IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, VOL. 17, 2024 3587
Sa elli e Rada Al ime y Suppo ing Coas al
Hyd ology: Case S udies o Guadalqui i Ri e
Es ua y and Eb o Ri e Del a (Spain)
Jesús Gómez-En i ,A.Alda ias ,R.Mule o-Ma ínez , S. Vignudelli ,M.B uno , R. Mañanes ,
A. Izquie do , and M. Fe nández-Ba ba
Abs ac —Coas al zones close o big i e sys ems a e hyd ody-
namically cha ac e ized by inland and open ocean wa e s. Sa elli e
al ime y da a ha e ecen ly been scien i ically exploi ed in coas al
zones. Howe e , mo e e o s a e s ill needed in e ms o using hese
da a in new applica ions in he coas al s ip. Ou s udy analyzes
he capabili y o he Eu opean Space Agency C yoSa -2 sa elli e o
measu e he sea le el ele a ion associa ed wi h bulge-like lens o
less sal y wa e s om i e discha ges, sp eading in o he adjacen
coas al zones. We analyze ou e en s o high i e eshwa e dis-
cha ges in wo i e es ua ies in Spain: Guadalqui i and Eb o. We
ob ain along- ack Absolu e Dynamic Topog aphy om C yoSa -2
sa elli e du ing hese e en s and compa e hem wi h pe iods o low
discha ge condi ions. In h ee o he e en s, he bulges inc ease he
sea le el be ween 5 and 10 cm wi h espec o he sea le el obse ed
in low / no mal discha ge condi ions. The ex ension o he lens is
a om he es ua y mou h: mo e han 25 km om he coas in he
case o he Guadalqui i Ri e and abou 50 km o he Eb o Ri e .
The wind egime and he su ace ci cula ion explain he lack o a
highe sea le el in one o he e en s analyzed.
Index Te ms—Absolu e dynamic opog aphy (ADT), coas al
hyd ology, C yoSa -2, i e discha ge, sa elli e al ime y.
I. INTRODUCTION
RIVERmou hsandes ua iesa eanimpo an elemen o he
clima e sys em, no only o ensu ing he link be ween he
a mosphe e and he ocean [1] bu also o a ec ing he dynamics
andbiogeochemis y o he adjacen a eas.F eshwa e discha ge
and incoming seawa e om he con inen al shel mainly con ol
hese a eas. Gene ally, i e discha ges ha e been adi ionally
igno ed as d i e s o sea le el a iabili y [2]; howe e , spo adic
bu high eshwa e discha ges migh con ibu e o he sea le el
Manusc ip ecei ed31July2023; e ised9No embe 2023and13Decembe
2023; accep ed 31 Decembe 2023. Da e o publica ion 4 Janua y 2024; da e o
cu en e sion 29 Janua y 2024. (Co esponding au ho : Jesús Gómez-En i.)
Jesús Gómez-En i, R. Mule o-Ma ínez, M. B uno, R. Mañanes, and A.
Izquie do a e wi h he Applied Physics Depa men , Uni e si y o Cadiz, 11510
Pue o Real, Spain (e-mail: [email p o ec ed]; [email p o ec ed];
[email p o ec ed]; [email p o ec ed]; [email p o ec ed]).
A. Alda ias is wi h he Physical Oceanog aphy Depa men , Span-
ish Ins i u e o Oceanog aphy (IEO-CSIC), 11006 Cadiz, Spain (e-mail:
[email p o ec ed]).
S. Vignudelli is wi h he Ins i u e o Biophysics (CNR), 16149 Pisa, I aly
(e-mail: [email p o ec ed].i ).
M. Fe nández-Ba ba is wi h he Depa men o Ecology and Coas al Man-
agemen , Ins i u e o Ma ine Science o Andalusia, 11510 Pue o Real, Spain
(e-mail: [email p o ec ed]).
Digi al Objec Iden i ie 10.1109/JSTARS.2024.3349692
a di e en imescales [3],[4],[5]. In addi ion, i e eshwa e
inpu s modi y he salini y and empe a u e, esul ing in a iable-
densi y su ace wa e s, which can gene a e an imbalance in he
p essu e g adien along he su ace o he ocean. The e o e,
hey can in luence he ci cula ion and coas al cu en s [6],[7],
[8]. Rega ding he biogeochemical ele ance o hese ansi ion
zones, hey a e biologically mo e p oduc i e han he i e s
and he adjacen oceanic wa e s as a esul o he p esence o
abundan suspended sedimen , o ganic and ino ganic ma e ials
[9]. Howe e , hey can also be he main sou ces o pollu an s
in o he ocean, educing wa e quali y, and a ec ing ac i i ies
such as ishe ies [10].
One e ec o i e plumes in coas al zones lies in he im-
pac o eshwa e discha ge on he sea le el posi i e anomaly
(bulge-like lens). This assumes ha he lens depends on i s
salini y anomaly wi h espec o he su ounding wa e s and o
heplumedep h[11].TheEa h’s o a ionmakes heou low u n
an i-cyclonic (No he n Hemisphe e), a o ing he de elopmen
o a coas al cu en [12]. This is pa icula ly in e es ing as he
wid h o hese bulges migh be se e al imes he wid h o he
coas al cu en . Fu he mo e, he sea le el di e ence be ween
helow-salini ylens and hesu oundingwa e scanbemeasu ed
by ada al ime e s on-boa d sa elli es as we will show in his
wo k. High eshwa e discha ges in adjacen a eas o he ocean
can also c ea e he so-called u bid plume due o he suspended
ma e ial discha ged. The di e en colo o he u bid plume wi h
espec o he su ounding wa e s can be obse ed wi h op ical
senso s as p e iously shown by Fe nández-Nó oa e al. [13] and
e e ences he ein.
Moni o ing i e mou hs and es ua ies is essen ial o ully
unde s and he dynamics o coas al zones and comp ehend how
hey will be a ec ed in u u e clima e scena ios. Fo ha eason,
he e is a g owing in e es in he scien i ic communi y in using
addi ional in o ma ion o complemen in si u and model da a.
As men ioned, his is he case o using al ime e da a o he
hyd odynamic cha ac e iza ion o coas al zones a ec ed by bo h
inland and open ocean condi ions, especially in a eas close o
la ge i e sys ems. The idea o a coas al hyd ology al ime y
communi y combines common in e es s o coas al and inland
a eas. The ad en o new al ime e concep s, such as he delay-
Dopple al ime y (DDA) [14], also known as syn he ic ape u e
ada al ime y (SAR), wi h imp o ed p ecision and accu acy,
can play a key ole in be e exploi a ion o hese da a close o he
© 2024 The Au ho s. This wo k is licensed unde a C ea i e Commons A ibu ion-NonComme cial-NoDe i a i es 4.0 License. Fo mo e in o ma ion, see
h ps://c ea i ecommons.o g/licenses/by-nc-nd/4.0/
3588 IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, VOL. 17, 2024
sho e. This is he main mo i a ion behind some ini ia i es, such
as he Eu opean Space Agency’s unded p ojec : “SAR/SARin
Rada Al ime y o Coas al Zone and Inland Wa e Le el” (Hy-
d ocoas al: h ps://www.sa oc.eu/p ojec s/hyd ocoas al/) aimed
a he exploi a ion o SAR/SARin DDA da a (C yoSa -2 and
Sen inel-3A/B missions) in he coas al zone and inland wa e s.
Some o he ‘his o ical’ limi a ions o coas al/inland ada al-
ime y [15],[16] ha e been mos ly o e come, hanks o he
e o s made in p e ious yea s by space agencies and esea ch
communi ies. Fo a mo e comple e lis o s udies ocused on
coas al/inland wa e s using sa elli e al ime y da a, he eade is
e e ed o some ecen wo ks, such as [17],[18],[19], and e -
e ences he ein. I is also wo h men ioning he ex ensi e wo ks
donein helas decades oanalyze hee ec o i e plumesin he
sea su ace salini y a ia ions using passi e mic owa e senso s
(e.g., ESTAR, STARRS, SMOS, Aqua ius, and SMAP). Mo e
de ailed in o ma ion on his is gi en in pape s [20],[21], and
e e ences he ein. Ano he in e es ing aspec is he combina ion
o sea su ace salini y om hese senso s and sea su ace heigh
om sa elli e al ime y [21] bu his is ou o he scope o ou
wo k.
The main objec i e o his a icle is o analyze he capabili y
o sa elli e ada al ime y o measu e he e ec o high i e
discha ges o eshwa e in he coas al zones close o i e
mou hs in e ms o sea le el a iabili y. The adi ional lack o
in si u long- ime se ies o sea le el measu emen s associa ed
wi h hese e en s, makes he use o accu a e al ime e da a
a powe ul ool o hei analysis and in e p e a ion. A ull
cha ac e iza ion o hese low salini y plumes using all a ailable
sa elli e, and in si u obse a ions o he sea su ace salini y and
empe a u e (among o he a iables) du ing pe iods o high i e
eshwa e discha ges,isou o hescopeo hisa icle.Howe e ,
some auxilia y ma e ial (including op ical image y) was used
o a be e in e p e a ion o he esul s ob ained. The analysis
ocused on wo la ge Spanish i e s: Guadalqui i and Eb o.
The o me lows in o he A lan ic Ocean and he la e in o he
Medi e anean Sea. To do his, we analyzed he spa io empo al
a iabili yo heseale el(o equi alen lychangesinalong- ack
Absolu e Dynamic Topog aphy: ADT, he eina e ) o he closes
acks o he i e es ua y mou hs a ailable wi hin he C yoSa -2
mission (CS2, hence o h).
The es o his a icle is o ganized as ollows: Sec ion II is de-
o ed o he s udy a eas, ollowed by Sec ion IIIwi h heda ase s
and me hod; he esul s a e p esen ed in Sec ion IV, and he
discussion and conclusions in Sec ions Vand VI, espec i ely.
II. STUDY AREAS
The e a e six majo i e s in he Ibe ian Peninsula: Miño,
Due o, Tajo, Guadiana, Guadalqui i , and Eb o. Acco ding o
Fe nández-Nó oa e al. [13], he Guadalqui i and Eb o Ri e s
(hence o h, GR and ER, espec i ely) showed he wides u bid
plume ex ensions h oughou he wa e yea . This was obse ed
unde bo h low and high i e discha ge condi ions. Hence,
assuming he in e se co ela ion be ween high u bidi y and low
salini y du ing high discha ge e en s, he GR and ER sys ems
we e selec ed as s udy a eas. This included conside ing he
Fig. 1. Loca ion o s udy a eas (a): The Guadalqui i Ri e Es ua y (GRE) in
he sou hwes e n coas o Spain (b): he ed do in he small igu e inse shows he
posi ion o he in si u i e discha ge s a ion, and he Eb o Ri e Del a (ERD) in
he no heas e n (c), wi h he ed do gi ing he posi ion o he discha ge s a ion.
The black squa e in (b) shows he loca ion o he GRE mou h. Ba hyme y (in m)
ex ac ed om [22].
TABLE I
MAIN CHARACTERISTICS OF GUADALQUIVIR AND EBRO RIVERS
es ua y mou h and su ounding wa e s in he A lan ic Ocean and
he Medi e anean Sea, co esponding o hei a eas o in luence
(see Fig. 1).
A. Guadalqui i Ri e Es ua y (GRE)
The GR is one o he main i e s in Spain, being he only
na igable i e in he egion. Some o i s main cha ac e is ics
a egi eninTableI.The GREs ua y(hence o h GRE)isloca ed
in he sou hwes o he Ibe ian Peninsula. The 5 km wide mou h
[seeFig.1(b)] ispa o he eas e nsub-basin o he Gul o Cadiz
(GoC, he eina e ), which connec s he A lan ic Ocean and he
Medi e anean Sea. As he main con ibu o o eshwa e in he
GoC, he high and spo adic eshwa e discha ges migh a ec
he sea le el a di e en imescales [3],[4],[5],[23]. This plays
an impo an ole in he eas e n GoC su ace ci cula ion [24].
Someo he i e s lowin o heGul bu wi hasigni ican lylowe
discha ge (mainly he Tin o-Odiel Ri e sys em and Guadiana
Ri e ). As p e iously no ed in he In oduc ion, high discha ge
e en s dec ease he salini y and he plume o less dense wa e
p oducesanele a iono hewa e le el.Thishasp e iouslybeen
no ed by some au ho s in he GRE using in si u measu emen s
o a ide-gauge loca ed in he es ua y mou h [7],[25],[26].
Somee o sha ealsobeenmadein in es iga ing hesee en s
wi h sa elli e-al ime e da a. Using A chi ing, alida ion, and
GÓMEZ-ENRI e al.: SATELLITE RADAR ALTIMETRY SUPPORTING COASTAL HYDROLOGY 3589
in e p e a ion o sa elli e oceanog aphic da a (AVISO), sea le el
anomaly g idded (mon hly) p oduc s (1/3° ×1/3° o g id cell
esolu ion) and a ide gauge loca ed in he es ua y mou h, [23]
obse ed a s ong e ec in he in si u measu emen s due o high
eshwa e discha ges. Howe e , no signals we e obse ed in
he closes al ime e poin o he es ua y mou h. Gómez-En i
e al. [3] used a dedica ed g idded (1/16° ×1/16° o g id size)
dailyp oduc [27] and ound e idence o mesoscalesigna u esin
he es ua y mou h due o high i e eshwa e discha ges. They
concluded ha he spa io empo al dis ibu ion o he acks used
o compu e he g idded p oduc is c ucial in i s abili y o obse e
mesoscalesignals.Mo e ecen ly,Gómez-En ie al.[4]analyzed
one e en o la ge uno and i s e ec on he sea le el anomaly
a ia ions using one CS2 ack close o he es ua y mou h. They
obse ed a dis inc i e signal ha could be due o he eshwa e
discha ge. Ou wo k aims o con i m his by using mo e CS2
acks o his e en by ex ending he analysis o one mo e e en
in he GRE and including wo e en s in he ERD.
B. Eb o Ri e Del a (ERD)
The ER Del a (ERD, hence o h) is loca ed on he Medi e -
anean coas o he Ibe ian Peninsula. Mo e han 350 ibu a ies
con ibu e o he i e basin along i s 970 km o leng h (see
Table I). The i e discha ges p esen a high in e annual a i-
abili y. Though he ocean ci cula ion nea he ERD is domina ed
by along-sho e la ge-scale dynamics [28], ep esen ed by he
Ligu o-P o ençal-Ca alan cu en , he buoyancy inpu in o-
duced by la ge i e eshwa e discha ges esul s in he de elop-
men o ci cula ion ea u es. Such ea u es include an icyclonic
eddies a ound he del a which, a he same ime, ein o ce he
gene al ci cula ion o he sou h [28],[29],[30]. The sou hbound
ci cula ion d i es he i e u bid plume owa ds he sou h o
he del a, whe e i can be ound mo e han 70% o he ime
[13]. Mo eo e , since he a ea is cha ac e ized by a mic o- idal
egime, he ERD plume ex ension and loca ion is no a ec ed by
idal oscilla ions [31]. The u bid plume was analyzed in de ail
by Fe nández-Nó oa e al. [13] using op ical sa elli e image y.
To he bes o ou knowledge, he e a e no consolida ed pape s
dealing wi h he e ec o high i e discha ges on he sea le el
a iabili y a ound he ERD mou h.
III. DATASETS AND METHODS
F om all he a ailable pas and p esen sa elli e al ime e
missions, CS2 was selec ed as i is he only sa elli e wi h acks
c ossing he s udy a eas close enough o he es ua y mou hs
du ing a ime pe iod long enough o cap u e a numbe o
e en s (CS2 da a a e a ailable o mo e han 10 yea s since
he beginning o i s mission). Along- ack da a a high pos ing
a e (numbe o along- ack measu emen s pe second) we e
selec ed ins ead o g idded p oduc s o wo easons: 1) he
spa io empo al dis ibu ion o he acks used o c ea e he maps
could comp omise he capabili ies o de ec ine-scale signals
(da a close o he coas a e usually lagged ou ); and 2) he lack
o a ba hyme y cons ain and smalle co ela ion scales in he
mapping p ocess, as p e iously no ed by Gómez-En i e al. [3].
TABLE II
INFORMATION ABOUT THE CS2 TRACKS DURING THE EVENTS SELECTED
ACCORDING TO TRACK AVAILABILITY AND RIVER DISCHARGE
In his sec ion, he da ase s (Sec ions III-A o III-F) and
me hods (Sec ion III-G) used a e p esen ed. Al ime y da a
(Sec ion III-A) and in si u i e discha ges (Sec ion III-B)a e
used o isola e he ime pe iods o high discha ges, and analyze
hei e ec s on sea le el a iabili y. Some in o ma ion abou
hese e en s is a ailable in Table II. Auxilia y ma e ial is also
used o suppo and in e p e ou indings. The compa ison be-
ween he ex ension o he low-salini y lens and he u bid plume
ismadeusingop ical image y(Sec ion III-C).Absolu edynamic
opog aphyandsu acecu en da a omahyd odynamicmodel
(Sec ion III-D) and model-based winds (Sec ion III-E) a e used
o make a quali a i e assessmen o he sea le el a ia ions ob-
se ed by CS2 da a, and also o discuss he esul s ob ained. This
is also he case o he ide gauge wa e le els (Sec ion III-F).
A. C yoSa -2 Da a
The Eu opean Space Agency CS2 mission began in Ap il
2010 a e he launch ailu e o i s p edecesso in 2005. The
sa elli eope a esinanonsunsynch onousquasi-pola e og ade
o bi (inclina ion espec o he equa o ial plane o 92°) a a
mean al i ude o 717 km and a empo al esolu ion o 369 days
(30-day sub-cycle). The mission objec i es we e ocused on he
pola zones [32], bu u he da a exploi a ion has allowed o he
applica ions in open ocean, coas al zones and inland wa e s.
The main ins umen , he Delay-Dopple SAR/In e e ome ic
Rada Al ime e -2 (SIRAL) was designed o ope a e in h ee
modes:syn he icape u e ada (SARM),low esolu ion(LRM),
and SAR-in e e ome ic (SARIn). A geog aphical mode mask
de ines he acquisi ion mode o ope a ion o SIRAL a any ime.
Ou s udy a eas ha e been co e ed by he SARM since he
beginning o he mission [33].
In SARM, he e ec i e oo p in size a e pos -p ocessing
is abou 380 m in he along- ack di ec ion ( his could change
as he al i ude a ies), ega dless o he wa e heigh , being
abou 1.5 km in he ac oss- ack di ec ion [33]. This imp o es
he ins umen pe o mance in e ms o educed noise in he
es ima ion o he geophysical pa ame e s and imp o ed ange
p ecision, bu also a educed a ea subjec o land/calm wa e
con amina ion in he wa e o m signal as he sa elli e app oaches
he land. A ull desc ip ion o he SAR mode o ope a ion is
a ailable in [34] (among o he s).
The p oduc e sion used in his wo k was om he Baseline
D. The majo imp o emen espec o he p e ious e sion was
he change o he sel -desc ibing Ne CDF o ma . The CS2
3590 IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, VOL. 17, 2024
Fig. 2. C yoSa -2 acks (do ed lines) analyzed a he GRE mou h o he i s
(a) and second (b) e en s. The same a he ERD o E en 3 (2.c) and E en 4
(2.d). The solid lines o e he acks show he segmen o ack analyzed in each
case. The black do a he GRE mou h (a) and (b) gi es he posi ion o he ide
gauge.
acks in he s udy a ea we e ob ained om [35].Weused he
p ede ined p ocessing se -up o coas al zones (see [34], o
mo e de ails), and ex ac ed along- ack da a a 20 Hz o pos ing
a e ( his co esponds o abou 350 m be ween consecu i e
measu emen s). Fig. 2gi es he acks used and hei loca ion in
he s udy a eas o he e en s analyzed: wo acks o E en
1 and wo mo e o E en 2 a GRE [see Fig. 2(a) and (b),
espec i ely] and wo acks o E en 3 and one o E en 4
a ERD [see Fig. 2(c) and (d)]. In all he cases, he selec ion
o he acks was made acco ding o he a ea o in luence o
he uno i e plume (du ing high discha ge condi ions). This
was assuming ha in gene al e ms, he plume sp ead adially
om he mou h wi h he Co iolis o ce de lec ing i o he igh
in he No he n Hemisphe e ( o an obse e looking seawa d),
o ming an an icyclonic bulge (also known as eddy) [13],[36]
and a coas al bounda y cu en [37].
B. Ri e Discha ge
A e age daily eshwa e discha ge (Qd) was ob ained om
wosou ces:1)Guadalqui i Ri e :Thenea es s a ionisloca ed
a Alcalá del Río dam (abou 108 km ups eam) wi h da a
a ailable in he Au oma ic Sys em o Hyd ological In o ma ion-
Guadalqu i [38]; and 2) Eb o Ri e : To osa s a ion (42 km
ups eam) a ailable in he ASHI-Eb o [39].
High discha ge e en s in he s udy a eas we e selec ed, ac-
coun ing o he pe iods wi h a high low egime (abou 99 h
pe cen ile o discha ge) and o he a ailabili y o CS2 acks in
he es ua y mou h o he i e s du ing hese e en s. In he high
eshwa e discha ge egimeandunde somecondi ions, he low
egime p e en s sal wa e in usion, allowing he gene a ion o
he plume [40]. As men ioned, we analyzed he a ailabili y o
CS2 o bi s du ing he e en s in bo h a eas om Ap il 2010 o
Decembe 2019 (see Fig. 2). A summa y wi h some de ails
ega ding he CS2 o bi s and Qdis gi en in Table II.The
emaining high discha ge e en s obse ed in he analyzed pe iod
(including hose spo adicin ime)we e no included inou wo k,
due o he una ailabili y o CS2 acks.
C. Op ical Image y
Red-g een-blue (RGB) ue colo (co ec ed e lec ance)
Aqua and e a mode a e esolu ion imaging spec o adiome e
(MODIS) sa elli e images we e ob ained om [41]. Aqua/Te a
MODIS sa elli es we e selec ed due o hei a ailabili y du ing
he analyzed pe iod. The pixel size o he images is 250 m; hey
we e selec ed conside ing hei da e o acquisi ion and he da e
o CS2 c ossing he a ea o in e es . The le el o cloud co e
was also used as an image selec ion c i e ion.
D. Ibe ian Biscay I ish (IBI) Model
The Ibe ian Biscay I ish (IBI) ocean eanalysis model p o-
ideshou ly alueso zonalandme idional eloci ycomponen s
and ADT (among o he a iables) a 50 le els ( om –5698.06 o
–0.51 m). The IBI model is based on he Nucleus o Eu opean
Modelling o he Ocean (NEMO) 3.6 ocean gene al ci cula ion
model (1/12° ho izon al esolu ion). NEMO assimila es sa elli e
(al ime e and in a ed senso s), in si u ( empe a u e and salini y
e icalp o iles)da a and idal ele a ions p o idedby woglobal
models: FES2004 [42] and TPXO7.1 [43]. I also includes daily
i e discha ges as la e al open bounda y condi ions o 33 i e s
[44],[45]. Daily discha ges ha e been ob ained om [38] and
[46]. The IBI ou pu s a e eely a ailable a [47]
E. Wind Da a
The IBI model is o ced by hou ly 10 m wind ields om he
i h-gene a ion Eu opean Cen e o Medium-Range Wea he
Fo ecas s a mosphe ic eanalysis o he global clima e model a
0.33° ×0.33° spa ial esolu ion (ERA5) [44].WeusedERA5
in he analysis o he compa ison o he ou pu s o he model
and CS2 in e ms o sea le el a iabili y, o assess he wind ield
condi ions du ing he discha ge e en s.
F. Tide Gauge Da a
In si u wa e le els a hou ly in e als we e ob ained om
[48]. The Bonanza ide gauge is loca ed a he mou h o GRE
[see Fig. 2(a) and (b)]. This is a MIROS ada ins umen and
he hou ly da a ha e been ob ained by a e aging he quali y-
con olled p oduc p o ided by Pue os del Es ado a 5-min
in e als. I is pa o he Spanish Red de Ma eóg a os, which
is in he Pe manen Se ice o Mean Sea Le el and Global Sea
Le el Obse ing Sys em.
G. C yoSa -2 Absolu e Dynamic Topog aphy (ADT) P o iles
The along- ack ADT was ob ained ollowing (1)
ADT = Al i ude −Range −(RangeCo ec ions
+ GeophysicalCo ec ions) −Geoid (1)
GÓMEZ-ENRI e al.: SATELLITE RADAR ALTIMETRY SUPPORTING COASTAL HYDROLOGY 3591
TABLE III
SAMPLING INTERVALS (SPACE AND TIME)OF THE PRODUCTS USED
whe e Al i ude is he dis ance be ween he sa elli e’s cen e
o mass and he e e ence ellipsoid (WGS84); Range is he
e acked ange SAMOSA+ e- acke , which is he SAMOSA
2 ailo ed o applica ion in he open ocean, coas al zone and ice
[49]; Range Co ec ions s ands o We /D y T oposphe ic Co -
ec ions and Ionosphe ic Co ec ion; Geophysical Co ec ions
emo e Tides: TPX09 [43], In e se Ba ome e , high- equency
a mosphe icsignals (a mosphe icp essu eand winds),SeaS a e
Bias Co ec ion (5% o Signi ican Wa e Heigh as in [4]. Las ly,
he geoid used was EIGEN-6C4 [50]. All he inpu alues in
(1) a e ob ained om he P-PRO se ice [35]. A e compu ing
ADT, high- equency noise was emo ed by applying a unning
mean il e o 7 km along he ack. This is app oxima ely he
spa ial esolu ion o he NEMO model a ou la i udes.
A summa y o he sampling in e als o he p oduc s used in
his wo k is gi en in Table III.
IV. RESULTS
In his sec ion, we p esen he main esul s ob ained in he
wo s udy a eas in e ms o sea le el a iabili y associa ed wi h
high i e discha ges. Some auxilia y ma e ial is also shown
o suppo he main indings. Bu i s , a quan i a i e analysis
o wo o he da ase s used is p esen ed. We es ima ed he
Pea son’s co ela ion coe icien be ween he ADT om CS2
and IBI (in e pola ed o he CS2 along- ack posi ions) o all
he acks analyzed. We ound a high co ela ion o 0.94 (99%
o con idence le el).
A. Guadalqui i Ri e Es ua y (GRE) Mou h
1) E en 1: Fig. 3shows he daily a e age eshwa e dis-
cha ge om he dam ±30 days o he da e o he i s sa elli e
ack (#0088) c ossing he s udy a ea. The ±30 day pe iod
was selec ed in o de o ully cap u e he a iabili y o he
discha ge be o e and a e he sa elli e c ossings. Two acks
(one ascending and one descending) c ossed he ou e limi s
o he es ua y wi h a di e ence in ime be ween hem o 4 o
5 days (see Table II). Two high discha ges (highe han he
99 h pe cen ile limi ) we e obse ed du ing Decembe 2010 [see
Fig. 3(a)]. Un o una ely, due o he limi ed a ailabili y o CS2
acks in he a ea, we could ocus only on he second one (less
in ense). No high i e discha ges we e obse ed in he es o
Fig. 3. Times se ies (60 days) o daily a e age discha ges om he Alcalá del
Río dam in he uppe limi o he GRE du ing he da es o CS2 o e passes (a)
o ( ) o he i s e en analyzed. The dashed e ical lines indica e he da es
o he CS2 o e passes o ack #0088 (black line), and ack #0153 (g ey line),
espec i ely. The ho izon al black dashed line shows he 99 h pe cen ile limi .
Fig. 4. Along- ack ADT du ing he i s e en o ack #0088 (a) and #0153
(b) o six 369-days CS2 passes o he same o bi . The da es o he sa elli e
passes o e he a ea a e also shown.
he pe iods analyzed [see Fig. 3(b)–( )] conside ing he da es o
a ailable CS2 passes o he wo acks analyzed.
This e en is pa icula ly in e es ing as i eached high dis-
cha ge (>99 h pe cen ile limi ) du ing 7 o he 30 days be o e
he i s sa elli e pass c ossed he a ea. Thus, he p esence o
a la ge i e plume would be expec ed in he es ua y mou h
sp eading om he es ua y mou h in o he gul . Fig. 4gi es he
along- ack ADT on he da es analyzed o ack #0088 [see
Fig. 4(a)] and #0153 [see Fig. 4(b)]. I should be no ed ha
some passes o he acks (unde low discha ge condi ions) we e
no a ailable in P-PRO and we e no included in his analysis.
Absolu e dynamic opog aphy shows only weak in e annual
a iabili y in bo h acks in mos o he pe iods analyzed, excep
o da es: Decembe 27, 2010 (#0088) and Janua y 1, 2011
(#0153). Du ing hese da es, he ADT is abou 5 o 10 cm highe
han he ADT p o iles obse ed in he o he sa elli e passes. This
bulge o wa e obse ed in ADT migh be ela ed o he lens o
eshwa e o he i e . In ac , he signal associa ed wi h he
i e eshwa e plume is obse ed a away om he es ua y
mou h, up o 25 km om he coas as shown by he loca ion
o he descending ack #0153 [see Fig. 2(a)], and he highe
ADT obse ed du ing he eshwa e discha ge [black line in
Fig.4(b)].The eason o heADTele a ionobse edonJanua y
2, 2013 (#0088) emains unclea as a low i e discha ge was

3592 IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, VOL. 17, 2024
Fig. 5. Red-G een-Blue op ical images o he closes da es o he sa elli e
passes du ing E en 1. The da es o he op ical image selec ion we e made
conside ing he cloud co e : Decembe 26, 2010 (a) and Decembe 29, 2010
(b). Whi e solid line shows he ack segmen s analyzed (#0088: Decembe 27,
2010 and #0153: Janua y 1, 2011).
obse ed days be o e he sa elli e pass. A ema kable dec ease
in ADT o ack #0153 on Janua y 21, 2018 is also obse ed.
Two MODIS RGB images we e selec ed du ing his e en
conside ing he closes da es o he sa elli e passes: Decembe
26, 2010 [see Fig. 5(a)] and Decembe 29, 2010 [see Fig. 5(b)].
The selec ion o hese da es was made acco ding o he cloud
co e o he images. The u bid plume is clea ly seen in bo h
cases, being mo e e iden on Decembe 26. The main u bid
plume obse ed in he op ical image y is de lec ed o he igh
as expec ed due o he ac ion o Co iolis accele a ion. The ack
segmen s a e inside he a ea co e ed by he u bid plume, bu
no a i s co e. He e, we assumed he co espondence be ween
he eshwa e bulge (obse ed in he ADT p o iles o he CS2
sa elli e) and he a ea co e ed by he u bid plume (con ain-
ing suspended and dissol ed ma e ials) as p e iously no ed
by Fe nández-Nó oa e al. [13] (and e e ences he ein). As
men ioned, he ADT signal obse ed ou side he co e o he
u bid plume migh indica e ha he eshwa e bulge ex ends
a away om he es ua y mou h o e en mo e han 25 km o
he coas . The dec ease in he i e discha ge 6 days be o e he
sa elli e pass #0153 (and hence a elaxa ion o he o cing ac o )
could explain he dec ease in ADT obse ed in he no he n
sec o o ha acksegmen wi h espec opass#0088.Howe e ,
his could also be due o he ac ha #0088 is close o he coas .
Aquali a i eassessmen o heseale elele a ionobse edby
he ada al ime e can be made wi h he ADT o he IBI model.
The maps o daily ADT o he model be ween Decembe 25,
2010 and Janua y 2, 2011 a e shown in Fig. 6. The selec ed
a ea was wide han p e iously shown in Fig. 1in o de o
ge a mo e gene al iew o he ocean a iabili y. The dec ease
in ADT obse ed in CS2 on Janua y 1, 2011 wi h espec o
Decembe 27, 2010 is clea ly obse ed in he model esul s.
I should be no ed ha a di e en geoid e e ence is used in
he model; his explains he di e ence in he magni ude o ADT
obse edbe weenbo hda ase s. Apiling-up o wa e isobse ed
on he coas no h o he es ua y mou h on mos o he days. This
could be a s ong e idence o he p esence o he plume. I is
known ha s ong and pe sis en eas e ly winds can con ibu e
o he piling-up o wa e on he coas o he s udy a ea [24],
[52]. We ob ained ERA5 daily 10 m wind speed and di ec ion
Fig.6. Absolu edynamic opog aphy (in cm) o he IBI model om Decembe
25, 2010 o Janua y 2, 2011 (a) o ( ). “CS2” (c) and (h) highligh s he ADT-IBI
maps on he da es he sa elli e c ossed he s udy a ea ( acks #0088 and #0153,
espec i ely). The whi e squa e in (a) shows he loca ion o he GRE mou h.
Fig. 7. Same as Fig. 6, bu o ERA5 daily 10 m wind.
in he same ime pe iod as in ADT-IBI. The esul s a e shown
in Fig. 7. Nei he s ong no pe manen eas e lies ha e been
obse ed, so he wind egime i sel does no seem o explain he
sea le el inc ease obse ed by CS2 and he model da a du ing
he discha ge e en .
2) E en 2: The daily eshwa e discha ge ±30 days o
he i s sa elli e pass ( ack #1319) is shown in Fig. 8.In his
case, high discha ges a e close o he sa elli e passes a ailable
GÓMEZ-ENRI e al.: SATELLITE RADAR ALTIMETRY SUPPORTING COASTAL HYDROLOGY 3593
Fig. 8. Same as Fig. 3, bu o E en 2 analyzed in he GRE mou h conside ing
he da es o CS2 acks: #1319 (black e ical dashed line) and #1384 (g ey
e ical dashed line).
Fig. 9. Same as Fig. 4, bu o E en 2 o acks #1319 (a) and #1384 (b).
in Ma ch / Ap il 2013 [see Fig. 8(b)]. This e en eached high
discha ge (>99 h pe cen ile limi ) du ing 4 o he 30 days be o e
he da e o he sa elli e pass. I is o less in ensi y han E en 1,
so he ex en o he eshwa e plume migh be expec ed o be
lowe . We also obse ed spo adic bu small discha ges in Ap il /
May 2014 [see Fig. 8(c)] and Ap il 2018 [see Fig. 8(e)], bu hey
a e con ined o a sho pe iod o ime and well be o e he da es
o he sa elli e passes. Fo his eason, hey will no be analyzed.
The ADT p o iles o ascending ack #1319 [see Fig. 9(a)]
show a highe in e annual a iabili y wi h espec o descending
ack #1384 [see Fig. 9(b)]. The sea le el signal associa ed
wi h he eshwa e discha ge measu ed du ing Ap il 2013 is
clea ly obse ed in he ADT co esponding o he ascending
ack [Ma ch 28, 2013 in Fig. 9(a)]. This is no he case o he
ADT ou days a e o he descending ack [Ap il 1, 2013:
Fig. 9(b)]. As men ioned, some passes o he acks we e no
a ailable in P-PRO [35] o his e en . The op ical RGB images
show ha he ex en o he u bid plume is lowe han he plume
in E en 1 (see Fig. 10). The bes image close o he da e o he
ascending ack was a ailable ou days be o e he sa elli e pass
[see Fig. 10(a)]. The u bid plume is clea ly de eloped wi h he
CS2 ack close o i . The image o Ap il 1, 2013 [see Fig. 10(b)]
indica es ha he descending ack #1384 is a away om he
u bid plume, so he bulge could also be a away om he ack
posi ion. This could explain he lack o a bulge-like signal in he
ADT p o ile. This will be u he discussed in Sec ion V.
Fig. 10. Same as Fig. 5, bu o E en 2. C yoSa -2 o e lew he a ea ou days
a e he da e o he MODIS image shown in (a) (Ma ch 24, 2013) and he same
day o he image shown in (b) (Ap il 1, 2013).
Fig. 11. Same as Fig. 6, bu o E en 2. “CS2” highligh s he images acqui ed
on he da es he sa elli e c ossed he s udy a ea ( acks #1319 and #1384,
espec i ely).
Figs. 11 and 12 show he same analysis as in Figs. 6and 7,bu
o E en 2 in e ms o ADT-IBI and ERA5-wind, espec i ely.
In his case, a less in ense bulge o wa e is clea ly obse ed
close o he es ua y mou h om Ma ch 25 h o 31s . Howe e ,
i is mo e con ined in o he mou h and sligh ly di e ed o he
sou h. This con i ms he di e en ADT eco ds displayed by
acks #1319 (Ma ch 28, 2013) and #1384 (Ap il 1, 2013). The
wind egime o ERA5 indica es p e ailing wes e lies, which
a o s he sou h-eas wa d ci cula ion o e he con inen al shel
[52]. This could explain he sligh de lec ion o he plume o he
sou h.
B. Eb o Ri e Del a
Ou esul s o GRE sugges a posi i e co ela ion be ween
high i e discha ges ecei ed by coas al a eas close o he es u-
a y and a sea le el ele a ion obse ed by CS2 ADT along- ack
p o iles. We comple e ou analysis wi h ERD, as i shows a
wide u bid plume unde high eshwa e discha ge [13].I is
3594 IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, VOL. 17, 2024
Fig. 12. Same as Fig. 7, bu o E en 2.
Fig. 13. Same as Fig. 3 o E en 3 (b) and 4 (e) in he ERD. Black e ical
dashed line deno es he da es o CS2 c ossing he a ea o ack #1152 and g ey
line o ack #1478.
hus a good candida e o con i m he e ec o high discha ges
on coas al sea le el a iabili y close o he del a. Conside ing
he a ailabili y o CS2 acks a ound he ERD and he da es o
high discha ges, wo e en s ha e been iden i ied: Ma ch/Ap il
2013 ( acks #1152 and #1478; E en 3) and Ap il 2018 ( ack
#1478; E en 4).
Six y days o daily discha ges o his i e a e gi en in Fig.13.
They a e less in ense wi h espec o GRE. Howe e , in some
cases, he daily egime is close o, o e en exceeds, he 99 h
pe cen ile (Figs. 13(b): E en 3 and 13(e): E en 4). In some
o he cases, high/mode a e discha ges a e obse ed a ew days
be o e/a e he sa elli e passes [see Fig. 13(c) and (d)], so hey
will no be conside ed o analysis.
Fig. 14 shows he along- ack ADT p o iles o bo h acks
du ing hese e en s. The highe ADTs co espond o he pe iods
wi h high discha ges ( he E en 3 o bo h acks) con i ming
Fig. 14. Absolu e dynamic opog aphy p o iles in he ERD o E en 3: T ack
#1152 (a) and ack #1478 (b) and E en 4: T ack #1478 (b).
he esul s ob ained in GRE. The ADT signal o ack #1478
on Ap il 23, 2018 (E en 4) shows a s ong inc ease a he sea
le el in he middle o he ack segmen ; his migh indica e he
ou e bounda y o he plume co esponding o he limi be ween
he less saline plume and he su ounding saline wa e s. The
con i ma iono hiswouldbepossiblewi hsuppo ingda a(e.g.,
in si u measu emen s) bu o he bes o ou knowledge hey a e
no a ailable o his speci ic e en , and also is ou o he scope
o his a icle. The lowe ADT can be obse ed in he sa elli e
passes du ing pe iods o low o mode a e i e discha ges (bo h
e en s).
The op imal RGB image, in e ms o cloud co e , was ob-
ained on Ma ch 14, 2013 (no shown he e) o sa elli e pass
#1152 (E en 3), and i shows a u bid plume con ined in he
del a mou h. The signal obse ed in he ADT shows a wide
ex ension o he eshwa e low-salini y plume. This con i ms
ou p e ious esul s o GRE, which showed a wide ex ension
o he eshwa e bulge wi h espec o he op ical plume. The
RGB image co esponding o pass #1478 (same e en ) was ha
o Ap il 6,2013.I shows awide u bid plumein hedel amou h
de lec ing o he sou h, bu i is a away om he sou he nmos
loca ion o he ack. The RGB image o Ap il 22, 2018 du ing
E en 4 (no shown) also p esen s a signi ican ly wide u bid
plume de lec ing o he sou h.
Absolu e dynamic opog aphy o he IBI model (see Fig. 15)
showssignalscompa ible wi h he i e plumesgene a ed du ing
E en s 3 and 4, con i ming he esul s shown by CS2 ADT da a
and op ical images. Howe e , he model desc ibes a lowe plume
on Ap il 7, 2013 [see Fig. 15(e)] espec o Ma ch 16, 2013
[see Fig. 15(b)]. S ong no hwes e ly Mis al (no shown he e)
a o s he widening o he plume du ing Ma ch (E en 3). The
weakening o he plume on Ap il 7, 2013 can be explained by
a weakening wind egime. E en 4 is s onge in e ms o i e
discha ge [see Fig. 13(e)] bu sho e in ime wi h espec o
E en 3. This could explain he lowes signal obse ed in he IBI
model ou pu s he day be o e he CS2 c ossing [see Fig. 15(g)]
du ing E en 4, wi h espec o he days be o e E en 3 [see
Fig. 15(a) and (d)]. As men ioned, he plume is de lec ed o he
sou h du ing bo h e en s. This is due o he Co iolis o ce and he
Ligu o-P o ençal-Ca alan cu en . The sou he n de lec ion o
heplume ein o ces hegene alci cula ion o hesou hobse ed
by [28],[29],[30].
GÓMEZ-ENRI e al.: SATELLITE RADAR ALTIMETRY SUPPORTING COASTAL HYDROLOGY 3595
Fig. 15. Same as Fig. 6 o E en 3 (a)–( ) and 4 (g)–(i). CS2 s ands o he
da e o he sa elli e c ossing he s udy a ea: T acks #1152 (b) and #1478 (e) and
(h).
V. DISCUSSION
Along- ack one-dimensional da a esol e ime and wa e-
leng h scales o he ocean signal sho e han hose esol ed
in g idded wo–dimensional maps, mainly due o he smoo hing
and in e pola ion p ocesses made o gene a e he maps. [11]
analyzed maps o ADT a ound he Ob-Yenisei plume (Ka a Sea
in he A c ic Ocean) and concluded ha hei use migh gi e a
misleading in e p e a ion o he esul s due o synop ic (wind
o cing) and seasonal (plume mixing wi h su ounding wa e s)
a iabili y. [53] combined di e en sa elli e-de i ed p oduc s
(sea su ace salini y and empe a u e, and al ime y measu e-
men s) o cha ac e ize he Sibe ian Lena Ri e in he ice- ee
pe iod be ween 1993 and 2020. Howe e , hey used maps o
ADT in hei analysis ins ead o along- ack p oduc s.
Ou s udy a eas ha e di e en geog aphical se ings, bu he
esul s (in e ms o along- ack ADT om CS2) demons a e
ha he bulge due o he high discha ge om he i e s has a
dis inc i e signal in he ADT p o iles, which a e o he same
o de o magni ude a bo h loca ions. One eason ha could
explain his is he ac ha bo h con inen al shel s a e simila in
e ms o ex ension and dep h. The e is one excep ion o his o
ack #1384 in he GRE mou h ( his will be discussed u he ).
Ano he aspec o in e es is ha he use o CS2 da a e ealed
ha he ex ension o he low-saline plumes is wide han he
op ic plumes a bo h loca ions. A ull cha ac e iza ion o he
li e ime and ex ension o he bulges could be made by inc easing
he numbe o acks om o he sa elli e al ime e missions. As
men ioned, he ime pe iod analyzed p ecluded he use o o he
sou ces o al ime e da a.
We discuss now possible explana ions o he beha io ob-
se ed in he ADT p o iles du ing E en s 1 and 2 in he GRE
mou h and ERD. Wi h ega ds o GRE mou h, he loca ion o
he descending acks: #0153 (E en 1) and #1384 (E en 2)
Fig. 16. Two days o hou ly in si u sea le el o Bonanza2 ide gauge s a ion
(blue line). Red do indica es he in si u wa e le el alue a he closes ime
be o e hesa elli eo e passes.Tidegaugemeasu emen a 05:00(17minbe o e)
(a) and a 12:00 (34 min be o e) (b). All imes a e in UTC.
in he es ua y mou h is ba ely he same. The along- ack ADT
p o ile o ack #1384 (Ap il 1, 2013) does no show a signi ican
a ia ion (in e ms o inc eased sea le el) despi e he p esence
o he eshwa e plume in he su ounding a ea due o he high
i e discha ge. The maps o ADT om he IBI model con i med
his. Acco ding o Sousa e al. [54] and Cy iac e al. [55], he e
a e se e al ex e nal agen s a ec ing he plumes. We di ided
hem acco ding o: 1) e en -speci ic agen s: Ri e discha ge,
wind s ess, ides, cu en s; and 2) none en -speci ic agen s:
Ou low ine ia, o a ion e ec s, bo om opog aphy, alignmen
o he coas line, and i e mou h shape. We analyzed only (1)
assuming ha he agen s in ol ed in (2) simila ly a ec all he
e en s analyzed.
The e ec o he i e discha ge was analyzed in e ms o he
amoun o eshwa e discha ged in o he GoC du ing he e en s.
Table IV shows he discha ge accumula ed du ing wo weeks o
bo he en s.Theaccumula eddischa gein hese wope iodsiso
simila magni ude, so his canno explain he di e en beha io
obse ed in he CS2 ADT p o iles.
Wind di ec ion and in ensi y we e analyzed in p e ious sec-
ions. Wes e lies (landwa d low) a e dominan du ing he da es
o he sa elli e pass on Janua y 1, 2011 [see Fig. 6(h)] and Ap il
1, 2013 [see Fig. 12(h)], bu hey a e mo e in ense in he second
case. This migh a ec he ex ension o he plume. Acco ding
o [55] hese landwa d winds migh con ine he plume o he
coas line,andinou case,dependingoni sin ensi ycouldde lec
i o he sou h. This is wha we obse ed and could pa ially
explain he di e en beha io in he ADT p o iles [see Figs. 4(b)
and 9(b)].
The impac o he idal cycle in he ex ension o he plume
is analyzed in he nex lines. The idal egime in he GoC is
meso idal, so i is expec ed o ha e an in luence in he plume
ex ension unde some condi ions. High / low ides a o smalle
/ la ge plume u bid ex ensions [13].Fig.16 shows wo days
o hou ly sea le el measu ed by he ide gauge loca ed a he
es ua y mou h [see Fig. 2(a) and (b)]. The idal phase is qui e