Full-text available
In the framework of the French MOOSE project (Mediterranean Ocean Observing System on Environment), the Mediterranean Institute of Oceanography is operating HF radars on the North Western Mediterranean coast. The surface circulation in this region is characterized by a large-scale flow (Northern Current) and by a broad range of other scales of variability induced by meteorological and tidal forcing. The ability of HF radars is to provide synoptic observation as sea surface current map every hour and over long distances. One site is already operational nearby Toulon for more than two years and a second one is in deployment around Nice. This paper gives an overview of the radars network, of the surface current mapping facility offered by the system, and of recent observation results and applications.
Book
Over the past decade the significant advances in real-time ocean observing systems, ocean modelling, ocean data assimilation and super-computing has seen the development and implementation of operational ocean forecasts of the global ocean. At the conclusion of the Global Ocean Data Assimilation Experiment (GODAE) in 2008 ocean forecasting services were being supported by 12 international centres. The book is about ocean forecasting - a maturing field which remains an active area of research, and includes the discussion on such topics as ocean predictability, observing system design, high resolution ocean modelling and ocean data assimilation. It presents the introduction to ocean forecasts which allow new opportunities in areas of coupled bio-geochemical forecasting and coupled atmosphere-wave-ocean forecasting. The book describes the research and development to improve forecast systems, determining how best to service the marine user community with forecast information as well as demonstrating impact to their applications. It also discusses operational centres which develop online graphical and data products for their user communities and obtain real-time feedback on the quality of this information. The contents of this book are aimed at early career scientists and professionals with an interest in operational oceanography and related ocean science. There are excellent opportunities for exciting research through operational forecasting careers in order to address current and future challenges in this field and to service a growing user community.
Conference Paper
The U.S. Integrated Ocean Observing System (IOOS®) partners have begun an effort to extend the use of high frequency (HF) radar for U.S. Coast Guard (USCG) search and rescue operations to all U.S. coastal areas with HF radar coverage. This project builds on the success of an IOOS and USCG-supported regional USCG search and rescue product created by Applied Science Associates (ASA), Rutgers University and University of Connecticut for the mid-Atlantic region. We describe the regional product and the expanded national product's two main components: optimally-interpolated velocity fields and a predicted velocity field.