The role of the Magellan Strait on the southwest South Atlantic shelf

Abstract

We analyze historical hydrographic data together with recent high-resolution hydrographic and underwater glider observations collected in the Magellan Strait to determine the water mass characteristics and exchanges between the continental shelves of the southeast South Pacific and southwest South Atlantic around southern South America. The near-surface salinity distribution and water mass analyses indicate a strong interoceanic connectivity associated with diluted subantarctic waters of the Pacific Ocean through the Magellan Strait which in turn are further diluted largely by inflows from the Almirantazgo Fjord via the Whiteside Channel. The lowest salinity waters reach the Atlantic shelf via the Magellan Strait. This core of low salinity waters (S < 31.5) are observed on the northern portion of the strait's Atlantic mouth and creates a strong baroclinic signal that leads to intense eastward geostrophic velocities suggesting a net Pacific to Atlantic transport (0.038–0.074 Sv, 1 Sv = 10⁶ m³ s⁻¹). This flow plays a significant role on the thermohaline characteristics and extent of the subantarctic shelf waters that occupy the Atlantic shelf. The low salinity inflow from the Magellan Strait combines with saltier inflows through the Le Maire Strait and farther east that feed the Atlantic shelf.

Full text

Estuarine, Coastal and Shelf Science 237 (2020) 106661
Available online 26 February 2020
0272-7714/Published by Elsevier Ltd.
The role of the Magellan Strait on the southwest South Atlantic shelf
Anahí A. Brun
a
,
b
,
c
,
*
, Nadin Ramirez
d
,
e
, Oscar Pizarro
d
,
e
,
f
, Alberto R. Piola
a
,
b
,
c
a
Departamento de Oceanografía, Servicio de Hidrografía Naval, Argentina
b
Universidad de Buenos Aires, Argentina
c
Consejo Nacional de Investigaciones Cientícas y T�
ecnicas, Argentina
d
Instituto Milenio de Oceanografía (IMO), Concepci�
on, Chile
e
COPAS Sur-Austral, Universidad de Concepci�
on, Chile
f
Departamento de Geofísica, Universidad de Concepci�
on, Chile
ARTICLE INFO
Keywords:
Pacic ocean
Atlantic ocean
Interoceanic uxes
Magellan Strait
Surface salinity
ABSTRACT
We analyze historical hydrographic data together with recent high-resolution hydrographic and underwater
glider observations collected in the Magellan Strait to determine the water mass characteristics and exchanges
between the continental shelves of the southeast South Pacic and southwest South Atlantic around southern
South America. The near-surface salinity distribution and water mass analyses indicate a strong interoceanic
connectivity associated with diluted subantarctic waters of the Pacic Ocean through the Magellan Strait which
in turn are further diluted largely by inows from the Almirantazgo Fjord via the Whiteside Channel. The lowest
salinity waters reach the Atlantic shelf via the Magellan Strait. This core of low salinity waters (S <31.5) are
observed on the northern portion of the strait’s Atlantic mouth and creates a strong baroclinic signal that leads to
intense eastward geostrophic velocities suggesting a net Pacic to Atlantic transport (0.038–0.074 Sv, 1 Sv ¼10
6
m
3
s
1
). This ow plays a signicant role on the thermohaline characteristics and extent of the subantarctic shelf
waters that occupy the Atlantic shelf. The low salinity inow from the Magellan Strait combines with saltier
inows through the Le Maire Strait and farther east that feed the Atlantic shelf.
1. Introduction
The Atlantic and Pacic continental shelves around southern South
America present distinct morphological and oceanographic character-
istics. The coast west of the Andes was broken down by quaternary
glaciation forming a vast archipelago extending from Chilo�
e Island
(41�47’ S) to Cape Horn (56�S) (Rabassa and Clapperton, 1990). At its
widest point near 43�S the Pacic shelf is only about 70 km wide
(Camus, 2001). In contrast, the Atlantic shelf is hundreds of km wide,
exceeding 800 km at about 51�S and only about 10 km wide southeast of
Tierra del Fuego (Parker et al., 1997).
The SE Pacic is inuenced by seasonal and latitudinal changes in
precipitation, which is nearly entirely controlled by the regional distri-
bution and intensities of the southern hemisphere westerly winds.
Furthermore, the morphology of this region contribute to a strong
annual mean precipitation ranging from 5000 to 10000 mm yr
1
(e.g.
Aracena et al., 2015), causing signicant continental runoff. Excess
precipitation over the ocean and continental runoff substantially dilute
the subantarctic water (Davila et al., 2002; Silva et al., 2009) derived
from the northern portions of the Antarctic Circumpolar Current.
Consequently, south of about 35�S the Pacic continental shelf and
neighboring open ocean are dominated by low surface salinity waters (S
<33.8; Deacon, 1937; Silva and Neshyba, 1977; Schneider et al., 2003).
As most of the humidity contained in the westerlies is lost in the up-
stream side of the Andes, the land east of the Andes is semiarid and the
continental freshwater runoff from the 1500 km coastline to the Atlantic
shelf south of 45�S is <1800 m
3
s
1
(see Global Runoff Data Center,
2009; http://www.bafg.de/GRDC/). In addition, the wide continental
shelf on the Atlantic side of South America receives very limited pre-
cipitation north of about 50�S and is dominated by excess evaporation
(~35 cmyr
1
; Uppala et al., 2005). Interestingly, however, and similar
to the Pacic shelf, the Atlantic shelf is also dominated by relatively low
salinities compared to the subantarctic waters found along the shelf
edge.
The source of low salinity waters observed over the Patagonian
Atlantic Shelf has been associated with a ow from the Pacic primarily
* Corresponding author. Departamento de Oceanografía, Servicio de Hidrografía Naval, Argentina.
E-mail addresses: anahibrun@gmail.com (A.A. Brun), nadin.ramirez@imo-chile.cl (N. Ramirez), oscar.pizarro@imo-chile.cl (O. Pizarro), apiola@hidro.gov.ar
(A.R. Piola).
Contents lists available at ScienceDirect
Estuarine, Coastal and Shelf Science
journal homepage: http://www.elsevier.com/locate/ecss
https://doi.org/10.1016/j.ecss.2020.106661
Received 8 July 2019; Received in revised form 20 February 2020; Accepted 22 February 2020

Estuarine, Coastal and Shelf Science 237 (2020) 106661
2
through the Magellan Strait and to lesser extent to the Le Maire Strait
(Bianchi et al., 1982; Panella et al., 1991; Guerrero and Piola, 1997;
Valdenegro and Silva, 2003; Combes and Matano, 2014). Climatological
sea surface salinity distributions suggest that the Magellan Strait are a
source of low salinity waters, which are presumably advected north-
eastward by the so-called Patagonian Current (Brandhorst and Castello,
1971). Realistic numerical simulations suggest that the northward
spreading of low salinity waters from the Magellan Strait is controlled by
the combined effects of the wind, tides and the western boundary cur-
rents (Palma and Matano, 2012). This tongue of low salinity seems to
play a signicant role in the oceanographic properties over the conti-
nental shelf off southern Argentina. Though the link between the
northern Drake Passage and the western boundary current along the
slope of the Argentine Basin is well established (e.g. Piola and Gordon,
1989; Peterson and Whitworth, 1989), the sources of the waters over the
continental shelf are still very poorly understood. The aim of this work is
to determine the water mass characteristics at the “entry points” of the
Pacic to Atlantic inows using a combination of historical and recent
hydrographic data and high resolution underwater glider measurements
within the Magellan Strait.
2. The Magellan Strait: background
The Magellan Strait (hereafter MS) separates continental South
America from the island of Tierra del Fuego. The MS is relatively narrow
and elongated channel that extends about 560 km from its western
mouth in the Pacic near 52.6�S to the relatively wide and shallow
mouth on the Atlantic side near 52.5�S (Fig. 1). The strait width is highly
variable; there are three well-dened narrow passages at Carlos III Is-
land (CI), Segunda Angostura (SA) and Primera Angostura (PA) a few
km wide. These constrictions and the distinct physiographic character-
istics divide the MS in three sub basins: the western basin spans from the
Pacic mouth to CI is up to 1100 m deep; the central basin from CI to SA,
a broad and deep V-shaped channel that presents maximum depths of
550 m and the eastern basin spans from SA to the Atlantic mouth, is
about 70 m deep (Panella et al., 1991; Valdenegro and Silva, 2003).
The tidal regime prevailing in the MS is mixed, mainly semidiurnal,
on the Pacic side (mean range 1.2m, Fierro, 2008) and semidiurnal on
the Atlantic side (mean range 7.1m). On the eastern basin the tidal wave
propagates westward with sharply decreasing ranges after PA and SA
(<1.2 m)(Medeiros and Kjerfve, 1988).The western mouth has large
annual rainfall between 2000 and 5000 mm. These rain and melt waters
deliver substantial amounts of freshwater to the interior zone (~6500
m
3
s
-1
, Global Runoff Database). This input together with the contri-
bution melting of glaciers in spring and summer also produce an
important input of fresh water into the Straits (Panella et al., 1991; Lutz
et al., 2016) that are presumably transported towards the eastern mouth
(Sassi and Palma, 2006).
3. Data and methods
This study is based on the analysis of historical hydrographic data
obtained from World Ocean Database (grey crosses in Fig. 1a and Boyer
et al., 2013) together with three hydrographic surveys: two
high-resolution CTD sections occupied across the eastern mouth of the
MS (GEFPAT-3, Charo and Piola, 2014, and ANTXXV-4, Provost, 2010,
Fig. 1. a) Location of historical hydrographic data
(grey crosses), CIMAR16 cruise collected in the
Magellan Strait in spring 2010 (red circles), un-
derwater glider sections collected in the central
basin of the Magellan Strait in June 2011 (black
and magenta lines), high-resolution hydrographic
stations occupied in the Atlantic mouth of the
Magellan Strait and across the Le Maire Strait
during the GEF III in September 2006 and
ANTXXV-4 in March 2009 cruises (green circles).
The black circle in the northern portion of the
former indicates the location where the ADCP
prole was collected (station 8 of ANTXXV-4
cruise), and the black star indicates the location
of the current meter. The background shading and
contours indicate display the bottom topography in
meters from GEBCO. Acronyms listed from Pacic
towards Atlantic: SPi: Southern Patagonia iceeld,
Skyring Sound (SS), Otway Sound (OS), CI: Carlos
III Island, DI: Dawson Island, CDi: Cordillera Dar-
win iceeld, WC: Whiteside Channel, AF: Almir-
antazgo Fjord, IB: Inútil Bay, PA: Punta Arenas, Pa:
Primera Angostura, Sa: Segunda Angostura, BC:
Beagle Channel, LM: Le Maire Strait, EI: Estados
Island. b) Bottom prole (m) along the Magellan
Strait derived from soundings and GEBCO and
location of stations collected during the CIMAR 16
cruise. (For interpretation of the references to
colour in this gure legend, the reader is referred to
the Web version of this article.)
A.A. Brun et al.

Estuarine, Coastal and Shelf Science 237 (2020) 106661
3
green circles in Fig. 1a) and one along-strait section (CIMAR 16, CONA,
2011, red circles in Fig. 1a). The latter survey is used to characterize the
along-channel property distributions while the two high-resolution
surveys were used to characterize the water mass properties and quan-
tify the inow to the Atlantic.
The GEFPAT-3 cruise was carried out on board R/V ARA Puerto
Deseado as part of the Global Environment Facility (GEF) Patagonia
project. The MS section was occupied on 17 September 2006 and a Le
Maire Strait section on 19 September 2006. The section consisted of
twelve full-depth hydrographic and nutrient stations across the Atlantic
mouth of the MS. At each station a vertical quasi-continuous CTD prole
was collected with a Sea-Bird Electronics model 911plus unit. The CTD
was mounted together with a Sea-Bird Electronics SBE32 carrousel with
24-bottle, each bottle of 5 L capacity (Charo and Piola, 2014). Water
samples were collected at selected levels for the determination of
salinity, dissolved oxygen and nutrients.
The ANT-XXV/4 cruise was carried on R/V Polarstern (Provost,
2010). The MS section was occupied on 22 March 2009 and consisted of
a 9-station section across the eastern mouth of the strait. Vertical proles
of horizontal velocity were collected only in the two northernmost sta-
tions (8 and 9) of this section. In addition, a Le Maire strait section was
occupied on 24 March 2009. At each station, hydrographic proles were
obtained with a Seabird CTD 911plus. Water samples at different levels
were collected using a Sea-Bird Electronics SBE32 carrousel with 24
bottle positions tted with 22 Niskin bottles, each of 12 L capacity.
Water samples were collected at selected levels for the determination of
salinity, dissolved oxygen, alkalinity and nutrients.
The CIMAR 16 cruise was carried on board R/V Abate Molina as part
of the CIMAR Program (Marine Research Cruises in Remote Areas) be-
tween 20 October and 13 November 2010 (CONA, 2011). The cruise
consisted of fteen stations occupied along the axis of the MS and a
section along the Almirantazgo Fjord (AF and red circles in Fig. 1a). At
each station, a vertical quasi-continuous CTD prole was collected with
a Sea-Bird Electronics model 911plus unit which reached 800 m depth.
The CTD was mounted with a 12-bottle water sampler holding 12 L
Niskin bottles. Water samples were collected at selected levels for the
determination of dissolved oxygen and nutrients.
We also used data from an underwater glider survey carried out on
the central part on MS (Fig. 1a). The glider was deployed near Mansa
Bay (53�370S, 70�530W) on June 24, 2011, for approximately 11 days.
The survey covered a large portion of the central basin with unprece-
dented resolution between 53.1�S off Punta Arenas and 53.8�S, west of
Dawson Island (DI in Fig. 1a). The tidal currents in this portion are
relatively weak (commonly <10 cm s
1
), allowing the proper operation
of underwater gliders. The glider missions collected 650 temperature
and salinity proles on three cross-channel (black circles in Fig. 1a) and
one along-channel section (magenta line in Fig. 1a).
Temperature, salinity and density data from the CIMAR 16 cruise are
used to analyze the oceanographic conditions along a Pacic-Atlantic
section along the MS and Whiteside Channel (WC in Fig. 1a). Bathym-
etry proles were prepared based on data of Panella et al., 1991 and
navigational chart 12550 of the Chilean Hydrographic Institute, by
plotting the best approximate depth at approximately 1 nautical mile
intervals along the channel axis. Both sections were contrasted against
the 30 arc sec gridded bathymetry from the General Bathymetric Chart
of the Ocean (GEBCO_2014 Grid, version 20150318, www.gebco.net,
Weatherall et al., 2015).
In addition, we use daily Optimal Interpolation sea surface temper-
ature (SST) data obtained from NOAA NCDC (Reynolds et al., 2007). To
minimize the effect of the frequent cloud cover in the region we used the
combination of Advanced Very High Resolution (AVHRR) and Advanced
Microwave Scanning Radiometer (AMSR). These data are available on a
0.25�latitude x 0.25�longitude grid over the period June 2002–October
2011. To characterize the temperature variability in the central and
eastern basins of the MS we spatially averaged four 0.25 �x 0.25�grid
points in each basin: between 68.5 and 69.5 �W in the eastern basin and
between 52.75 and 53.5 �S in the central basin. Annual climatologies
were estimated based on the time average between 1 January 2003 and
31 December 2010. Surface wind stress during the occupation of the two
high-resolution hydrographic sections in the eastern mouth of the MS
are derived from Cross-Calibrated Multi-Platform (CCMP) winds avail-
able at NOAA’s Pacic Marine Environmental Laboratory, Live Access
Server (https://ferret.pmel.noaa.gov/uaf/las/).
4. Results
4.1. Regional salinity distribution
The near-surface long-term mean salinity distribution (Fig. 2) is a
useful indicator of extent of the impact of the inter-oceanic connectivity.
Continental runoff is a key contributor to the low salinity waters of the
continental shelf of southern Chile, thus, signicant seasonal variations
associated with snow and glaciers melting in summer are expected.
However, due to the scarcity of observations only an annual mean dis-
tribution is presented. The seasonality however will be discussed at
specic locations where more data are available. We choose the 20 m
depth salinity to prepare and describe the horizontal salinity distribu-
tion, but the overall patterns do not differ signicantly from the distri-
butions at shallower levels except very near inputs of continental
discharge of freshwater. Various altimeter derived mean surface dy-
namic topographies and surface drifter-based circulation analyses indi-
cate that south of 40�S the narrow continental shelf of southern Chile is
fed by eastward owing upper layer subantarctic waters from the ocean
interior (e.g. Maximenko et al., 2009; Lumpkin and Johnson, 2013, and
references therein). The subantarctic waters are sharply diluted by the
large excess precipitation and continental runoff. Consequently, the
lowest salinities (S <29) observed around the southern tip of South
America are located over the shelf, particularly at the numerous fjords
found along the coast of southern Chile (Fig. 2). The Skyring and Otway
sounds (SS and OS in Figs. 1a and 2), located near 52.5�S are sources of
the lowest salinity waters (S <29) to the western mouth of the straits.
The regional salinity distribution further reveals an additional low
salinity inow within the MS, at Almirantazgo Fjord (AF in Figs. 1a and
2).
The continental shelf south of Tierra del Fuego and the Beagle
Channel are also characterized by low salinity waters (<32.5), partic-
ularly within the Beagle Channel, where input of freshwater from the
Cordillera Darwin iceeld (CDI in Fig. 1a) lead to salinities as low as
29.5 (Fig. 2). Further details on the water mass structure within the MS
are based on the analyses of high-resolution hydrographic observations
described in Section 3.3. For a more detailed description of the hydro-
graphic characteristics of the Chilean fjord region the readers are
referred to Valdenegro and Silva (2003) and Palma and Silva (2004).
On the Atlantic shelf Fig. 2 reveals a well-dened low salinity plume
(S <31.5) emanating from the northern side of the MS mouth. The
plume extends northward close to the coast, gradually becoming more
salty. Along this path the MS plume is reinforced by the contribution of
the Santa Cruz river (~690 m
3
s
1
, BDHI, http://www.bdhi.hidric
osargentina.gov.ar/), which creates a narrow and elongated coastal
band with salinities lower than 32.5 and extending northeastward
nearly 250 km from the river mouth. Waters fresher than ~32.75 reach
Cape Blanco at 47.2�S (Fig. 2). North of Cape Blanco the coast describes
a sharp westward turn forming the San Jorge Gulf, while the plume
continues its northeastward extension and thus separates from the coast.
On average the MS plume is clearly identied in by near surface salinity
<33.5 to about 42.75�S, while further north there is a wide mid-shelf
region with salinities lower than 33.7. At this latitude an outow from
the northern San Matias Gulf creates a wedge of salty (S >34.1) near-
coastal waters (Scasso and Piola, 1988; Lucas et al., 2005). Mixtures
with the signicant discharge of freshwater from the Rio de la Plata onto
the shelf at around 35�S precludes unequivocally identifying the MS
plume north of about 39�S based on the salinity distribution.
A.A. Brun et al.

Estuarine, Coastal and Shelf Science 237 (2020) 106661
4
4.2. Water masses
The temperature – salinity diagrams from high-quality observations
collected in austral summer and winter in the Pacic and Atlantic
mouths of the Magellan Strait and in the Le Maire Strait (LM) (Fig. 3)
reveal the distinct hydrographic characteristics of each of these regions.
In the western basin of the MS only waters less dense than
σ
θ
~24.7 kg
m
3
were included because, as will be shown later, the bottom topog-
raphy appears to block the exchange of denser waters with the central
basin (see Fig. 4c and Panella et al., 1991). In summer the western
(Pacic) side the MS presents the lowest surface salinities (~28) and
substantial thermal and haline stratication (open blue symbols in
Fig. 3). As pointed out earlier, the freshest upper-layer waters observed
in the western MS are primarily derived from the Skyring and Otway
sounds (Fig. 2). There are few winter observations in the western MS and
in Fig. 3 we only included two stations collected in early October 1998,
which may be somewhat warmer and fresher than in the peak of the
winter (solid blue symbols in Fig. 3). These observations present upper
layer waters about 2.5 �C colder than in summer and have similar sa-
linities. The seasonal temperature stratication vanishes for waters
denser than
σ
θ
~24.5 kg m
3
.
In contrast, the eastern mouth of the MS presents a much larger
seasonal temperature range (10 �C in summer versus ~ 4.2 �C in winter)
compared to the Pacic mouth and only a moderate temperature strat-
ication (orange symbols in Fig. 3). Throughout the year a relatively
wide salinity range (31.10–~32.57) is observed in the eastern mouth. As
will be shown in the following section the freshest waters in the Atlantic
side occupy the northern sector of the MS mouth (also see Fig. 2). In
summer the freshest waters observed in the LM (open green symbols in
Fig. 3) present nearly identical characteristics to the saltiest waters
observed in the MS (~32.5 and 9.2 �C), suggesting a continuous near-
coastal path between these regions, despite the >350 km distance be-
tween the straits. In contrast, in winter the LM waters are substantially
colder than the MS (~6.2 �C) and present no apparent salinity changes
compared to the summer observations. Thus, the connectivity inferred
from the summer observations is not apparent in winter. The preceding
water mass analysis clearly indicates that the MS is the source of the
lowest salinity waters entering the Atlantic Patagonia continental shelf.
4.3. Magellan Strait
The inter-ocean transition of water masses within the MS is further
described based on along-strait vertical sections of temperature, salinity,
and density spanning the entire extent MS from the Pacic to the
Atlantic oceans (Fig. 4), the underwater glider sections in the central
basin (Fig. 5) and hydrographic data collected within the Almirantazgo
Fjord (Fig. 6). The analysis focuses on the upper 400 m of the water
column because, as shown below, topographic features block the ex-
change of deep waters between basins. The temperature and salinity
sections along the MS (Fig. 4a and b) depict a relatively sharp front close
to the Pacic entrance, associated with the inow of cold, low salinity
waters from local continental discharge, which lead to a temperature
and salinity decrease from 8 to 7 �C and from ~32 to ~29, respectively.
The low salinity waters observed in the western basin have been referred
to as estuarine waters (Sievers and Silva, 2008). Relatively warm-salty
Subantarctic waters are found in the western basin below an intense
Fig. 2. Climatological salinity distribution at 20 m around the southern tip of South America based on historical and recent hydrographic data. Also shown are the
station locations (grey crosses) and the 200 and 2000 m isobaths (heavy grey lines). ChI: Chilo�
e Island, Skyring Sound (SS), Otway Sound (OS), AF: Almirantazgo
Fjord, IB: Inútil Bay, VP: Vald�
es Peninsula, SMG: San Matías Gulf.
A.A. Brun et al.

Estuarine, Coastal and Shelf Science 237 (2020) 106661
5
halocline (see Valdenegro and Silva, 2003; Vargas et al., 2018). The
penetration of these dense oceanic waters to the MS is controlled by
relatively shallow sills within the continental shelf (60–110 m, Palma
and Silva, 2004). In turn, the exchange of dense (
σ
θ
>24.7 kg m
3
),
relatively warm (~7.5 �C) and salty (>31) waters is blocked by the
shallow passages around Carlos III Island (Palma and Silva, 2004). East
of Carlos III Island the surface waters are relatively cold (6.5–7 �C) and
fresh (~30), while the deep waters present a weak thermal stratication
and a moderate salinity increase with depth, reaching 31.2 at 350 m
(Fig. 4). Consequently, the deep waters in the central basin are signi-
cantly less dense than in the western basin. Since the deep waters in the
western basin are about 1.5–2 �C colder than in the central basin the
density contrast below the sill depth is mostly due to the salinity
contrast.
The underwater glider observations provide much higher spatial
resolution than the rather sparse sampling (20–30 km) of the available
hydrographic sections. The glider surveys collected three high-
resolution cross-channel sections; the northernmost of these sections is
shown in Fig. 5. The glider sections present a nearly 100 m thick cold-
fresh (~6 �C, ~30.3) surface mixed layer overlying a warm-saline
lower layer (7.5 �C, 31). Thus, the vertical stability of the water col-
umn is sustained by the salinity stratication. The warm-saline deep
layer is conned to waters deeper than 200 m. Thus, as the bottom rises
northward (toward the Atlantic) the cold-fresh upper layer occupies the
entire water column in the northernmost portion of the section. It is this
less dense cold-fresh water that can ow northward towards the
Segunda Angostura narrow, which is only about 50 m deep. In the cross-
channel section the isopycnals slope downward towards the western side
of the strait, this is particularly evident in the western half of the section
(Fig. 5). The slope is consistent with a northward owing baroclinic
current in geostrophic balance (toward the Atlantic Ocean). The cross-
strait density contrast is caused by the low-salinity waters that occupy
the western portion of the section (Fig. 5). Because the glider section
Fig. 3. Temperature - Salinity diagrams of the Atlantic (orange) and Pacic
(blue) sectors of the Magellan Strait and the Le Maire Strait (green) (see Fig. 1a
for station locations). Open symbols represent summer (JFM) observations and
solid symbols winter observations (JASO). In the western MS only data less
dense than
σ
θ
~24.7 kg m
3
are included. (For interpretation of the references
to colour in this gure legend, the reader is referred to the Web version of
this article.)
Fig. 4. Vertical distribution of temperature (
�C, a), salinity (b) and density
anomaly (kg m
3
, c) along the upper 400 m of the MS as observed during the
CIMAR16 cruise (see Fig. 1 for station locations). The observations were
collected between October 20 (station 1) and November 1, 2010 (station 60)
and are indicated by the red dots in Fig. 1a. (For interpretation of the references
to colour in this gure legend, the reader is referred to the Web version of
this article.)
A.A. Brun et al.

Estuarine, Coastal and Shelf Science 237 (2020) 106661
6
does not reach the coast, the extent of this low-salinity waters is not well
resolved by the observations.
The Almirantazgo Fjord which is located at the northeastern corner
of Dawson Island and close to the southern end of Inútil Bay (IB in
Fig. 1a) is an additional source of low salinity waters to the central basin
via the Whiteside Channel (Aracena et al., 2015). Calving glaciers form
Cordillera Darwin iceeld (CDI in Fig. 1a, e. g. Silva and Prego, 2002;
Aracena et al., 2015) and runoff fresh waters (e.g., Antezana, 1999)
salinities ~ 28 are observed at the head of the fjord. Within the fjord the
upper 60 m of the water column salinity increases towards the MS to
about 30.5 (Fig. 6). Below this surface layer a relatively cold (~5.8 �C)
and saltier layer (<30.6; less than 24 kgm
-3
) is observed. The deeper
layers (down to ~ 200 m), in turn, is warmer (6.2 �C) and saltier
(>30.7), and is possibly originated in the mouth of Whiteside Channel
(st.55). The fjord inuence manifests as a local salinity minimum
(<30.5) around 350 km in the along passage section (Fig. 4). Along the
MS axis salinity increases eastward, reaching 30.7 between the SA and
PA, and then more rapidly towards the Atlantic mouth. However note
that stations 1 and 3 are nearly 90 km apart and do not capture the
along-strait thermohaline structure near the eastern end of the MS. In
contrast with the central and eastern basins the Atlantic mouth presents
vertically homogeneous temperature, salinity and density stratication,
particularly east of the Primera Angostura. Given that this region close
to the Atlantic mouth is mostly shallower than 50 m, the lack of vertical
stratication may be partly associated with tidal mixing promoted by
the interaction of the intense tidal currents and the bottom topography
(Medeiros and Kjerfve, 1988; Michelato et al., 1991; Sassi and Palma,
2006). The eastern basin also presents a different seasonal sea surface
temperature pattern from the central and western basins, with about 0.3
�C warmer waters than the central basin in austral summer and 2.2 �C
colder waters in winter (Fig. 7). Assuming that the net sea-air heat uxes
in both basins are similar, the distinct SST amplitudes are possibly a
response to different mixed layer depths. Interestingly, the CIMAR 16
data display a small temperature difference (<0.2 �C) across Segunda
Angostura, primarily because this region was sampled in mid-October,
when the temperature difference is relatively small (Fig. 7). Though
the transition between the relatively warm-fresh waters of the central
basin and the cold-salty eastern basin is difcult to identify based on the
relatively sparse in-situ data, the similarity between the annual cycles of
satellite derived sea surface temperature at either side of PA suggests
that the transition takes place at SA.-
The hydrographic characteristics of the Atlantic mouth of the MS are
portrayed by two high-resolution salinity cross-sections (3–5 km station
spacing) occupied in mid-September 2006 and late March 2009 (Fig. 8).
Since the temperature variations are minor even in summer (9.3–10 �C,
see Fig. 3) only the salinity sections are presented in Fig. 8. In late winter
2006 salinity between 31.14 in the upper layer on the northern margin
and ~32.5 throughout most of the mouth. In late March 2009 salinity
uctuates between 31.024 on the northern margin and 32.596 in the
deep central and the southern slope of the channel. Thus, both sections
present nearly homogeneous temperature distributions and a narrow
band (~5 km) of low salinity waters located against the northern shore.
These waters are the source of lowest salinity to the Atlantic shelf
excluding those regions under the direct inuence of continental
Fig. 5. Vertical distribution of temperature
(
�C, top), salinity (middle) and density
anomaly (kg m
3
, bottom) along (left) and
across (right) the central basin of the MS as
observed during the underwater glider sur-
vey (see Fig. 1a for location). The observa-
tions were collected between June 24 and
July 5, 2011 and are indicated in Fig. 1a by
purple and black dots for the along-strait and
cross-strait sections, respectively. (For
interpretation of the references to colour in
this gure legend, the reader is referred to
the Web version of this article.)
A.A. Brun et al.

Estuarine, Coastal and Shelf Science 237 (2020) 106661
7
discharges, such as the mouth of the Santa Cruz river (see Fig. 2).
Though the available hydrographic observations indicate waters are
vertically mixed in the eastern basin of the MS, the scarcity of these data
prevents determining the possible source of the low salinity waters
observed along the northern portion of the Atlantic mouth.
4.4. The low salinity outow
The low salinity plume emanating from the MS is clearly observed in
the regional salinity distribution (Fig. 2). The plume extends northward,
rst along the shore and displaces offshore north of Cape Blanco
(~47�S). Off Valdes Peninsula (~43�S), about 1200 km from the MS the
plume salinity minimum is located in the mid-shelf region and reaches
33.5 (Fig. 2).
To better understand the nature and quantify the transport of the
outow from the MS to the Atlantic shelf we estimated the geostrophic
ow based on the two high-resolution cross-channel sections. The
geostrophic velocities are referenced to the deepest common level of
each station pair. As described earlier, both sections present a well-
dened salinity minimum close to the northern coast. The location of
the buoyant plume along the northernmost portion of the strait is
consistent with a buoyant discharge under the effects of rotation,
owing with the coast at its left (e.g. against the northern shore of the
strait) in the southern hemisphere. The cross-strait salinity gradients
observed between the low salinity plume and the middle and southern
portions of the mouth are ~0.4 km
1
and ~0.2 km
1
during the
September 2006 and March 2009 occupations, respectively. The
reduced salinity gradient in late austral summer is partly due to the
wider and deeper low salinity plume at that time. As suggested by the T-
S diagrams presented in Fig. 3, the density eld at the eastern mouth of
the strait is mostly determined by salinity. Thus, the low salinity in the
north end of the Atlantic mouth creates a low density plume at that
location, with a cross-channel density gradient of ~0.15 kg m
3
km
1
at
the transition with the mid portion of the section. Despite the relatively
shallow depth of the channel, this large density gradient leads to strong
baroclinicity and an intense surface geostrophic jet of nearly 50 cm s
1
.
In the March 2009 occupation, the velocity away from the plume front is
substantially lower, though still signicant (~15 cm s
1
, lower panels in
Fig. 8). The geostrophic velocities within the jet in March 2009 are
slightly higher than in September 2006, and are also more detached
from the northern shore, in agreement with the more offshore extension
of the low salinity plume. The estimated volume transports (Q
v
) within
the low salinity jets in September 2006 and March 2009 are 38,500 and
74,500 m
3
s
1
, respectively. The nearly doubled transport estimate in
2009 is associated with a wider and deeper low-salinity plume. Though
in March 2009 the geostrophic velocities in the mid-channel region are
not negligible (15 cm s
1
), the volume transport within the low salinity
plume accounts for more than 80% of the total geostrophic transport
through the Atlantic mouth of the MS. We therefore argue that the low
salinity jet in the northern shore of the MS mouth is the main source
feeding the low salinity plume that extends northward along the Atlantic
shelf.
Fig. 6. Vertical distribution of temperature (
�C, left), salinity (middle) and density anomaly (kg m
3
, right) from the head of Almirantazgo Fjord (AF) to the Magellan
Strait (MS) as observed during the CIMAR16 cruise indicated by red dots in Fig. 1a. The observations were collected between October 2 (station 6) and October 22,
2010 (station 51). (For interpretation of the references to colour in this gure legend, the reader is referred to the Web version of this article.)
Fig. 7. Climatological (2003–2010) sea surface temperature variability based
on daily data from NOAA NCDC in the central (red) and eastern (blue) basins of
the Magellan Strait. These basins are separated by the Segunda Angostura. The
light shadings indicate daily standard deviations around the mean SST. (For
interpretation of the references to colour in this gure legend, the reader is
referred to the Web version of this article.)
A.A. Brun et al.

Estuarine, Coastal and Shelf Science 237 (2020) 106661
8
5. Discussion
The regional salinity distribution (Fig. 2) suggests a strong inter-
oceanic connectivity between the Pacic and Atlantic shelves. Despite
the complex conguration of the coast and the bottom topography the
Magellan Strait appears to be the most effective path of the resulting low
salinity waters to the Atlantic shelf (Figs. 2 and 3).
5.1. Pacic sector
At a regional scale the salinity minimum along the southern coast of
Chile (S <29; Fig. 2) forms a narrow band, as oceanic subantarctic
waters are only separated from the fjord system by the narrow (~10–20
km) continental shelf (Antezana, 1999). Recent analyses of historical
hydrographic and Argo proling oat data together with satellite ob-
servations show that the low salinity plume plays a more signicant role
in vertical stratication of the outer shelf and slope north of 47�S,
whereas temperature variations are more important further south (Sal-
días et al., 2019). The low salinity coastal waters create an offshore
density gradient, possibly reinforcing the regional southward ow in the
upper layer south of 45�S (Davila et al., 2002). This surface intensied
coastal current thus feeds the low salinity waters southward, where they
become available for export towards the Atlantic via the MS and around
the southern coast of Tierra del Fuego. Though the majority of the low
salinity waters from the southern portion of the southeast Pacic con-
tinental shelf are probably transferred to the Atlantic south of Tierra del
Fuego via the Cape Horn Current (Guihou et al., 2020) the surface
salinity distribution clearly shows that the MS is the source of the lowest
salinity waters to the Atlantic. This ow receives additional freshwater
runoff from the southern ank of the Cordillera Darwin iceeld into the
Beagle Channel. Near coastal surface observations collected close to the
north shore of the Beagle Channel display salinities as low as ~22 (e.g.
Aguirre et al., 2012). Despite these additional freshwater inows to the
waters owing south of Tierra del Fuego, the salinity observed in the Le
Maire Strait is higher than 32.5, suggesting relatively intense mixing
with saltier waters found in the continental shelf.
5.2. Magellan Strait
Though tidal mixing plays a major role in mixing the waters within
the MS (e.g. Medeiros and Kjerfve, 1988), idealized and realistic simu-
lations suggest that the sea-level difference between the Pacic and
Atlantic mouths together with regional wind forcing largely control the
ow through the MS (Sassi and Palma, 2006; Guihou et al., 2020).The
inter-ocean transition of water masses along the MS in spring 2010
(Fig. 4) presents distinct oceanographic characteristics in the different
basins. The along-channel salinity distribution (Fig. 4b) is qualitatively
similar to the salinity distribution derived from a realistic simulation
forced by tides, winds and large scale sea level differences (see Sassi and
Palma, 2006, their Fig. 8). The western basin presents relatively cold
(~7 �C), low salinity (<29.0) waters in the upper 50 m of the water
column and a sharp halocline at around that depth. Subsurface and
deeper layers (down to ~ 800 m), in turn, are warmer (>8 �C) and saltier
(>32), and relatively homogeneous (Fig. 4). East of the CI sill, the
central basin presents cold (<7 �C) and fresh (<30.7) waters down to
about 100 m depth. The low surface salinity in the central basin is due to
the local input of continental runoff and discharge of cold meltwater
from Almirantazgo Fjord through the Whiteside Channel (Figs. 4 and 5)
and Inútil Bay (e.g., Panella et al., 1991; Mazzocchi et al., 1995; Val-
denegro and Silva, 2003; Aracena et al., 2015). Estimates of river runoff
within the central basin are low, 118 m
3
s
1
(Ministerio de Obras Púb-
licas, Chile, http://snia.dga.cl/BNAConsultas/), but given the intense
Fig. 8. High resolution salinity (top) and geostrophic velocity (cm s
1
, bottom) sections occupied across the Atlantic mouth of the Magellan Strait on 17 September
2006 (left) and 22 March 2009 (right). The stations are indicated by green dots in Fig. 1a. (For interpretation of the references to colour in this gure legend, the
reader is referred to the Web version of this article.)
A.A. Brun et al.

Estuarine, Coastal and Shelf Science 237 (2020) 106661
9
precipitation, averaging about 2000 mm yr
1
(Aracena et al., 2015),
other potential sources of freshwater (e.g. meltwater from the Darwin
iceeld) need to be considered. Integration of this precipitation over the
drainage area of central basin leads to a total runoff of about 10
3
m
3
s
1
.
Though the vertical temperature stratication in the central basin
observed during the underwater glider survey is different from that
observed during the CIMAR 16 survey, suggesting substantial time
variability, the upper layer properties in all available observations
suggest the above mentioned inow of low salinity water from the
Almirantazgo Fjord. Salinity, temperature and density show a fairly
homogeneous mixed layer (~6 �C, ~30.3, ~23.7 kg m
3
; Fig. 4). More
homogeneous thermohaline characteristics are observed in the eastern
basin (<6.8 �C and <32). These properties are similar to the observa-
tions reported by Panella et al. (1991). The vertically homogeneous
thermohaline characteristics are the result of the intense mixing by the
strong tidal currents present at Primera Angostura (4.5 m s
1
) and
Segunda Angostura (3 m s
1
) narrows in response to the propagation of
the large semidiurnal tidal waves entering from the Atlantic (Michelato
et al., 1991). In agreement with the ndings of Panella et al. (1991)
these characteristics are comparable to those found in the coastal water
of the Atlantic Patagonia continental shelf (Fig. 2); this water is thought
to originate from mixing of subantarctic waters found north of the
Subantarctic Front, with more diluted waters derived from the MS and
spreading over a large area of the shelf (Fig. 2).
Valdenegro and Silva (2003) proposed a vertical circulation scheme
in the MS consisting of surface ows of low-salinity waters from the
central basin towards both the Pacic and the Atlantic oceans, which are
partially compensated by inows at depth. Silva and Vargas (2014)
proposed a similar circulation pattern in the western and central basins
of the MS and also illustrate the low-salinity inow from Almirantazgo
fjord. The along-strait salinity distribution reveals a substantial input of
low-salinity waters in the western basin, which does not seem to pene-
trate eastward beyond the Carlos III sill, nor westward towards the Pa-
cic as suggested by these studies (Fig. 4). Note also that salinities lower
than 30 at 20 m depth are conned to the MS near the eastern mouth of
the strait (Fig. 2). However, the waters from 75 to 100 m depth in the
western basin appear to ow over the Carlos III sill and occupy the
subsurface layer of the central basin, in agreement with the deep part of
the circulation schemes proposed by Valdenegro and Silva (2003) and
Silva and Vargas (2014). These waters may mix with low-salinity waters
derived from Almirantazgo fjord and feed the ow to the eastern basin
which eventually enters the Atlantic shelf. Valdenegro and Silva (2003)
suggest that despite the intense tidal circulation in the eastern basin the
excess precipitation and runoff lead to a net ow towards the Atlantic. A
more detailed description of the water mass distribution and the ex-
changes between the MS basins would require sampling at signicantly
higher resolution than presently available.
5.3. Atlantic sector
The regional salinity distribution (Fig. 2) suggests that the sources of
low salinity waters to the Atlantic shelf are derived mainly from the
eastern basin of the MS (orange dots, Fig. 3) and from the Le Maire Strait
(LM; green dots Fig. 3). Numerical simulations indicate that the majority
of the inow to the Atlantic shelf is through the Le Maire Strait and also
east of Estados Island (e.g., Combes and Matano, 2014; Guihou et al.,
2020). Our surface salinity distribution (Fig. 2) and water mass analysis
(Fig. 3) shows that the latter inows carry waters saltier than 32.5, while
the waters associated with the weaker inow through the MS present the
lowest salinities, within the 31.1 and 32.5 range, and are the source of
lowest salinity waters observed over the Atlantic shelf. The core of low
salinity waters (S <31.5) observed on the northern portion of the
Atlantic mouth of the MS eventually mixes with the saltier waters
entering through the LM Strait to create the core of the low salinity
plume that extends ~1800 km northeastward (Fig. 2). This low salinity
waters are a key element of the so-called Subantarctic Shelf Waters that
ll most of the Atlantic Patagonia continental shelf (Piola et al., 2000).
Numerical models indicate that the northernmost extension of the
low salinity plume is set by the discharge of saltier waters from the San
Matias Gulf (41.5�S, see Fig. 2) and that without this outow the core of
low salinity would extend up to the Río de la Plata River near 35�S
(Palma and Matano, 2012). Moreover, the models indicate that changes
in the magnitude of the low salinity outow from the MS substantially
impact the downstream penetration of the plume and its buoyancy
transport, which is modulated by the net ow through the MS mouth
and its salinity difference with the ambient Atlantic waters. It is there-
fore important to determine the properties and quantify the inow of MS
waters to the Atlantic shelf.
Our geostrophic velocity estimates have several shortcomings, which
are discussed below. First, the instantaneous circulation is most likely
inuenced by intense tidal currents. Direct current observations at 17 m
depth from a current meter moored about 16 km southeast of the hy-
drographic section on the southern ank of the channel that extends
southeastward from the MS mouth into the Atlantic continental shelf
(bottom depth 55 m, see Fig. 1) present semi-diurnal reversing along-
isobath velocities of ~35 cm s
1
, which are of the same order of
magnitude of the estimated geostrophic velocities along the edge of the
low-salinity plume. This current meter record is used to estimate the
tidal currents at the time of the hydrographic occupations based on 7
tidal harmonics. During the September 2006 occupation intense (40 cm
s
1
) southeastward (toward the Atlantic) tidal currents are estimated
during the sampling of the low-salinity plume (stations 24 and 25). The
currents reversed by the time of the occupation of the southward end of
the section (station 31) and presented similarly intense (40 cm s
1
)
northwestward ow. In contrast, during the March 2009 survey of the
low-salinity plume (stations 8 and 9) the estimated tidal ow was
relatively weak (6–18 cm s
1
) and northwestward (towards the Pacic).
During the occupation of southern portion of the section the tidal cur-
rents were southeastward (25 cm s
1
). It is unclear whether or not the
tidal currents may impact on the salinity distribution across the Atlantic
mouth, and thus on the geostrophic ow estimates. Interestingly, the
estimated eastward geostrophic transport is larger when the stations
were occupied during a westward owing tidal current. Despite the
nearly 180�out of phase tidal conditions during the occupation of the
northern portion of both cross-channel sections, the observations sug-
gest that the presence of the low salinity plume against the northern
coast of the mouth is a permanent feature. This is further supported by
yet another observation of low salinity waters against the northern shore
of the Atlantic mouth, and similarly intense cross-channel salinity gra-
dients (~0.26 km
1
) revealed by a set of lower-resolution hydrographic
stations occupied across the MS mouth in December 1974 (not shown).
Since the salinity gradient is the main cause of the intense cross-channel
pressure gradient that leads to the geostrophic ow towards the Atlantic,
it seems reasonable to speculate that the inferred geostrophic ow may
be a real and quasi-permanent feature.
In addition to the above discussed tidal effects, the intense (0.5 m
s
1
) and narrow (5 km) currents estimated based on the geostrophic
balance imply a relatively high Rossby number (~0.9), suggesting that
inertial terms may play a signicant role in the cross-channel mo-
mentum balance, and introducing additional uncertainties in the
derived geostrophic velocities. However, the geostrophic shear and the
magnitude of the bottom-referenced geostrophic velocity estimated at
the northernmost station pairs (stations 7–9) during the March 2009
occupation is in surprisingly good agreement with concomitant direct
velocity observations from a vessel mounted acoustic Doppler current
proler that operated during the occupation of stations 8 and 9 (i.e. over
the low salinity plume, Fig. 9). For this comparison the predicted tidal
current was used to remove the tidal signal from the observed ADCP
velocity prole. Thus, despite the aforementioned shortcomings, and
given the lack of other observationally based estimates of the volume
transport through the MS we argue that it is useful to carry out these
transport calculations as an order of magnitude estimate.
A.A. Brun et al.

Estuarine, Coastal and Shelf Science 237 (2020) 106661
10
Finally, relying in relative geostrophic velocities is questionable as it
does not account for a barotropic contribution, precisely the main
forcing used in numerical models (Sassi and Palma, 2006; Palma and
Matano, 2012). Given the lack of observations required to estimate the
barotropic ow we had no choice but to ignore its possible contribution
to the Magellan throughow.
Numerical simulations of a realistic channel geometry forced by
Pacic-Atlantic sea level differences, regional wind forcing and tides
suggest that the net mean MS outow to the Atlantic is about 85,000 m
3
s
1
(Sassi and Palma, 2006). This estimate is in relatively good agree-
ment with the 74,500 m
3
s
1
geostrophic transport of low salinity waters
through the northern portion of the Atlantic mouth estimated in March
2009. That study also predicted a relatively strong sensitivity of the
Pacic to Atlantic transport through the MS to the intensity of the
westerly wind stress, with a transport reduction of 20% when the wind
stress is decreased 50% and a 40% transport increase when the wind
stress is doubled. We estimated the surface wind stress around the
Atlantic mouth of the MS derived from CCMP winds during both occu-
pations. During the September 2006 survey the daily averaged westerly
wind stress was 0.06 Pa, while in March 2009 it increased to 0.13 Pa (not
shown). Thus, a large fraction of the nearly doubling of volume transport
(from 38,500 to 74,500 m
3
s
1
) may be associated with the increased
westerly wind stress as predicted by the model sensitivity experiments of
Sassi and Palma (2006).
Numerical models further suggest that the northward penetration of
the low salinity plume on the Atlantic shelf is modulated by the buoy-
ancy ux anomaly introduced by the low salinity discharge at the strait’s
mouth. For instance, the 85,000 m
3
s
1
model discharge combined with
a salinity difference (ΔS) of 1.4 relative to an ambient salinity (S
0
) of
33.7, leads to a buoyancy ux anomaly (Q
b
¼Q
v
ΔS/S
0
) of ~3500 m
3
s
1
(Palma and Matano, 2012). Because the observed salinity difference on
both occupations is larger than that imposed in the models: ~ 2.2
(31.5–33.7), the model-derived buoyancy ux can be attained with a
smaller volume transport (53,000 m
3
s
1
), which falls between the
geostrophic transports estimated on the hydrographic occupations re-
ported here. Conversely, the observationally-derived buoyancy ux
anomalies are 2300 and 4400 m
3
s
1
, in September 2006 and March
2009, respectively. Based on the model results such changes in buoyancy
ux anomaly may imply signicant changes in the northward extension
of the Magellan low salinity plume. However, there are no observations
to evaluate these changes at the time of the MS sections occupations.
6. Conclusions
The combined analysis of historical hydrographic data collected
around southern South America together with high-resolution sections
reveal that the Magellan Strait is the source of lowest salinity waters to
the Atlantic Patagonian continental shelf. The highly complex geo-
morphology of the Magellan Strait and the strong seasonal and lat-
itudinal patterns in precipitation and continental freshwater discharge
are likely to modulate the nature of the low salinity anomaly into the
Atlantic. The western basin of the strait, bordering the Pacic, is char-
acterized by low salinities (<28) in the upper 50 m associated with the
inow of cold, low salinity runoff waters capping a warm-salty subsur-
face layer of oceanic Subantarctic waters. In the much shallower central
basin (east of CI) the temperature and density show a fairly homoge-
neous surface mixed layer compared to the western basin and suggest
the inow of lower salinity waters from Almirantazgo Fjord. Intense
tidal currents in the eastern basin of the MS effectively mix the waters in
the vertical. Though it is difcult to infer the circulation patterns from
the property distributions within the MS, the across-strait density
structure revealed by sections occupied in the central basin and the
Atlantic mouth suggest a net ow of low-salinity waters towards the
Atlantic. In these sections the core of low-salinity waters is found against
the western and northern shores, respectively. These waters create a
sharp salinity and density front and a strong baroclinic signal that leads
to eastward geostrophic velocities and a net Pacic to Atlantic transport.
Though the ow of low salinity waters may be a minor contributor to the
total northeastward mass transport over the Atlantic shelf inferred from
numerical models, the associated buoyancy ux may play a signicant
role on its thermohaline characteristics and on the extent of the sub-
antarctic shelf waters.
Declaration of competing interest
The authors declare that they have no known competing nancial
interests or personal relationships that could have appeared to inuence
the work reported in this paper.
CRediT authorship contribution statement
Anahí A. Brun: Conceptualization, Writing - original draft, Visual-
ization, Writing - review & editing, Formal analysis. Nadin Ramirez:
Writing - review & editing, Formal analysis, Resources. Oscar Pizarro:
Supervision, Writing - review & editing, Formal analysis, Resources.
Alberto R. Piola: Supervision, Funding acquisition, Investigation,
Conceptualization, Methodology, Writing - original draft.
Acknowledgements
This research was possible thanks to grant CRN3070 from the Inter-
American Institute for Global Change Research, which is funded by the
US National Science Foundation grant GEO-1128040. Funding for the
GEF Patagonia observations was provided by the Global Environmental
Facility grant UNDP ARG/02/018-GEF IBRD 28385-AR, while the
Polarstern observations were funded by the French Centre National
d’Etudes Spatiales through the OST/ST altimetric program and by CNRS
Institut des Sciences de l’Univers. The glider survey was partially funded
by COPAS Sur-Austral (CONICYT PIA BASAL PFB31 and CCTE
AFB170006). We thank Guadalupe Alonso (SHN, Argentina) for
providing the tidal current predictions near the Atlantic mouth of the
MS. Comments from Paola D�
avila, Elbio Palma and two anonymous
Fig. 9. Vertical proles of along-channel velocity (cm s
1
) from a vessel
mounted ADCP (station 8, thin black line) after removing the predicted tidal
currents and geostrophic velocity referenced to the bottom between stations 7
and 9 (heavy grey line).
A.A. Brun et al.

Estuarine, Coastal and Shelf Science 237 (2020) 106661
11
reviewers on a previous version of this manuscript are greatly
appreciated.
References
Aguirre, G.E., et al., 2012. Seasonal variability of metazooplankton in coastal sub-
Antarctic waters (Beagle Channel). Mar. Biol. Res. 8 (4), 341–353, 5.
Antezana, T., 1999. Hydrographic features of Magellan and Fuegian inland passages and
adjacent Subantarctic waters. Sci. Mar. 63 (Suppl. 1), 23–34.
Aracena, C., et al., 2015. Holocene variations in productivity associated with changes in
glacier activity and freshwater ux in the central basin of the Strait of Magellan.
Palaeogeogr. Palaeoclimatol. Palaeoecol. 436, 112–122.
Bianchi, A.A., et al., 1982. Analisis estadistico-volumetrico de las caracteristicas T-S del
sector austral de la plataforma continental Argentina. Acta Ocean Arg. 3 (1), 93–117.
Boyer, T.P., et al., 2013. World Ocean Database 2013. In: Levitus, S., Mishonov, A.,
Technical (Eds.), NOAA Atlas NESDIS, vol. 72, p. 209.
Brandhorst, W., Castello, J.P., 1971. Evaluaci�
on de los recursos de anchoíta
(Engraulisanchoita) frente a la Argentina y Uruguay, vol. 29. I Proy Des Pesq FAO,
Publication, p. 63.
Camus, P., 2001. Biogeografía marina de Chile continental. Rev. Chil. Hist. Nat. 74,
587–617.
Charo, M., Piola, A., 2014. Hydrographic data from the GEF Patagonia cruises. Earth
Syst. Sci. Data 6, 265–271.
Combes, V., Matano, R.P., 2014. A two-way nested simulation of the oceanic circulation
in the Southwestern Atlantic. J. Geophys. Res. Oceans 119. https://doi:10.1002/201
3JC009498.
CONA, 2011. Crucero CIMAR 16 Fiordos: 11 de octubre al 19 de noviembre 2010.
Informes Preliminares, Valparaíso, Chile. http://www.cona.cl/pub/pub_cimar.ph
p#.
Davila, P., Figueroa, D., Müller, E., 2002. Freshwater input into the coastal ocean and its
relation with the salinity distribution off austral Chile (35-55�s). Continent. Shelf
Res. 22 (3), 521–534. https://doi.org/10.1016/S0278-4343(01)00072-3.
Deacon, G.E.R., 1937. The hydrology of the southern ocean. Discov. Rep. 7, 171–238,
1933.
Fierro, J., 2008. Tides in the austral Chilean channels and fjords. Avances en el
conocimiento oceanogr�
aco de las aguas interiores chilenas, Puerto Montt a cabo de
Hornos. In: Silva, N., Palma, S. (Eds.), Comit�
e Oceanogr�
aco Nacional. Ponticia
Universidad Cat�
olica de Valparaíso, Valparaíso, pp. 63–66.
Global Runoff Data Centre, 2009. Surface Freshwater Fluxes into the World Oceans/
GRDC. Federal Institute of Hydrology (BfG), Koblenz, Germany.
Guerrero, R., Piola, A., 1997. Masas de agua. In: Boschi, E. (Ed.), El Mar Argentino Y sus
Recursos Pesqueros, Tomo 1. Instituto Nacional de Investigaci�
on y Desarrollo
Pesquero, Mar del Plata, pp. 107–118.
Guihou, K., Piola, A.R., Palma, E.D., Chidichimo, M.P., 2020. Dynamical connections
between large marine ecosystems of austral south America: a numerical analysis.
Ocean Sci. Discuss. https://doi.org/10.5194/os-2019-99. Ocean Science, (in press).
Lucas, A.J., et al., 2005. Coastal oceanographic regimes of the northern argentine
continental shelf (34–43�S). Estuar. Coast Shelf Sci. 5, 405–420.
Lumpkin, R., Johnson, G.C., 2013. Global ocean surface velocities from drifters: mean,
variance, El Ni~
no–Southern Oscillation response, and seasonal cycle. J. Geophys.
Res. Oceans 118 (6), 2992–3006.
Lutz, V., et al., 2016. Bio-optical characteristics along the straits of magallanes.
Continent. Shelf Res. 119, 56–67.
Maximenko, N., et al., 2009. Mean dynamic topography of the ocean derived from
satellite and drifting buoy data using three different techniques. J. Atmos. Ocean.
Technol. 26 (9), 1910–1919.
Mazzocchi, M.G., et al., 1995. Copepods. In: Guglielmo, L., Ianora, A. (Eds.), Atlas of
Marine Zooplankton. Strait of Magellan. Springer, Berlin, pp. 1–279.
Medeiros, C., Kjerfve, B., 1988. Tidal characteristics of the strait of magellan. Continent.
Shelf Res. 8, 947–960.
Michelato, A., Accerboni, E., Berger, P., 1991. Current meter observations in the eastern
and central sectors of the Strait of Magellan. Boll. Geos. Teor. Appl. 9, 261–271.
Palma, E.D., Matano, R.P., 2012. A numerical study of the Magellan Plume. J. Geophys.
Res. 117, C05041.
Palma, S., Silva, N., 2004. Distribution of siphonophores, chaetognaths, euphausiids and
oceanographic conditions in the fjords and channels of southern Chile. Deep Sea Res.
Part 2 Top. Stud. Oceanogr. 51, 513–535.
Panella, S., et al., 1991. A preliminary contribution to understanding the hydrological
characteristics of the Strait of Magellan: austral spring 1989. Boll. Geos. Teor. Appl.
9, 107–126.
Parker, G., Paterlini, C.M., Violante, R.A., 1997. El fondo marino. In: Boschi, E. (Ed.), El
Mar Argentino y sus Recursos Marinos, INIDEP, Mar del Plata, pp. 65–87. Argentina.
Peterson, R.G., Whitworth III, T., 1989. The Subantarctic and Polar Fronts in relation to
the deep water masses through the South- western Atlantic. J. Geophys. Res. 94 (C8),
10817–10838.
Piola, A., et al., 2000. Subtropical shelf front off eastern South America. J. Geophys. Res.
105, 6566–6578.
Piola, A.R., Gordon, A.L., 1989. Intermediate waters in the southwest South atlantic.
Deep-Sea Res. 36 (1), 1–16.
Provost, C., 2010. The expedition of the Research Vessel "Polarstern" to the antarctic in
2009 (ANT-XXV/4), reports on polar and Marine Research, Bremerhaven, 616.
Alfred Wegener Institute for Polar and Marine Research, p. 81. https://doi.org/
10.2312/BzPM_0616_2010. https://epic.awi.de/id/eprint/29628/1/Pro2010a.pdf.
Rabassa, J., Clapperton, C.M., 1990. Quaternary glaciations of the southern Andes. Quat.
Sci. Rev. 9, 153–174. https://doi.org/10.1016/0277-3791(90)90016-4.
Reynolds, R.W., et al., 2007. Daily high-resolution blended analyses for sea surface
temperature. J. Clim. 20, 5473–5496.
Saldías, G.S., Sobarzo, M., Qui~
nones, R., 2019. Freshwater structure and its seasonal
variability off western Patagonia. Prog. Oceanogr. 174, 143–153.
Sassi, M.G., Palma, E.D., 2006. Modelo Hidrodin�
amico del Estrecho de Magallanes;
Asociaci�
on Argentina de Mec�
anica Computacional. Mec. Comput. XXV (16),
1461–1477, 11. https://cimec.org.ar/ojs/index.php/mc/article/view/577/550.
Scasso, L.M., Piola, A.R., 1988. Intercambio neto de agua entre el mar y la atm�
osfera en
el Golfo San 684 Matías. Geoacta 15 (1), 13–31.
Schneider, W., et al., 2003. Characteristics and formation of eastern South Pacic
intermediate water. J. Geophys. Res. 30 (11), 1581.
Sievers, H.A., Silva, N., 2008. Water masses and circulation in austral Chilean channels
and fjords. In: Silva, N., Palma, S. (Eds.), Progress in the Oceanographic. Knowledge
of Chilean Interior Water, from Puerto Montt to Cape Horn. Comit�
e Oceanogr�
aco
Nacional-Ponticia Universidad Cat�
olica de Valparaíso, pp. 53–58.
Silva, N., Neshyba, S., 1977. Corrientes superciales frente a la costa austral de Chile.
Cienc. Tecnol. del Mar 3, 37–42.
Silva, N., Prego, R., 2002. Carbon and nitrogen spatial segregation and stoichiometry in
the surface sediments of Southern Chilean inlets (41–56oS). Estuar. Coast Shelf 55,
763–775.
Silva, N., Rojas, N., Fedele, A., 2009. Water masses in the Humboldt Current System:
properties, distribution, and the nitrate decit as a chemical water mass tracer for
Equatorial Subsurface Water off Chile. Deep-Sea Res. II 56, 1004–1020.
Silva, N., Vargas, C.A., 2014. Hypoxia in Chilean patagonian fjords. Prog. Oceanogr. 129,
62–74.
Uppala, S.M., et al., 2005. The ERA-40 re-analysis. Q. J. Roy. Meteorol. Soc. 131,
2961–3012.
Valdenegro, A., Silva, N., 2003. Caracterizaci�
on oceanogr�
aca física y química dela zona
de canales y ordos australes de Chile entre el Estrecho de Magallanes y Cabo de
Hornos (CIMAR 3 Fiordos). Cienc. Tecnol. Mar (CONA) 26 (2), 19–60. http://www.
cona.cl/ctmol2/vol26-2/Silva/Silva.pdf.
Vargas, C.A., Cuevas, L.A., Silva, N., Gonz�
alez, H.E., De Pol-Holz, R., Narv�
aez, D.A.,
2018. Inuence of gla-cier melting and river discharges on the nutrient distribution
and DIC recycling in the Southern Chilean Patagonia. J. Geophys. Res.:
Biogeosciences 123, 256–270. https://doi.org/10.1002/2017JG003907.
Weatherall, P., et al., 2015. A new digital bathymetric model of the world’s oceans. Earth
Space Sci. 2, 331–345. https://doi:10.1002/2015EA000107.
A.A. Brun et al.