emo e sensing
A icle
Quali y Assessmen and P ac ical In e p e a ion o
he Wa e Pa ame e s Es ima ed by HF Rada s in
NW Spain
Ana Basañez 1,∗, Pablo Lo en e 2, Ped o Mon e o 3, En ique Ál a ez-Fanjul 2and
Vicen e Pé ez-Muñuzu i 1
1Ins i u e CRETUS, G oup o Nonlinea Physics, Uni e si y o San iago de Compos ela,
15782 San iago de Compos ela, Spain; [email p o ec ed]
2Pue os del Es ado, 28042 Mad id, Spain; [email p o ec ed] (P.L.); [email p o ec ed] (E.Á.-F.)
3INTECMAR, Xun a de Galicia, 36611 Vilaga cía de A ousa s/n, Spain; pmon e o@in ecma .gal
*Co espondence: anabme [email p o ec ed]
Recei ed: 28 Janua y 2020; Accep ed: 7 Feb ua y 2020; Published: 11 Feb ua y 2020
Abs ac :
High- equency (HF) ada s a e e icien ools o measu ing as a eas and ga he ing ocean
pa ame e s in eal- ime. Howe e , he accu acy o hei wa e es ima es is unde analysis. This pape
p esen s a new me hodology o analyzing and alida ing he wa e da a es ima ed by wo CODAR
SeaSonde ada s loca ed on he Galician coas (NW Spain). App oxima ely one and a hal yea s o
wa e da a (Janua y, 2014–Ap il, 2015) we e ob ained o en ange cells employing wo di e en
sampling imes used by he ada so wa e. The esul ing da a we e sc eened by an upda ed me hod,
and hei abundance and quali y we e desc ibed o each ada ange cell and di e en wa e egime;
he la e we e de ined using he spec al signi ican wa e heigh (Hm0) and mean wa e di ec ion
(Dm) es ima ed by wo buoys and h ee SIMAR poin s (SImulación MARina in Spanish, om he
wa e eanalysis model by Pue os del Es ado (PdE)). The co ela ion be ween he esul s and he
pa icula i ies o he di e en sea s a es (b oadband o bimodal), he wind and he ope a ion o he
de ices a e discussed. Mos HF ada wa e pa ame e s’ e o s occu o wa es om he NNE and
highe han 6 m. The bes ag eemen be ween he Vilán ada and he Vilano-Sisa gas buoy wa e
da a was ob ained o he dominan wa e egime ( om he no hwes ) and he sou hwes wa e
egime. Howe e , ele an con adic ions ega ding wa e di ec ion we e de ec ed. The possibili ies
o educing he wa e pa ame e s’ p ocessing ime by one hou and inc easing he numbe s o ange
cells o he ada s ha e been alida ed.
Keywo ds: HF ada ; wa e egime; wa e modeling; emo e sensing
1. In oduc ion
The accu acy o wa e egime assessmen and o ecas ing is one o he g ea es challenges o
he de elopmen and sa e y o ma ine ac i i ies. Un o una ely, adi ional in si u wa e gauges,
such as buoys, he mos ex ended one, a e single-poin measu ing de ices which also ha e o he
limi a ions, such as he di icul y o making an accu a e moo ing ins alla ion o a oid in e e ence in
wa e measu emen s and he isk o su e ing acciden s o andalism on he wa e su ace. Toge he
wi h he di icul y o doing main enance and epai s du ing some seasons due o ha sh wea he
condi ions, his all usually leads o many pe iods wi hou da a.
The e o e, new emo e sensing echnologies ha e been adap ed o de eloped o analyze and
measu e he wa es. Among hese, high- equency (HF) ada s ha e become an impo an al e na i e
as hey a e able o co e la ge a eas o he ocean’s su ace, ga he ing cu en , wa e and wind da a [
1
–
5
].
Remo e Sens. 2019,12, 598; doi:10.3390/ s12040598 www.mdpi.com/jou nal/ emo esensing
Remo e Sens. 2019,12, 598 2 o 24
HF ada echnology is based on he Dopple shi o he ada ’s wa e backsca e a e colliding
wi h he ocean g a i y wa es [
6
]. While cu en pa ame e s a e ex ac ed om he i s -o de peaks o
he Dopple -spec um, which a e p oduced by ocean wa es o leng h one hal o he ada ’s leng h,
and wi h di ec ion ei he away o owa ds he ada , he comple e wa e di ec ional spec um is
based on he second-o de peaks, p oduced by he in e ac ion be ween any pai o ocean wa es
no malized by he p e ious i s -o de spec a [
7
,
8
]. Se e al designs o HF ada s ha e been de eloped
o ob aining ocean wa es’ in o ma ion [
9
–
12
], and he one used in his pape is a b oad-beam di ec ion
inding o CODAR, whose basis o es ima ing he di ec ional wa e spec um has been ex ensi ely
explained [
8
,
13
,
14
]. The p esen model o hese ada s, he SeaSonde, can use elec omagne ic wa es o
equencies be ween 4.4–50 MHz, bu in oceanog aphy, he mos ypically used a e 5, 12 and 25 MHz.
One o he bene i s o hese ada s is o ha e only one o wo monopole an ennas ( o emi ing and
ecei ing) which acili a e he loca ion and ins alla ion [15,16].
Howe e , HF ada s ha e some limi a ions [
17
,
18
]; hei wo k equency de e mines a h eshold
o he a io be ween he wa e backsca e and he noise loo o he signal ha limi s he minimum
wa e heigh ha can be measu ed by he ada , and de e mines also a maximum wa e heigh ha ,
when exceeded, ensu es he second-o de peaks a e me ged wi h he i s -o de ones, so he wa e
spec um canno be de ined [
19
]. Addi ionally, o la ge su ace cu en eloci ies, he i s -o de peaks
inc ease hei wid h, hiding he second-o de peaks [
14
]. Mo eo e , HF ada wa e pa ame e s can
be modula ed by ides [
4
] and by ine ial cu en s, al hough his las p ocess can also pe u b he
calcula ion o he buoys’ wa e pa ame e s, o he HF ada case is no clea ly de ined ye [20].
Fu he mo e, HF ada s-de i ed wa e pa ame e s a e no usually ob ained di ec ly om he
wa e spec um [
21
]; ins ead, he non-di ec ional wa e spec um is i ed o a Pie son-Moskowi z
model speci ic o each ange cell and some p ede ined coas line limi s, which a e adap ed o each
ada loca ion. F om his model, he spec al signi ican wa e heigh (Hm0) and cen oid pe iod
(Tc) a e ex ac ed. Then, mean wa e di ec ion (Dm) is ex ac ed om he di ec ional wa e spec a,
whose di ec ion ac o is calcula ed assuming a ca dioid dis ibu ion [
22
]. Al hough heo e ically, wa e
pa ame e s can be ob ained o an i egula ba hyme y and unde di e en sea s a e condi ions [
22
],
o he momen , he so wa e o CODAR ada s wo ks wi h he basis o open wa e s and unimodal
and homogeneous wa e condi ions h oughou he measu ed a ea, which induces some e o s o he
Hm0 as i s o e es ima ion du ing bimodal sea s a es [
23
,
24
] o gene a ing a e age wa e pa ame e s
om an a ea pa ched up wi h di e se wa e condi ions [4].
Despi e his, many au ho s ha e e alua ed he quali y o such ada -de i ed wa e pa ame e s
wo ldwide, concluding ha HF ada s a e a eliable sou ce o wa e in o ma ion [
25
–
28
]. The ada s,
he objec i e o his p ojec , a e pa o he HF ada ne wo k deployed on he Galician coas (Figu e 1),
consis ing o ou HF ada s wo king wi h a equency o 4.86 MHz, by which he highes sa u a ion
limi is 24 m [
29
], bu o he signal o he second o de peaks, ising o e he noise loo , he minimum
wa e heigh can be la ge depending on he si ua ion; hence, hese ada s wo k be e wi h high wa es
han wi h small ones [
14
,
30
]. Besides, wa es should ha e a leas a pe iod o 5 s, o longe i he noise
loo aises, o be in e p e ed by he ada [25].
P e ious alida ion o wa e da a om he Galician ada ne wo k concluded ha Sillei o ada
Hm0 is in good ag eemen wi h buoy da a, bu also e lec ed some p oblems es ima ing Tc when
bimodal sea s a e occu ed [
31
]. On he o he hand, Lo en e e al. [
30
] showed, o a longe da a ime
se ies, some di e ences in wa e heigh es ima ion be ween he HF ada s om Sillei o and Vilán and
he closes buoys, which could be due o he in luence o some oceanog aphic condi ions, such as ides,
cu en s and bimodal sea s a es, and also di e ences in wa e di ec ion p obably induced by he dep h
a a eas close o he sho eline. This wo k also e ealed ha he de ia ion o he wa e di ec ion was
di e en o each ada .
Remo e Sens. 2019,12, 598 3 o 24
VILA
CW
VB
VBW
S24
S20
VILA
SILL
S02
SB
Figu e 1.
No hwes e n coas o Galicia (NW Spain) wi h se en mean di ec ion wa e oses and wo
wind oses ob ained om di e en da a sou ces. Sampling om Janua y, 2014–Ap il, 2015.
Wa e
oses: VB
: Vilano-Sisa gas buoy.
VILA
: Vilán high- equency (HF) ada —wa e ose o ange cell a
10 km (RC 10 km).
S24
: SIMAR poin 3002024.
S20
: SIMAR poin 3004020.
SB
: Sillei o buoy.
SILL
:
Sillei o HF ada —wa e ose o RC 10 km.
S02
: SIMAR poin 3014002.
Red buoy symbol
: buoys
loca ion.
Red iangle
: SIMAR poin s loca ion.
Blue squa es
: ada s loca ion.
Con inuous black
ci cles
: ada s ange cells (un il 25 km).
Con inuous s aigh line
: ada coas line limi s (CL).
Wind
oses: CW
: Cama iñas me eo ological s a ion ( ue da a sou ce loca ion a ed squa ed symbol).
VBW
:
Vilano-Sisa gas buoy.
Conside ing hese ac o s, he complex wea he and he cha ac e is ics o he Galician coas , ou
s udy aims o elabo a e an exhaus i e me hodology o desc ibing and in e p e ing he eliabili y o
wa e pa ame e s es ima ed by he Sillei o and Vilán HF ada si es, so hei da a may be used di ec ly
by he end-use s.
Thus, all ada s’ spec a we e e-p ocessed; hen, he ypes o wa e egimes no measu ed by he
ada s we e iden i ied, and he new clean wa e da a o he en ange cells we e alida ed wi h nea by
buoys. The weaknesses and s eng hs o he ada es ima ions o he usual wa e egimes o he a ea
we e es ablished. In o de o suppo he discussion o he esul s, he wa e egime o he a ea was
desc ibed by employing he Hm0 and Dm calcula ed by in si u wa e-gauges and eanalysis modeling
da a a di e en loca ions along he coas .
Addi ionally, we a emp ed o upda e he p esen me hodology o wa e da a sc eening de ined
by he Cope nicus Ma ine In Si u Team [
32
] in he amewo k o he Cope nicus Ma ine En i onmen
Moni o ing Se ice (CMEMS). Such sc eening does no include a me hod o deal wi h wa e p oduc s
om HF ada s. Finally, we we e able o educe by one hou , he wa e spec um p ocessing ime.
2. Ma e ials and Me hods
The ada s selec ed o he analysis we e he Vilán ada (VILA), owned by In ecma -Xun a de
Galicia and he Sillei o HF ada (SILL) owned by Pue os del Es ado (PdE). Bo h a e long- ange
CODAR SeaSonde HF ada s, composed o 2 an ennas deployed only a ew me e s apa : one is o
Remo e Sens. 2019,12, 598 4 o 24
emi ing and o he is a composi ion o 3 c ossed an ennas ha independen ly ecei e he backsca e
signal, which is used du ing he wa e di ec ion- inding p ocess. The wo king equency is 4.86 MHz,
and he ansmission signal’s sweep wid h is
≈
29.41 KHz. The backsca e signal is spa ially sol ed
in o concen ic ings a ound he ada loca ion, named ange cells (see Figu e 1). These a e 5 km
hick, which is he spa ial esolu ion de ined by he sweep wid h [
16
]. Thi y minu es o he ecei ed
signal a e p ocessed by he p op ie a y CODAR so wa e in o Dopple c oss spec a which a e sa ed
in iles named CSS (C ossSpec a Sho ime) [
33
]. Then, by means o he in e sion i ing wi h he
Pie son-Moskowi z model and a ca dioid di ec ional ac o , a p e- ixed sampling pe iod, de ined by
a ce ain numbe o CSS iles, is used o calcula e he wa e di ec ional spec um [22].
Fo his wo k, CSS iles om Janua y, 2014–Ap il, 2015 we e e-p ocessed using an upda e o he
elease 7 o he CODAR SeaSonde so wa e, and he ollowing con igu a ion was applied:
•
Selec ion o sampling pe iod o be in oduced in he model i ing, which will de e mine he
delayed ime be ween he wa e de ec ion and he equa ion o he spec al wa e pa ame e s—Tc,
Hm0 and Dm. The CSS iles o Vilán ada we e p ocessed wice, using 180 and 120 min. Sillei o
ada CSS iles we e p ocessed o e 180 min.
•The pe iodici y o he esul ing wa e pa ame e s was ixed o hi y minu es.
•
De ini ion o coas line limi s (CL) o he co e age a ea, and he wa e bea ing limi s (WB) acco ding
o he sho eline and he open sea su ounding he ada . In his case, due o he coas line acing
open sea and he wa es om land being conside ed negligible, CL and WB coincide. Hence,
o Vilán ada hese we e ixed o 221
◦
and 41
◦
(all di ec ions a e de ined as clockwise om
he ue no h) and o Sillei o ada o 180
◦
and 350
◦
. On he o he hand, o de ine he ange,
he numbe o ange cells was ixed o en. Since he i s 5 km om he ada ’s loca ion a e no
included in he sampled a ea, he o al ada ange is 55 Km.
•
Minimum and maximum wa e pe iod we e ixed o 5 and 17 s, ollowing CODAR
ecommenda ions [33].
The esul ing wa e pa ame e da a se s we e sc eened by a me hod based on he quali y con ol
o wa es de eloped by he Cope nicus Ma ine In Si u Team [
32
]. The upda e consis ed o aking in o
conside a ion he eno mous numbe o samples lagged as nulls by he ada so wa e o de ec s uck
alues and jumps.
Wi hin he HF ada ’s oo p in a e wo pe manen ly deployed SeaWa ch buoys (Figu e 1), which
we e deployed by PdE: Vilano-Sisa gas buoy (VB), 40 km o he no h o Vilán Cape, and Sillei o
buoy (SB), 60 km in on o he Sillei o ada . These buoys ake mo ion measu emen s o wen y six
minu es and gene a e spec al wa e pa ame e s iles each hou [
34
]. Fo his wo k, mean di ec ion
(Dm), peak di ec ion (Dp), peak pe iod (Tp) and spec al signi ican heigh (Hm0) we e used o
alida e he HF ada s’ wa e pa ame e s. Besides, he mean wind eloci y and di ec ion measu ed by
Vilano-Sisa gas buoy we e also used (VBW, Figu e 1).
Fo he wa e egime desc ip ion all h ough he co e age a ea o bo h ada s, wa e da a om
h ee SIMAR poin s we e used: 3002024 (S24); 3004020 (S20) and 3014002 (S02) (Figu e 1). These a e
pa o he subse WANA, based on he eanalysis o he WAM and Wa eWa ch models un by PdE.
The inal p oduc s a e ou da a se s o hou ly basis spec al wa e pa ame e s: one ex ac ed om
he o al wa e spec um, wo om he spec a o wo possible swells and one om he wind wa es
spec um [35].
Wind da a om Cama iñas me eo ological s a ion (CW, Figu e 1) we e ob ained om Me eoGalicia
wea he se ice [36].
All da a se s we e ime pai ed on an hou ly basis. The alida ion o he ada wa e pa ame e s
was ca ied ou mainly by compu ing he Pea son co ela ion index (R); he mean absolu e pe cen
e o (MAPE); he oo mean squa ed e o (RMSE); he bias; and in he case o he wa e di ec ion,
he ci cula R and ci cula mean absolu e e o (MAE).
The collec ed da a spans om 30 Janua y 2014–30 Ap il 2015. Howe e , he abundance o samples
om each de ice is di e en due o he deploymen schedules o mal unc ions du ing ope a ion.
Remo e Sens. 2019,12, 598 5 o 24
Hence, he e a e some ele an sampling gaps: 1–19 Augus 2014; 31 Janua y–2 Feb ua y 2015; and 31
Ma ch–9 Ap il 2015—p obably due o a VILA mal unc ion. 17 Janua y–20 Feb ua y 2014; and 21
Ap il–22 July 2015—due o a b eak on VB deploymen . Finally, a gap om 21–25 Augus 2015 in he
SILL da a se .
3. Resul s and Discussion
Fo me alida ion o VILA ange cell a 25 km (RC 25 km) and Vilano-Sisa gas buoy Hm0 by
Lo en e e al. [
30
] e ealed a consis en co ela ion. Howe e , ada wa e da a had a lo o ou lie s
which had a high impac in he s a is ical esul s (R
≈
0.75). Hence, o his wo k, VILA da a we e
e-p ocessed wi h a mo e ecen elease o he ada ’s so wa e which is mo e accu a e, p oducing
ewe ou lie s. Mo eo e , he sc eening applied inc eased he numbe o ou lie s de ec ed; hence, R go
a alue o ≈0.88 o he same ange cell.
Figu e 2shows he good ag eemen be ween he VILA Hm0 ime se ies o RC 10 km and he
co esponding buoy pai ed da a. Howe e , he e a e many gaps which ex end om some hou s o
many days ha do no co espond o he ope a ion gaps desc ibed in he p e ious sec ion.
Figu e 2.
Spec al signi ican wa e heigh (Hm0) ime se ies: RC 10 km o Vilán ada (VILA, blue) and
Vilano-Sisa gas buoy (VB, ed). Hm0 co ela ion index: 0.88.
Table 1shows ha he highes loss o da a co esponds o ada samples lagged as nulls by he
so wa e, which in some cases can each mo e han 70
%
o he samples. This lagging occu s when he
numbe o backsca e Dopple poin s a e no enough o accu a ely calcula ing he Dopple spec um,
and hence he wa e pa ame e s [33].
Remo e Sens. 2019,12, 598 6 o 24
Table 1.
Numbe o Vilán ada (
VILA
) ange cells and Vilano-Sisa gas buoy (
VB
) samples.
Ideal
samples
: heo e ical numbe o samples o he sampling pe iod.
To al samples
: eal samples
gene a ed.
Gaps
: pe iods wi hou da a.
Nulls
: VILA samples lagged as nulls.
Sc eened ou lie s
:
VILA ou lie s lagged by ou own sc eening and VB ou lie s lagged by PdE sc eening.
Time pai ed
aw samples
: samples a e selec ing he hou ly coinciden samples o VB and VILA, including nulls
and ou lie s.
Time pai ed clean samples
: samples a e selec ing he hou ly coinciden samples o VB
and VILA, excluding nulls and ou lie s.
01/01/2014–30/04/15 VILA Range Cells
VB 5 Km 10 Km 15 Km 20 Km 25 Km 30 Km 35 Km 40 Km 45 Km 50 Km
Ideal 1 h/30’ samples 11,640 23,279 23,279 23,279 23,279 23,279 23,279 23,279 23,279 23,279 23,279
To al samples 10,580 21,826 21,826 21,826 21,826 21,826 21,826 21,826 21,826 21,826 21,826
% o gaps 9.10% 6.44% 6.44% 6.44% 6.44% 6.44% 6.44% 6.44% 6.44% 6.44% 6.44%
% o nulls – 61.13% 52.99% 52.25% 53.83% 55.79% 60.33% 62.87% 66.37% 68.84% 72.04%
Sc eened ou lie s 25 99 55 67 93 64 90 106 131 86 67
Time pai ed aw samples 9833 9833 9833 9833 9833 9833 9833 9833 9833 9833 9833
Time pai ed clean samples – 3449 4276 4345 4173 3976 3510 3245 2916 2685 2399
3.1. Analysis o HF-Rada Da a Loss
Such a pe cen age o ada samples ep esen s a g ea loss o he wa e in o ma ion ha will be
excluded om he s a is ical analysis. Wi h he aim o inding ou whe he he e is any end in his
loss besides he ope a ional limi s o he ada s, a co ela ion be ween hese nulls and he wa e egime
desc ibed by he ela ed samples o he buoys was ca ied ou .
The esul s o VILA case, ep esen ed in Figu e 3a, shows a s ong co espondence be ween he
loss o da a and he wa es be ween 1–3 m, bu also o he highe wa es. Since he absence o Dopple
poin s mos ly happens when he second-o de peaks canno su pass he noise le el, he small wa es
a e mo e suscep ible o no being de ec ed by he ada . Ano he possibili y is ha he noise le el
inc eases, masking hen, he signal, e en o wa es la ge han 2 m [
24
]. The his og am (Figu e 3b)
shows how ele an his loss is, especially o wa es om 2 o 4 m in heigh , and he displacemen
be ween he modes o he VILA and VB Hm0 alues. Excep o RC 5 km, inc easing he dis ance
o he ada inc eases he da a loss (Table 1). This could be due o a dec easing o he s eng h o
he backsca e signal wi h he dis ance. In con as , he u hes ange cells desc ibe sligh ly mo e
wa es highe han 6 m, p obably as a esul o he sea s a e he e ogenei y a ound he ange cells, o
mo e likely due o he ada ’s so wa e end o p oduce mo e Hm0 ou lie s a he u hes ange cells
(see sc eened ou lie s in Table 1).
Remo e Sens. 2019,12, 598 7 o 24
(b)
(a)
Figu e 3.
Compa ison be ween ime pai ed da a o Vilán ada (VILA) and Vilano-Sisa gas buoy (VB).
(
a
) Pe cen age o VILA nulls a each combina ion o VB mean wa e di ec ion (Dm) and Hm0 ( esolu ion
5
◦×
1 m). (
b
) Hm0 his og ams o VB (black), VILA RC 5 km (blue), RC 10 km ( ed) and RC should
says 30 km.
SILL nulls a e 20
%
less han o VILA case, and co espond mos ly wi h wa es smalle han 2 m,
and in lowe pe cen ages o highe wa es (Figu e A1).
3.2. Wa e Pa ame e s Valida ion in Te ms o Rada Range Cells
Valida ion o wa e pa ame e s was examined be ween pai ed samples o VILA and VB excluding
he ada nulls and o he e o s de ec ed du ing he sc eening, so mo e han 50
%
o he wa e
in o ma ion was no included in he s a is ical analysis.
Figu e 4a– show wa e alida ion o each ange cell da a se . No e ha o he ange cells om
10–30 km he e is a s ong Hm0 co ela ion, wi h R
≈
0.88 (a) and low MAPEs (b). The alue o R o
he las ou ange cells alls o 0.78 (a). Albei he linea co ela ion be ween he pe iods is no so high
Remo e Sens. 2019,12, 598 8 o 24
(
≈
0.7), he MAPEs emain below 12
%
(c, d). Rega ding Dm, ci cula R emains below 0.7 o all ange
cells, bu again, mean e o s a e small, wi h all MAEs emaining below 26◦(e, ).
(c) (e)
( )
(d)
(b)
(a)
Figu e 4.
Valida ion o wa e pa ame e s o en Vilán ada (
VILA
) ange cells (5–50 km) wi h
Vilano-Sisa gas buoy es ima es: (
a
) Hm0 co ela ion index (R); (
b
) Hm0 mean absolu e pe cen e o
(MAPE); (c) pe iod R; (d) pe iod MAPE; (e) Dm ci cula R; ( ) Dm mean absolu e e o (MAE).
Al hough he di e ences be ween he alues o hese s a is ical pa ame e s a e small, in mos
cases disc epancies be ween he ada and he buoy inc ease a e RC 30 km, which coincides wi h he
las ange cell ecommended by CODAR o he long- ange ada s [33].
Howe e , Figu e A2 shows ha u he om he coas , di e ences be ween Sillei o ada and
Sillei o buoy do no inc ease. Mo eo e , he co ela ion o Dm alues imp o es wi h dis ance, p obably
because he a hes ange cells a e mo e exposed o he open sea and ha e less in luence o he
coas line, which is less ugged compa ed o VILA si e.
Fu he alida ion o wa e pa ame e s was ocused on VILA RC 10 km, as i is he one wi h he
highes numbe o samples and highes ag eemen wi h he buoy da a. This dis ance om he coas is
also he a ea wi h mo e po en ial o ishing and ma ine ac i i ies (leisu e, enewable ene gy, e c).
Figu e 5a–c show he eg ession lines be ween VILA and i s co esponding VB wa e pa ame e s.
The eg ession line o Hm0 has a e y small in e cep and slope, which ein o ces he signi icance
o he alue o R
≈
0.88 (a). Howe e , da a dispe sion is ele an , and some spu ious poin s can be
de ec ed. Besides, a e pai ing he wo da a se s, all he buoy alues below 1 m ha e been excluded;
he e o e, all he ada wa es below 1 m a e unde es ima ed when compa ed o buoy da a. Mo eo e ,
he la ges buoy wa es a e also unde es ima ed by he ada . Linea eg ession o he wa e pe iods
(b) esul s in a smalle bu ele an R
≈
0.78; howe e , in his case, he in e cep is la ge due o he
minimum ada ope a ion pe iod o 5 s; he sca e plo desc ibes also a ada unde es ima ion o many
buoy pe iods be ween 10 and 15 s.
The ci cula co ela ion index o Dm only eaches 0.66 (Figu e 5c). The mos ele an di e ences
occu when he buoy desc ibes no h and no heas wa e di ec ions, while he ada shows no hwes
di ec ions. The e is also a signi ican sou hwa ds de ia ion in he ada es ima es which includes
an accumula ion o poin s a he sou h ada coas line limi (221◦).
In gene al, alida ion o Sillei o ada wi h Sillei o buoy desc ibed a sligh ly lowe co ela ion
han VILA wi h VB o he h ee wa e pa ame e s (Figu e A3a– ), especially, as Lo en e e a . [
30
]
desc ibed, he co ela ion be ween pe iods (c). He e, he sou hwa ds end o he ada Dm alues
is e y no iceable (c). The exclusion o 2014 win e had a ele an in luence on hese esul s when
compa ed o hose by Lo en e e al. [30].
Remo e Sens. 2019,12, 598 9 o 24
(b) (c)
Figu e 5.
Wa e pa ame e s sca e plo s o Vilán ada (
VILA
) RC 10 km s. Vilano-Sisa gas buoy (
a
)
Hm0 eg ession line; R = 0.88; slope = 0.96. (
b
) Reg ession line be ween VILA cen oid pe iod (Tc) and
VB peak pe iod (Tp); R = 0.78; slope = 0.73. (c) Dm sca e plo ; Ci c.R = 0.66
These Hm0-R alues a e simila o hose desc ibed in ea lie alida ion wo ks o SeaSonde HF
ada s, such as Long [
26
] wi h a 12–13 MHz SeaSonde (R
≈
0.85–0.91), and Al onso [
31
] who alida ed
win e -pe iod wa e da a o Sillei o ada (R = 0.89). Ano he HF ada s, wi h a wo k equency
o 25 MHz, desc ibed a sligh ly smalle co ela ion (R
≈
0.65–0.71 [
25
], R
≈
0.78 [
27
]). In gene al,
he ag eemen be ween HF ada s and buoy pe iods is di icul since he ada cen oid pe iod does
no coincide ei he wi h buoys Tp o Tm [
4
]. Long and Sa iano’s wo ks [
26
,
37
] compa ed Tc wi h
Tm and Tp, demons a ing ha Tc akes alues be ween he wo o hem. Rega ding wa e di ec ion,
he me hod used by hese ada s, he di ec ion- inding, is condi ioned by he ca dioid model [
14
] and
by an accu a e an enna pa e n measu emen [
38
]. In addi ion, and as i is explained la e , in gene al,
he h ee pa ame e s a e condi ioned o an unimodal sea s a e and an assumed uni o mi y h ough all
ange cells.
3.3. Wa e Regime Desc ip ion
The explana ion o mos o he obse ed di e ences be ween he ada s and he buoys da a can
be clima e- ela ed o ela ed o he di e en measu ing and wa e spec a compu ing me hodology o
each de ice. To cla i y he incidence o hese causes, he wa e egime o he a ea has been analyzed
using a ailable wa e da a sou ces.
Coas al wa e egime, in con as o he p e ious alida ions, is based on he “ ime pai ed aw
samples” (Table 1), so all da a se s desc ibe he same sampling pe iod and size (9634 samples). Since
ada s nulls a e included, he wa e egimes desc ibed by hem ha e less in o ma ion han he o he
da a se s.
In gene al, he wa e oses shown in Figu e 1desc ibe h ee main wa e egimes: NW egime,
which includes all anges o wa e heigh s, and i is clea ly he dominan in his egion; he NNE egime,
wi h small and medium-size wa es; and he SW egime. Focusing on he speci ic da a si es, NNE wa es
a e only well desc ibed by VB, whose loca ion was exposed o be he Can ab ian sea, so any o he da a
sou ce loca ed sou he n om his posi ion will shade he NE wa es’ con ibu ion. Addi ionally, he NW
wa es change hei app oaching angle when ge ing close o he coas . Hence, o he da a sou ces such
as S20 and S02 desc ibe hese wi h a mo e wes e n di ec ion han VB. The e o e, S02 p ac ically does
no de ec NNE wa e sys ems and es ima es mos o he NW wa es as almos wes e ly. Rega ding he
SW wa e egime, whils he buoys only desc ibe small and medium-size wa es, he SIMAR poin s
show he occu ence o some high wa es. In con as , VILA clea ly unde es ima es bo h NNE and SW
wa e egimes, excep o a sa u a ion a i s sou he n CL. On he o he hand, SILL does no desc ibe
e en he no he ly wa es because i s no he n CL is limi ed o 350
◦
, and i unde es ima es he SW
wa e egime, excep , as well as VILA, o he sa u a ion a i s sou he n CL.
Al hough VILA shows a conside ably highe p opo ion o ex eme wa es han he o he s,
especially han he SIMAR poin s, his is mainly due o mos o he small wa es no being included
Remo e Sens. 2019,12, 598 16 o 24
Figu e 11.
Cama iñas me eo ological s a ion (CW) wind oses: (
a
) CW samples pai ed wi h VILA
SW wa e egimes. (
b
) CW samples pai ed wi h VILA wa es wi h Dm = 221
◦
—sa u a ion a VILA
sou he n CL.
The wa e egime alida ion pe o med on he Sillei o ada ou comes e ealed an agg ega ion
o la ge Hm0 and Dm e o s o a speci ic wa e egime (Hm0 > 4 m and Dm be ween 180
◦
–235
◦
,
Figu e A5) which co esponds o he ed poin s o he sca e plo s in Figu e A3a–c.
3.5. Reduc ion o Da a P ocessing Time
Figu e 12 shows he s a is ical compa ison o VILA and VB wa e pa ame e s, using 120 o 180 min
o CSS iles o calcula e he wa e spec um du ing VILA ope a ion. No e ha he di e ence be ween
bo h da a se s alida ion is less han 3% o all he wa e pa ame e s.
Valida ion o wa e pa ame e s o RC 10 km as a unc ion o he nine wa e egimes de ined in he
p e ious sec ion is shown in Figu e 13. Again, he di e ences obse ed o all he s a is ics a e small.
Howe e , p ocessing wi h a ime ange o 120 min inc eases by 3
%
he numbe o samples lagged
as null by he ada ’s so wa e.
Remo e Sens. 2019,12, 598 17 o 24
(c) (e)
( )
(d)
(b)
(a)
Figu e 12.
Compa ison be ween he alida ions o he wa e pa ame e s o he en Vilán ada (VILA)
ange cells wi h Vilano-Sisa gas buoy es ima es, p ocessed using 120 (blue ci cles) o 180 min ( ed
squa es) o c oss spec a (CSS) iles. (
a
) Hm0 co ela ion index (R); (
b
) Hm0 mean absolu e pe cen
e o (MAPE); (c) pe iod R; (d) pe iod MAPE; (e) Dm ci cula R; ( ) Dm mean absolu e e o (MAE).
(c) (e)
( )
(d)
(b)
(a)
Figu e 13.
Compa ison be ween he alida ions o he wa e pa ame e s o he 9 wa e egimes (WRs)
o VILA RC 10 km, p ocessed using 120 (blue ci cles) and 180 min ( ed squa es) o CSS iles. (
a
) Hm0
co ela ion index (R); (
b
) Hm0 mean absolu e pe cen e o (MAPE); (
c
) pe iod R; (
d
) pe iod MAPE; (
e
)
Dm ci cula R; ( ) Dm mean absolu e e o (MAE).
4. Conclusions
In his wo k, he accu acy o he wa e da a gene a ed by Vilán and Sillei o HF ada s (VILA,
SILL) ha e been ho oughly analyzed conside ing he cha ac e is ics o hei co e age a eas. Rada
c oss spec a iles om Jan 2014–Ap il 2015, we e ep ocessed using he same las upda e o he R7 o
CODAR SeaSonde so wa e. This so wa e imp o ed he quali y o he VILA-de i ed wa e pa ame e s
compa ed wi h hose es ima ed in si u wi h p e ious so wa e e sions [
30
]. Besides, an upda ed
sc eening me hod based on CMEMS ecommenda ions was implemen ed, which helped o de ec
ou lie s among ex ensi e pe iods o null da a.
Da a analysis e ealed ha VILA gene a ed nulls samples o all he wa es wi h spec al
signi ican heigh (Hm0)
≤
1 m; mos o he wa es <3 m; abou hal o wa es be ween 3–4 m; and abou
10
%
o hose be ween 4 and 6 m coming om he no h-no hwes . These pe cen ages o nulls inc eased
wi h he dis ance o he ada . SILL p oduced 20
%
ewe nulls han VILA, which mainly co esponded
Remo e Sens. 2019,12, 598 18 o 24
o wa es below 2 m. This as numbe o nulls canno be comple ely jus i ied by he HF ada
limi a ions, e en less so when SILL has signi ican ly ewe nulls han VILA. The e o e, o he ac o s
could be a ec ing signal ecep ion.
The analysis o he emaining no null HF ada s da a, e ealed ha VILA p oduces eliable wa e
pa ame e s o he i s six ange cells—un il 35 km—and SILL un il 55 km, imp o ing i s ag eemen
wi h Sillei o buoy, especially ega ding mean di ec ion (Dm), wi h he dis ance o he coas .
Fu he alida ion o VILA ange cell a 10 km e ealed ha he bes ag eemen wi h he buoy
occu s o wa es be ween 2 and 6 m coming om he no hwes (Hm0-R
≈
0.8) and ha he ada
ends o unde es ima e he bigges and he smalles wa es—Hm0>8 m and be ween 1 and 4 m
espec i ely. Rega ding Dm, VILA ends o allo mo e wes wa d di ec ions han he buoy o he NW
wa e egimes—mo e pe pendicula s o he sho eline—and he wa es ha he buoy es ima es wi h NE
di ec ion a e in e p e ed as NNW di ec ions by he ada .
The speci ic analysis o he VILA-de i ed wa e egimes e ealed ha wa es wi h Hm0 >
6 m and Dm be ween 330
◦
and 45
◦
—clockwise—co espond o spu ious da a. Addi ionally, VILA
sou hwes e ly wa es ha e a s ong ag eemen wi h buoy es ima es ega ding Hm0, bu in many cases,
he buoy es ima es NW wa e di ec ions. The analysis o he SILL-de i ed wa e egimes e ealed
ha wa es >4 m and Dm be ween 180
◦
and 235
◦
co espond o spu ious alues; howe e , ongoing
s udy has e ealed ha hese a e less a he u hes ange cells. Rega ding wa e pe iod, bo h ada s
unde es ima ed he buoys peak pe iods be ween 10 and 13 s, and SILL also o e es ima ed he peak
pe iods be ween 5 and 8 s.
The mos ele an di e ences be ween he es ima es o VILA and Vilano-Sisa gas buoy (VB) ha e
been explained by he cha ac e is ics o he wa e egime a he wes e n coas o Galicia. This has been
desc ibed using he wa e da a o bo h HF ada s, he wo buoys and h ee SIMAR poin s loca ed in he
a ea. The wa e egime is composed o wa es om NNE, NW and SW, wi h di e en anges o Hm0
and wi h a signi ican a iabili y along he coas , acco ding o hei app oaching angle and dis ance o
his. Fo ins ance, VILA’s co e age a ea is only pa ially open o he NE wa es—un il 41
◦
—and i s
p oximi y o he coas induces he pe pendicula i y—mo e wes wa ds—o he NW wa es. Mo eo e ,
as he SIMAR wa e pa ame e s and VB wa e spec a e ealed, in many occasions, hese wa e sys ems
can be simul aneous, leading o b oadband and bimodal sea s a es which a e no accu a ely ep esen ed
by he spec al wa e pa ame e s es ima ed ei he by he buoys o he ada s. In addi ion, he esul s o
he HF ada s wa e da a alida ion can be a ec ed when pa o he buoy’s wa e spec um is ou o
hei ange—ei he e y small wa e heigh o pe iod—bu also, because he HF ada -de i ed wa e
pa ame e s a e he a e age o he sea condi ions along each ange cell, while he buoys’ es ima es
a e he esul o measu ing a single poin . All hese ac o s also explain some di e ences be ween
VILA and SILL alida ion esul s: SILL’s coas is mo e s aigh han VILA’s and is in he shade o NE
wa es, so he a iabili y along i s ange cells is lesse han in VILA’s ones. Hence, he p oximi y o he
a hes ange cells o he buoy induces mo e simili ude ins ead o mo e disag eemen , as happens in
he VILA a ea.
This pape has es ablished ha VILA and SILL can educe by one hou , he calcula ion o he
wa e pa ame e s, by mean o using ewe samples du ing he model i ing, assuming an inc emen o
3% in he null pe cen age and a sligh dec ease in he co ela ion wi h he buoy es ima es.
The me hod desc ibed he e is a use ul ool o analyzing and classi ying he HF ada s’ wa e
ou comes, in o de o be in e p e ed and employed by inal use s. Addi ionally, i is use ul o
de ec ing whe he any adjus men should be done o he ada ; o example, o educe he high
pe cen age o nulls [39], o a oid he shade o ele an wa e egimes o o a oid he he e ogenei y o
he sea s a e condi ions h oughou he ange cell. Rega ding he las issue, he solu ion goes h ough
de eloping new so wa e able o es ablish di e en coas line limi s o each ange cell and sol ing
he wa e pa ame e s in shallow wa e s. Addi ionally, he de elopmen o mul i- equency ada s o
de ec ing accu a ely ei he small o high wa es is necessa y [
14
,
39
]. Finally, new so wa e upda es, as
desc ibed in Lipa [
24
], could ind he solu ion o sol ing complex sea s a es as bimodals. Addi ionally,
Remo e Sens. 2019,12, 598 19 o 24
he accu acy o e ie ing di ec ional wa e spec a could be imp o ed wi h many published, inno a e
me hods o o he HF ada models [5,40,41].
Au ho Con ibu ions:
Concep ualiza ion, A.B.; me hodology, A.B.; so wa e, A.B. and P.L.; alida ion, A.B. and
P.M.; in es iga ion, A.B.; esou ces, P.M. and E.Á.-F.; da a cu a ion, A.B.; w i ing—o iginal d a p epa a ion, A.B.;
w i ing— e iew and edi ing, V.P.-M.; supe ision, V.P.-M. All au ho s ha e ead and ag eed o he published
e sion o he manusc ip .
Funding:
This esea ch was unded by In e eg A lan ic A ea p ojec MyCOAST (EAPA 285/2016) and
INTERREG V-A Spain-Po ugal (POCTEP) p ojec RADAR
_
ON
_
RAIA co- unded by he Eu opean Regional
De elopmen Fund (ERDF) (EU). V.P.-M. and A.B. acknowledge inancial suppo by CRETUS s a egic pa ne ship
(ED431E2018/01), co- unded by he ERDF (EU) h p://www.usc.es/c e us/.
Acknowledgmen s:
The au ho s g a e ully acknowledge INTECMAR—Conselle ía do Ma —Xun a de Galicia
h ough RAIA Coas al Obse a o y o p o iding he VILA HF ada da a, Pue os del Es ado o p o iding
Sillei o HF ada da a and he SIMAR poin s and buoys da a (REDEXT).
Con lic s o In e es : The au ho s decla e no con lic o in e es .
Appendix A
In his sec ion he igu es wi h he mos ele an esul s o he alida ion be ween he wa e
pa ame e s es ima ed by Sillei o ada and Sillei o buoy a e exposed, as is a igu e wi h all he wa e
oses u ilized o compa ing he nine Vilán ada wa e egimes wi h he Vilán buoy wa e da a.
Figu e A1.
Pe cen age o Sillei o ada nulls a each combina ion o Dm and Hm0 ( esolu ion 1
◦×
1 m), o he ime pai ed da a o Sillei o buoy (SB).
Remo e Sens. 2019,12, 598 20 o 24
(c) (e)
( )
(d)
(b)
(a)
Figu e A2.
Valida ion o wa e pa ame e s o en Sillei o ada (SILL) ange cells (5–50 km) wi h
Vilano-Sisa gas buoy es ima es: (
a
) Hm0 co ela ion Index (R). (
b
) Hm0 mean absolu e pe cen e o
(MAPE). (c) Pe iod R. (d) Pe iod MAPE. (e) Dm ci cula R. ( ) Dm mean absolu e e o (MAE).
(b) (c)
Figu e A3.
Wa e pa ame e s sca e plo s o Sillei o ada (SILL) RC 10 km s. Sillei o buoy (VB). (
a
)
Hm0 eg ession line; R = 0.82; Slope = 0.93. (
b
) Reg ession line be ween VILA cen oid pe iod (Tc)
and VB peak pe iod (Tp); R = 0.70; slope = 0.67. (
c
) Dm sca e plo ; Ci c.R = 0.5.
Red poin s:
samples
ela ed o wa es >4 m and di ec ion be ween 180–235◦wi h e y la ge MAPEs.
Remo e Sens. 2019,12, 598 21 o 24
VILA NE2 VB Dm VB Dp
VILA NE4 VB Dm VB Dp
VILA NE6 VB Dm VB Dp
VILA NW2 VB Dm VB Dp
VILA NW4 BV Dm VB Dp
VILA NW6 VB Dm VB Dp
VILA SW2 VB Dm VB Dp
VILA SW4 VB Dm VB Dp
VILA SW6 VB Dm VB Dp
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
Figu e A4.
Wa e oses o pai ed clean samples o Vilán ada (VILA) wa e egimes (
1–9
) and
Vilano-Sisa gas buoy (VB). (
a
) VILA wa e egime. (
b
) VB mean di ec ion wa e ose. (
c
) VB peak
di ec ion wa e ose.
Remo e Sens. 2019,12, 598 22 o 24
Figu e A5.
Hm0 mean absolu e pe cen age e o (MAPE) be ween Sillei o ada (SILL) and Sillei o
buoy (SB) o each SILL wa e egime de ined by Hm0 and Dm (1 m ×5◦).
Re e ences
1.
Paduan, J.D.; Kim, K.C.; Cook, M.S.; Cha ez, F.P. Calib a ion and alida ion o di ec ion- inding
high- equency ada ocean su ace cu en Obse a ions. IEEE J. Ocean. Eng.
2006
,31, 862–875. [C ossRe ]
2.
Wya , L.R.; G een, J.; Middledi ch, A.; Moo head, M.D.; Howa h, J.; Hol , M.; Keogh, S. Ope a ional wa e,
cu en , and wind measu emen s wi h he Pisces HF ada . IEEE J. Ocean. Eng.
2006
,31, 819–834. [C ossRe ]
3.
Cochin, V.; Ma ie e, V.; Ga ello, R. Sea su ace cu en –wa e–wind in e ac ions measu ed by coas al g ound
wa e HF ada s. IEEE Geosci. Remo e Sens. Le . 2008,5, 227–230. [C ossRe ]
4.
Lipa, B.; Ba ick, D.; Alonso-Ma i ena, A.; Fe nandes, M.; Fe e , MI.; Nyden, B. The B ahan P ojec high
equency ada ocean measu emen s: Cu en s, winds, wa es and hei in e ac ions. Remo e Sens.
2014
,6,
12094–12117. [C ossRe ]
5.
López, G.; Conley, D.; G ea es, D. Calib a ion, alida ion and analysis i an empi ical algo i hm o he
e ie al o wa e spec a om HF ada sea echo. J. A mos. Ocean. Technol. 2016,33, 245–261. [C ossRe ]
6. C ombie, D. Dopple spec um o sea echo a 13.56 MHz. Na u e 1955,175, 681–682. [C ossRe ]
7.
Hasselmann, K. De e mina ion o ocean wa e spec a om Dopple adio e u n om he sea su ace.
Na . Phys. Sci. 1971,229, 16–17. [C ossRe ]
8.
Lipa, B.; Ba ick, D. Wa e P opaga ion Labo a o y. Analysis me hods o Na ow-Beam High-F equency
Rada Sea Echo. In NOAA Technical Repo ERL 420-WP 56; U.S. Dep . o Comme ce; Na ional Oceanic and
A mosphe ic Adminis a ion; En i onmen al Resea ch Labo a o ies: Boulde , CO, USA, 1982.
9.
He on, M.L.; Dex e , P.E.; Mc. Gann, B.T. Pa ame e s o he ai -sea in e ace by high- equency g ound wa e
Dopple ada . Ma . F eshwa e Res. 1985,36, 655–670. [C ossRe ]
10.
Wya , L.R. P og ess in he in e p e a ion o HF sea echo: HF ada as a emo e sensing ool. IEE P oc.
1990
,
137, 139–147. [C ossRe ]
11.
Chapman, R.D.; G abe , H.C. Valida ion o h ada measu emen s. Oceanog aphy
1997
,10, 76–79. [C ossRe ]
12.
Gu gel, K.W.; An onischki, G.; Essen, H.H.; Schlick, T. Wellen Rada (WERA): A new g ound-wa e HF ada
o ocean emo e sensing. Coas . Eng. 1999,37, 219–234. [C ossRe ]
13.
Lipa, B. In e sion o second—o de ada echoes om he sea. J. Geophys. Res.
1978
,83-C2, 959–962.
[C ossRe ]
14.
Lipa, B.; Nyden, B. Di ec ional wa e in o ma ion om he SeaSonde. IEEE J. Ocean. Eng.
2005
,30, 221–231.
[C ossRe ]
Remo e Sens. 2019,12, 598 23 o 24
15.
Fo ney, R.; Roa y, H.; Glenn, S. Measu ing Wa es wi h a Compac HF Rada . In P oceedings o he OCEANS
2015 - MTS/IEEE, Washing on, DC, USA, 19–22 Oc obe 2016.
16.
CODAR. Ocean Senso s SeaSonde
R
Remo e Uni Sys em Speci ica ion. Ve sión 6, Re ision 06/2017.
A ailable online: h p://www.coda .com/images/p oduc s/SeaSonde/1B-CODARspec_SSRS-100_ 6-
20170609.pd (accessed on 20 Janua y 2020).
17.
Wya , L.R. Limi s o he in e sion o HF Rada backsca e o ocean wa e measu emen . J. A mos. Ocean.
Technol. 1999,17, 1651–1665. [C ossRe ]
18.
Gu gel, K.W.; Essen, H.H.; Schlick, T. An empi ical me hod o de i e ocean wa es om second-o de b agg
sca e ing: P ospec s and limi a ions. IEEE J. Ocean. Eng. 2006,31, 804–811. [C ossRe ]
19.
Hisaki, Y. Quali y con ol o su ace wa e da a es ima ed om low signal- o-noise a io HF ada
Dopple -Spec a. J. A mos. Ocean. Technol. 2009,26, 2444–2461. [C ossRe ]
20.
Ba ick, D.; Lipa, B. When a e HF- ada obse ed wa e heigh s modula ed by pe iodic idal and ine ial
cu en s? In P oceedings o he IEEE/OES Ele e h Cu en , Wa es and Tu bulence Measu emen (CWTM),
S . Pe e sbu g, FL, USA, 2–6 Ma ch 2015.
21.
Ba ick, D.E. Ex ac ion o wa e pa ame e s om measu ed HF ada sea-echo Dopple spec a. Radio Sci.
1977,12, 415–424. [C ossRe ]
22.
Lipa, B.; Ba ick, D. Ex ac ion o sea s a e om HF ada sea echo: Ma hema ical heo y and modeling.
Radio Sci. 1986,21, 81–100. [C ossRe ]
23.
To oli, A.; Ono a o, M.; Monbaliu, J. Wa e s a is ics in unimodal and bimodal seas om a second-o de
model. Eu . J. Mech. B/Fluids 2006,25, 649–661. [C ossRe ]
24.
Lipa, B.; Daugha y, M.; Fe nandes, M.; Ba ick, D.; Alonso-Ma i ena, A.; Roa y, H.; Dicopoulos, J.; Whelan,
C. De elopmen s in Compac HF ada Ocean Wa e Measu emen s. In Physical Senso s, Senso Ne wo ks
and Remo e Sensing. Book Se ies: Ad ances in Senso s: Re iews,Vol. 5; Yu ish, S., Ed.; In e na ional F equency
Senso Associa ion Publishing (IFSA): Cas ellde els, Ba celona, Spain, 2018; Chap e 20, pp. 469–495.
25.
Kohu , J.; Roa y, H.; Lic henwalne , S.; Glenn, S.; Ba ick, D.; Lipa, B.; Allen, A. Su ace Cu en s and
Wa e Valida ion o a Nes ed Regional HF ada Ne wo k in he Mid-A lan ic Bigh . In P oceedings o he
IEEE/OES/CMTC 9 h Wo king Con e ence on Cu en Measu emen Technology, Cha les on, SC, USA,
17–19 Ma ch 2008.
26.
Long, R.; Ba ick, D.; La gie , J.; Ga ield, N. Wa e obse a ions om cen al Cali o nia: SeaSonde Sys ems
and in si u wa e buoys. J. Sens. 2011,2011, 1–18. [C ossRe ]
27.
A an, R.; Goggins, J.; Ha ne , M.; Nash, S.; Agos inho, P. Assessmen o Ex eme Wa e Heigh E en s in
Galway Bay Using High F equency Rada (CODAR) Da a. In P oceedings o he 1s In e na ional Con e ence
on Renewable Ene gies O sho e (RENEW), Lisbon, Po ugal, 24–26 No embe 2014.
28.
Falco, P.; Buonoco e, B.; Cianelli, D.; De Luca, L.; Gio dano, A.; Ie mano, I.; Kalampokis, A.; Sa iano, S.;
U ie i, M.; Zamba dino, G.; e al. Dynamics and sea s a e in he Gul o Naples: Po en ial use o
high- equency ada da a in an ope a ional oceanog aphic con ex . J. Ope a . Oceanog .
2016
,9(Suppl. 1),
33–45. [C ossRe ]
29. CODAR. The Unique Na u e o HF Rada . A ailable online: h p://www.coda .com/in o_h _ ada .sh ml
(accessed on 20 Janua y 2020).
30.
Lo en e, P.; Basañez Me cade , A.; Pied acoba, S.; Pé ez-Muñuzu i, V.; Mon e o, P.; So illo, M.G.;
Ál a ez-Fanjul, E. Long- e m skill assessmen o SeaSonde ada -de i ed wa e pa ame e s in he Galician
coas (NW Spain). In . J. Remo e Sens. 2019,10, 9208–9236. [C ossRe ]
31.
Al onso, M.; Ál a ez-Fanjul, E.; López, J.D. Compa ison o CODAR SeaSonde HF Rada Ope a ional Wa es and
Cu en s Measu emen s wi h Pue os Del Es ado Buoys; Final Repo o Pue os del Es ado; Pue os del Es ado:
Mad id, Spain, 2006.
32.
Cope nicus Ma ine in Si u Team. Cope nicus in Si u TAC, Real Time Quali y Con ol o WAVES; Cope nicus in
Si u TAC: Toulouse, F ance, 2017; pp. 1–19.
33.
CODAR. Con igu a ion File Fo ma s. A ailable online: h p://suppo .coda .com/ (accessed on 20 Janua y
2020).
34.
Pue os del Es ado. Conjun o de Da os REDEXT. A ailable online: h p://calipso.pue os.es/BD/in o mes/
INT_REDEXT.pd (accessed on 20 Janua y 2020).
35.
Pue os del Es ado. Conjun o de Da os SIMAR. A ailable online: h p://calipso.pue os.es/BD/in o mes/
INT_8.pd (accessed on 20 Janua y 2020).
Remo e Sens. 2019,12, 598 24 o 24
36.
Me eogalicia. His ó ico de la Red Me eo ológica. A ailable online: h ps://www.me eogalicia.gal/
obse acion/es acionshis o ico/his o ico.ac ion?idEs =10800 (accessed on 20 Janua y 2020).
37.
Sa iano, P.; Kalampokis, A.; Zambianchi, E.; U ie i, M. A yea -long assessmen o wa e measu emen s
e ie ed om an HF ada ne wo k in he Gul o Naples (Ty henian Sea, Wes e n Medi e anean Sea).
J. Ope a . Oceanog . 2019,12, 1–15. [C ossRe ]
38.
Teague, C.; Vesecky, J.; Fe nandez, D. HF Rada ins umen es, pas o p esen . Oceanog aphy
1997
,10-2,
40–44.
39.
Fo ge , P. Noise p ope ies o HF ada measu emen o ocean su ace cu en s. Radio Sci.
2015
,50, 764–777.
[C ossRe ]
40.
Lopez, G.; Conley, D. Compa ison o HF Rada ields o di ec ional wa e spec a agains in si u measu emen s
a mul iple loca ions. J. Ma . Sci. Eng. 2019,7, 271. [C ossRe ]
41.
Ha dman, R.; Wya , L. In e sion o HF Rada Dopple spec a using a neu al ne wo k. J. Ma . Sci. Eng.
2019,7, 255. [C ossRe ]
c
2019 by he au ho s. Licensee MDPI, Basel, Swi ze land. This a icle is an open access
a icle dis ibu ed unde he e ms and condi ions o he C ea i e Commons A ibu ion
(CC BY) license (h p://c ea i ecommons.o g/licenses/by/4.0/).
