Content uploaded by G. Caniaux
Author content
All content in this area was uploaded by G. Caniaux on Oct 08, 2018
Content may be subject to copyright.
Gyre-scale deep convection in the subpolar North
Atlantic Ocean during winter 2014–2015
A. Piron
1
, V. Thierry
1
, H. Mercier
2
, and G. Caniaux
3
1
Ifremer, Laboratoire d’Océanographie Physique et Spatiale, UMR 6523 CNRS-IFREMER-IRD-UBO, Plouzané, France,
2
CNRS,
Laboratoire d’Océanographie Physique et Spatiale, UMR 6523 CNRS-IFREMER-IRD-UBO, Plouzané, France,
3
Centre National
de Recherches Météorologiques, UMR 3589 Météo-France-CNRS, Toulouse, France
Abstract Using Argo floats, we show that a major deep convective activity occurred simultaneously in the
Labrador Sea (LAB), south of Cape Farewell (SCF), and the Irminger Sea (IRM) during winter 2014–2015.
Convection was driven by exceptional heat loss to the atmosphere (up to 50% higher than the climatological
mean). This is the first observation of deep convection over such a widespread area. Mixed layer depths
exceptionally reached 1700 m in SCF and 1400 m in IRM. The deep thermocline density gradient limited the
mixed layer deepening in the Labrador Sea to 1800 m. Potential densities of deep waters were similar in the
three basins (27.73–27.74 kg m
3
) but warmer by 0.3°C and saltier by 0.04 in IRM than in LAB and SCF,
meaning that each basin formed locally its own deep water. The cold anomaly that developed recently in the
North Atlantic Ocean favored and was enhanced by this exceptional convection.
1. Introduction
Deep convection is a key oceanic process that ventilates the lower limb of the Meridional Overturning
Circulation (MOC), thus contributing to the storage of heat, anthropogenic carbon and oxygen in the deep
ocean [Pérez et al., 2013]. Monitoring and understanding variability in volumes and properties of the water
masses produced by deep convection is required to better model and predict future changes in deep con-
vection in a warming world. It would also allow a better assessment of the link between deep convection
and the MOC intensity measured at mid-latitudes [Cunningham et al., 2007] and subpolar latitudes [Mercier
et al., 2015].
The Labrador Sea Water (LSW) is formed by deep convection in the western North Atlantic Ocean subpolar
gyre. In the last decades several sites of LSW production were identified, each site showing large interannual
to decadal variability in convection strength. In the Labrador Sea, the deepest winter mixed layers were
observed between the late 1980s and the mid-1990s [Yashayaev, 2007] due to the recurrence of severe win-
ters during this high North Atlantic Oscillation (NAO) index period [Hurrell, 1995]. They reached 2300–2400 m
in 1993–1995 [Kieke and Yashayaev, 2015]. Since the mid-1990s, the convection was shallower due to milder
winters and the maximum mixed layer depths (MLDs) ranged from 500 to 1500 m, except in 2008 and 2014
when they reached 1850 and 1700 m, respectively [Våge et al., 2009; Yashayaev and Loder, 2009; Kieke and
Yashayaev, 2015]. Based on hydrographic data collected over the period 1900–2000, Pickart et al. [2003a]
showed that the core of the deep convection area is localized west of 52°W and south of 59°N. The deep con-
vection area extended to 48°W and 60°N during the deep convective years of the early 1990s [Pickart et al.,
2003a] and retreated westward in the early 2000s [Våge et al., 2009].
In the last decade, the Irminger Sea was also recognized as a deep convection site forming LSW under favor-
able conditions such as positive NAO index, preconditioning of the water column, and strong heat loss to the
atmosphere due to high wind speed events occurring east of Cape Farewell (the so-called Greenland Tip Jets)
[Bacon et al., 2003; Pickart et al., 2003b; Våge et al., 2009; de Jong et al., 2012; Piron et al., 2016]. The end of
winter MLDs in the Irminger Sea reached 1000 m in 2008 and 2012 [de Jong et al., 2012; Piron et al., 2016]
and 1400 m in 2015 [de Jong and de Steur, 2016; Fröb et al., 2016], which are the deepest mixed layers
observed in that region. Based on a one-dimensional (1-D) mixed layer model, Våge et al. [2008] and
Pickart et al. [2003a] suggested that convection might have reached 1500–2000 m in 1994 and 1995.
South of Cape Farewell, Bacon et al. [2003] and Piron et al. [2016] observed MLDs of 900–1000 m in 1997 and
2012, respectively. While this region was considered in some studies as part of the Irminger Sea [Piron et al.,
2016; Bacon et al., 2003], it could be a third distinct deep convection site. Indeed, using the minimum
PIRON ET AL. NORTH ATLANTIC 2014–2015 DEEP CONVECTION 1439
PUBLICATION
S
Geophysical Research Letters
RESEARCH LETTER
10.1002/2016GL071895
Key Points:
•Exceptional heat loss caused
exceptional convection at subpolar
gyre scale during winter 2014–2015
•Exceptionally deep mixed layers
directly observed south of Cape
Farewell (1700 m) and in the Irminger
Sea (1400 m)
•This exceptional convection was
favored by and enhanced the cold
anomaly that developed recently in
the North Atlantic Ocean
Supporting Information:
•Supporting Information S1
Correspondence to:
V. Thierry,
vthierry@ifremer.fr
Citation:
Piron, A., V. Thierry, H. Mercier, and
G. Caniaux (2017), Gyre-scale deep con-
vection in the subpolar North Atlantic
Ocean during winter 2014–2015,
Geophys. Res. Lett.,44, 1439–1447,
doi:10.1002/2016GL071895.
Received 10 NOV 2016
Accepted 27 JAN 2017
Accepted article online 29 JAN 2017
Published online 13 FEB 2017
©2017. American Geophysical Union.
All Rights Reserved.
potential vorticity at 1000 m as a footprint of the deep convective activity, Pickart et al. [2003a] showed that
during the 1989–1997 period there was a separate area of minimum potential vorticity south of Cape
Farewell. In this latter area, a near-surface stratified layer capping a homogenous layer down to 1800 m in
1991 and 1600 m in 2008 was observed by Pickart et al. [2003a] and Våge et al. [2009], respectively. Those
results suggest that MLDs could extend deeper than 1000 m south of Cape Farewell.
Våge et al. [2009] described the deep convection at the subpolar gyre scale for the winter 2007–2008 (winter
(N1)–Nwill be referred to as WinterN) owing to Argo data. The sampling in the Irminger Sea and south of
Cape Farewell was, however, insufficient to allow comparison of the deep convection properties (depth,
spatial extent, and thermohaline properties) in the different convection sites and to relate them to the
Figure 1. (a) Positions of Argo profiles available over January–April 2015. Black dots: MLD <700 m; colored dots:
MLD >700 m. Superimposed are the 1000 m isobath (thin black line) and the Absolute Dynamic Topography (cm) aver-
aged over January–April 2015 (gray contours). (b) Positions of the late winter (March–April) MLD >1000 m in the LAB (blue
dots), SCF (green dots), and IRM (red dots) boxes. Black dots: the deepest mixed layer in each box (1790, 1700, and 1400 m,
respectively). Yellow dots: profiles with MLD >1000 m outside the boxes. Black circles: profiles with oxygen data shown in
Figure 3. Black contours: winter (“”DJFM) anomalies (in %) of air-sea heat fluxes relative to the 1979–2015 climatological
value. Positive numbers indicate stronger than average air-sea heat loss to the atmosphere.
Geophysical Research Letters 10.1002/2016GL071895
PIRON ET AL. NORTH ATLANTIC 2014–2015 DEEP CONVECTION 1440
atmospheric forcing and oceanic preconditioning. This is now possible for Winter15 because of an adequate
Argo sampling of the subpolar gyre. The number of profiles available in the three convection sites during
winter (DJFM) was 2 to 3 times larger in Winter15 than in Winter08 (183 profiles against 80). Winter15 was
characterized by deep mixed layers [Fröb et al., 2016] and a NAO index close to 2 corresponding to the
highest value observed since the early 1990s suggesting a strong atmospheric forcing. This motivated the
investigation of deep convection in the subpolar gyre for Winter15.
2. Data and Methods
Argo data were used to describe the mixed layers for Winter15. We selected Argo temperature, salinity, and
oxygen profiles available in the subpolar gyre of the North Atlantic Ocean (52–66°N; 30°W–66°W) between 1
September 2014 and 30 April 2015 [Argo, 2015]. Oxygen data were corrected as suggested by Takeshita et al.
[2013] by using as a reference either the oxygen profile collected at float deployment if available or the World
Ocean Atlas 2009 [Garcia et al., 2010]. As in Piron et al. [2016], MLDs were calculated for each profile by com-
paring the split-and-merge [Thomson and Fine, 2003] and threshold [de Boyer Montégut et al., 2004] methods
and a visual control. Mixed layers isolated from the surface [Pickart et al., 2002] are not considered here to
guarantee that the estimated MLDs correspond to mixed layers formed locally.
We used the surface wind stress, air-sea heat flux, and sea surface temperature (SST) provided by the ERA-
Interim reanalysis [Dee et al., 2011]. We used the Absolute Dynamic Topography (ADT) data provided by
AVISO, and Sea Ice data from the National Snow and Ice Data Center (NSIC) [Fetterer et al., 2016].
3. The 2014–2015 Deep Convection
To characterize the deep convection of Winter15, we considered Argo profiles with MLD exceeding 700 m
(MLD >700 m) because it is the minimum depth to be reached for LSW renewal [Yashayaev, 2007;
Figure 2. (a–c) Late winter (March–April) profiles with MLD >1000 m in the LAB (Figure 2a), SCF (Figure 2b), and IRM
(Figure 2c) boxes. Thick lines: potential density (kg m
3
). Thin dashed lines: oxygen saturation (%). Black line: profiles
with the deepest mixed layers. Yellow line: profile located outside the boxes. MLD is indicated with dots on the density
profiles. (d) Properties of the late winter profiles with MLD >1000 m in LAB (blue dots), SCF (green dots), and IRM (red dots).
The black circles identify properties of the deepest mixed layers. The yellow dots identify profiles located outside the boxes.
Geophysical Research Letters 10.1002/2016GL071895
PIRON ET AL. NORTH ATLANTIC 2014–2015 DEEP CONVECTION 1441
Piron et al., 2016]. Those profiles were obtained between 24 January and 22 April 2015. They cover a
continuous wide area extending from the interior of the Labrador Sea (east of 55.4°W) to the Irminger Sea
(west of 32.2°W) (Figure 1a) and are all located within the cyclonic circulation surrounding the Labrador
Figure 3. (a–c) Time series of the mixed layer heat content (HC) and associated error (gray shading) and of the cumulative
sum from 1 September 2015 of the heat fluxes in LAB (Figure 3a), SCF (Figure 3b), and IRM (Figure 3c). Q
tot
is the sum of the
radiative fluxes (Q
rad
), latent (Q
lat
), sensible (Q
sen
), and Ekman transport heat fluxes (Q
Ek
). (d) Positions of the late summer
profiles used to calculate a mean late summer potential temperature profile in each box. (e) Quantity of heat to extract from
each box (HC
ext
) to homogenize this late summer profile down to depth Z.
Geophysical Research Letters 10.1002/2016GL071895
PIRON ET AL. NORTH ATLANTIC 2014–2015 DEEP CONVECTION 1442
and Irminger Seas as identified by the ADT contour 55 cm [Våge et al., 2011; Piron et al., 2016]. The spatial
coverage is relatively homogeneous, and the 88 profiles with MLD >700 m represented 23% of the profiles
located within this ADT contour during January–April 2015.
In the Irminger Sea (east of 42°W), most of the MLDs varied between 1000 and 1400 m. The deepest MLDs
(1400 m) were observed on 17 March and 16 April. In the Labrador Sea and south of Cape Farewell, the
MLDs were deeper than in the Irminger Sea and many profiles had MLDs deeper than 1400 m. The deepest
MLDs observed in the Labrador Sea and south of Cape Farewell were 1790 m (18 March) and 1700 m (15 and
29 March), respectively. To go further in the description of the convection, we considered late winter
MLD >700 m (March–April) and calculated median (Me), first (Q1), and third (Q3) quartiles for the three boxes
defined in Figure 1b and referred to as LAB (Labrador Sea), SCF (south of Cape Farewell), and IRM (Irminger
Sea). Fifty percent of the MLD >700 m were deeper than 1400 m in LAB (Me = 1399 m), and 25% of them
exceeded 1500 m (Q3 = 1502 m). The maximum MLD reached in LAB is similar to the deepest MLDs observed
in the last decade in this area. Similar statistics were obtained for SCF (Me = 1468 m, Q3 = 1603 m) where the
maximum MLD observed (1700 m) is deeper than past observations. For IRM, 75% of the MLDs >700 m were
deeper than the deepest MLD ever observed in that area (1000 m) and 25% of the profiles even exceeded
1300 m (Q1 = 995 m, Q3 = 1316 m).
Potential density and oxygen of the vertical profiles with the deepest (MLD >1000 m) late winter MLDs were
homogeneous from the surface to the base of the mixed layer (Figures 2a–2c). As argued in Piron et al. [2016],
this testifies that the deep convection occurred locally. For those profiles, the mixed layer properties (poten-
tial temperature, salinity, and potential density) were determined at the potential vorticity minimum [Piron
et al., 2016]. The potential density ranged between 27.73 and 27.75 kg m
3
in the three boxes (Figure 2d).
The range is reduced to 27.73–27.74 kg m
3
for the deepest MLDs. In IRM, the MLDs >1000 m were warmer
and saltier than in the Labrador Sea and south of Cape Farewell (3.55–3.70°C against 3.2–3.4°C; 34.88–34.90
against 34.84–34.86). The mixed layers in LAB were also slightly colder (~0.1°C) and fresher (~0.01) than those
located in SCF.
To relate the formation of the deep mixed layers to the atmospheric forcing, we compared the temporal evo-
lution of the mixed layer heat content (HC) and of the cumulative heat flux Q
tot
split into the shortwave and
longwave (Q
rad
), latent (Q
lat
), and sensible (Q
sens
) heat fluxes and horizontal Ekman transport heat fluxes from
September 2014 to April 2015 in the three boxes (Figures 3a–3c). de Boisséson et al. [2010] and Piron et al.
[2016] showed the suitability of ERA-Interim data for such comparison in the subpolar gyre of the North
Atlantic Ocean. The Ekman heat flux is induced by the horizontal Ekman advection acting on horizontal tem-
perature gradients, which are computed, following Piron et al. [2016], from SST data. For convenience, all
terms were initialized at 0 in September. The Ekman component represented at most 10% of the total heat
loss. The latent heat loss dominated in IRM and SCF and represented about 60% of the total heat loss. In LAB,
the latent and sensible heat fluxes represented about half of the total heat flux each. HC was estimated from
the calculation of the mixed layer heat content variation (HCV) between two consecutive months in consid-
ering a MLD equal to the maximum MLD value observed over the 2 months. The procedure is similar to that
used by de Boisséson et al. [2010] along floats trajectories.
The temporal evolution of the mixed layer heat content is explained at first order by the sum of the air-sea
and Ekman transport heat fluxes (Figures 3a–3c). The agreement between HC and Q
tot
is striking in LAB. In
IRM, the total heat loss to the atmosphere was stronger than the mixed layer heat content variations. The dif-
ferences may be due to biases on the air-sea forcing and to other processes that were not taken into account
like entrainment into mixed layer of water from below [Lilly et al., 1999] or horizontal advection of the warm
Atlantic waters from the boundary current [Lazier et al., 2002]. This latter mechanism possibly explains the
mixed layer warming observed in November 2014.
4. Discussion
In agreement with the high NAO index characterizing Winter15, the oceanic air-sea heat loss over the subpo-
lar gyre during Winter15 was 20% to 50% larger than the 1979–2015 climatology (Figure 1b) and 20 to 40%
higher over the three convection sites. Anomalies were not due to anomalously cold (and dry) air tempera-
ture and large sea ice extent (Figure S1 in the supporting information) as in 2007–2008 [Våge et al., 2009] but
Geophysical Research Letters 10.1002/2016GL071895
PIRON ET AL. NORTH ATLANTIC 2014–2015 DEEP CONVECTION 1443
resulted from positive wind speed
anomalies in the 53–63°N latitude
band associated with an intensifica-
tion of the meridional surface pres-
sure gradient [Duchez et al., 2016].
Winter15 atmospheric conditions for
the three deep convection sites are
further characterized by determining
the number of strong wind events
(Figure 4). A strong wind event was
defined by a wind stress larger than
3 times the wind stress standard
deviation estimated over the period
1990–2015 between 1 September
and the date when the net air-sea
heat flux to the ocean starts to be
positive. The calculation was done in
LAB, SCF, and IRM and for all winters
since 1991. Note that in the Irminger
Sea strong wind events correspond
generally but not only to the westerly
Greenland Tip Jets [Pickart et al.,
2003b].
Compared to the other winters of the
period 1990–2015, Winter15 was
exceptional in terms of winter heat
loss and strong wind event occur-
rences (Figure 4). For the three boxes,
Winter15 is the second winter after
Winter93 in terms of strong heat loss
anomaly. It was also among the top
three winters in terms of large number
of strong wind events. Those excep-
tional atmospheric conditions caused
the formation of exceptionally deep
MLDs south of Cape Farewell and in
the Irminger Sea and explain the fact
that deep convection occurred over
an incomparably wide region in the subpolar North Atlantic as already noted by Fröb et al. [2016].
However, the atmospheric conditions alone cannot explain why the late winter MLDs were shallower in the
Irminger Sea than in the Labrador Sea and south of Cape Farewell despite similar winter heat loss amplitudes
in the three boxes (3 × 10
9
J; Figures 3a–3c); why the MLD deepening stopped at 800–1000 m in 2007–2008
[Våge et al., 2008; de Jong et al., 2012] while Winter08 was as exceptional as Winter15 in terms of heat loss and
number of strong wind events (Figure 4); and why the exceptional air-sea forcing did not lead to exception-
ally deep MLDs in the Labrador Sea. To address those questions, we assessed the role of the preconditioning
of the water column for each box by computing the amount of heat that should be extracted (HC
ext
) from an
end of summer profile to homogenize it down to a given depth (Figures 3d and 3e). The computation was
carried out using mean potential temperature profiles calculated from all profiles available in September
2014 in LAB and IRM and in August 2014 in SCF because no profiles were available in this box in
September (Figures 3d and 3e). This supposes that the mixed layer deepening is mainly a one-dimensional
process governed by winter heat loss. This is true at first order for Winter15 (Figures 3a–3c), and this was
proven to be true for other winters in the Labrador Sea [Lilly et al., 1999; Yashayaev and Loder, 2009] and in
the Irminger Sea [Våge et al., 2008].
Figure 4. Number of strong wind events as a function of total heat flux
anomalies cumulated from 1 September to the date when the net air-sea
heat flux to the ocean starts to be positive. The total heat flux is the sum of
air-sea and Ekman heat fluxes. Anomalies were computed relative to the
climatological value estimated over 1979–2015 and equal to 2.6, 2.4, and
2.7 × 0
9
J in (a) LAB, (b) SCF, and (c) IRM, respectively. Years indicated in
Figures 4a–4c correspond to the year of the end of winter.
Geophysical Research Letters 10.1002/2016GL071895
PIRON ET AL. NORTH ATLANTIC 2014–2015 DEEP CONVECTION 1444
In IRM, the HC
ext
curve predicts a MLD of 1400 m that matches the observed maximum MLD when consider-
ing the observed Winter15 heat loss (Figure 3e). The predicted MLD corresponds also to the top of the deep
thermocline that acts as a barrier for a further deepening of the MLD. A heat loss almost twice as large as the
one observed during Winter15 would have been necessary to reach a MLD in the Irminger Sea as deep as
those observed in the Labrador Sea (1800 m). Considering the already exceptional character of Winter15,
1400 m is obviously the deepest MLD that can be reached with the current preconditioning. By contrast,
the barrier effect due to the deep thermocline was not the limiting factor for the further deepening of the
mixed layers in Winter08 because, as argued by de Jong et al. [2012], the density gradient at the base of
the end of winter mixed layers was weak. The exceptional heat loss of Winter08 in IRM (Figure 4c) removed
the strong initial temperature stratification induced by the mild winters of the early 2000s. The surface cool-
ing ceased or weakened sufficiently in March for restratification by advection of warmer waters from
surrounding boundary currents to take place, and no further deepening of the mixed layers was observed
[de Jong et al., 2012]. In LAB and SCF, the maximum Winter15 MLDs were shallower by about 150 m than
the predicted MLD (Figure 3e) but matched the depth of the deep thermocline. In addition, late winter deep
mixed layers that reached 1700–1800 m in those boxes were colder than the water mass found at the same
levels in August–September 2014 by about 0.15°C. The heat extracted from the ocean in those two boxes
during the whole winter was used to deepen the mixed layers down to the top of the deep thermocline
and to cool it.
A cooling of the upper part of the water column is observed between September 2014 and September 2015
in the three basins (Figure S2). It is consistent with the cold SST anomaly computed compared to the period
2000–2015 that was present in June 2015 over the whole subpolar gyre and that was attributed to the strong
atmospheric forcing that prevailed during Winter15 [Duchez et al., 2016; de Jong and de Steur, 2016].
Interestingly, a cold SST anomaly of about 1–2°C was already present in the subpolar gyre in November
2014 [Duchez et al., 2016] and has likely favored the positive NAO phase observed in Winter15 as such SST
anomaly was proven to influence the phase of the NAO [Hurrell and Deser, 2009; Scaife et al., 2014]. This cold
anomaly was considered as a reemergence of subsurface anomaly from the preceding winter and was
present down to 700 m (Figure S2; Duchez et al., 2016). Being not salinity compensated, it caused a density
increase in the upper 700 m between September 2013 and September 2014 (not shown) and acted as a
preconditioning of the water column for deep convection, especially in IRM. Indeed, HC
ext
computed from
2013 summer conditions but with a heat loss of 3 × 10
9
J as that observed in Winter15 suggests that MLDs
would have only reached 800 m without the preconditioning by the cold anomaly (Figure 3e). The cold
anomaly that developed recently in the subpolar gyre of the North Atlantic Ocean favored the intense
Winter15 deep convection, which in turn enhanced this cold anomaly.
Despite the exceptional character of Winter15, which was similar in terms of atmospheric conditions to those
of the early 1990s, deep convection in LAB was shallower during Winter15 than during the early 1990s
winters when it reached 2400 m. To understand this difference, we computed HC
ext
from summer profiles
collected in 1993 and 1994 along the AR7W line in LAB [Yashayaev, 2007] (not shown). We found that, with
similar air-sea heat loss, the vertical extent of the deep convection could be much larger in the early 1990s
than in Winter15 because the deep thermocline lay at 2300–2400 m during this period. In addition, the heat
loss required to deepen the mixed layers down to the deep thermocline depth in the early 1990s (about
1×10
9
J) was much less than the observed heat loss (more than 3 × 10
9
J) (Figure 4a). This would explain
the LSW cooling trend observed in the early 1990s [Yashayaev, 2007].
5. Conclusions
During winter 2014–2015, a major deep convective activity forced by exceptional turbulent air-sea fluxes
occurred over an area extending from the interior of the Labrador Sea to the Irminger Sea. It was favored
by and enhanced the cold anomaly that developed recently in the subpolar gyre of the North Atlantic
Ocean. The MLDs reached 1800 m in the Labrador Sea, which is similar to the deepest MLDs observed since
the mid-1990s but shallower than those observed in the early 1990s. The MLDs reached 1700 m and 1400m
south of Cape Farewell and in the Irminger Sea, respectively, which are the deepest MLD directly
observed in those two regions. The three basins formed LSW of same density but with different tempera-
tures and salinities. In the Irminger Sea, the MLDs extended down to the base of the LSW pool (1400 m)
Geophysical Research Letters 10.1002/2016GL071895
PIRON ET AL. NORTH ATLANTIC 2014–2015 DEEP CONVECTION 1445
[Daniault et al., 2016] so that the convective patches contributed to the renewal of the LSW over its full depth
range. The mixed layer deepening in the Labrador Sea was limited by the deep thermocline density gradient,
and a significant cooling of the LSW compared to previous year was observed. By contrast, the MLDs observed
south of Cape Farewell and in the Irminger Sea nearly correspond to the MLDs predicted by applying surface
fluxes to the end of summer heat content. We cannot exclude that such exceptional widespread deep
convection occurred before, most likely in the early 1990s when winters were as exceptional as winter
2014–2015 and when convection intensity was at a maximum in the Labrador Sea [Yashayaev, 2007], but
sparse observations prevent to confirm it. One may also wonder whether this enhanced LSW production
was associated with an intensification of the MOC [Rahmstorf et al., 2015].
If harsh winters regularly occur in the future as observed since 2008, we can expect that the deep thermocline
will continue to be eroded and cooled and that a new dense and cold LSW variety will be formed. Those
changes in the strength of deep convection and on the properties of the LSW are not surprising in the
subpolar gyre of the North Atlantic, which is known to have exhibited significant variability over the past five
decades [Yashayaev, 2007]. However, we might also expect that reduced winter oceanic heat loss and
increased inflow of freshwater from Greenland and the Nordic Seas in a warming world could modify the
natural deep convection cycle [Yang et al., 2016; Brodeau and Koenigk, 2016] and impact the MOC
[Rahmstorf et al., 2015]. Some climate model simulations even predicted the complete extinction of deep
convection in the Labrador Seas after the 2020s [Brodeau and Koenigk, 2016]. In this context, using models
with the Argo observations of deep convection as benchmarks could lead to better understanding of the role
of the various parameters involved in deep convection, which is more than ever required to improve their
representation in climate models.
References
Argo (2015), Argo float data and metadata from Global Data Assembly Centre (Argo GDAC)—Snapshot of Argo GDAC of May, 8th 2015.
SEANOE, doi:10.17882/42182#42336.
Bacon, S., W. J. Gould, and Y. Jia (2003), Open-ocean convection in the Irminger Sea, Geophys. Res. Lett.,30(5), 1246, doi:10.1029/
2002GL016271.
Brodeau, L., and T. Koenigk (2016), Extinction of the northern oceanic deep convection in an ensemble of climate model simulations of the
20th and 21st centuries, Clim. Dyn.,46, 2863–2882, doi:10.1007/s00382-015-2736-5.
Cunningham, S. A., et al. (2007), Temporal variability of the Atlant ic meridional overturning circulation at 26.5 N, Science,317, 935–937.
Daniault, N., et al. (2016), The northern North Atlantic Ocean mean circulation in the early 21st Century, Prog. Oceanogr.,146, 142–158,
doi:10.1016/j.pocean.2016.06.007.
de Boisséson, E., V. Thierry, H. Mercier, and G. Caniaux (2010), Mixed layer heat budget in the Iceland Basin from Argo, J. Geophys. Res.,115
C10055, doi:10.1029/2010JC006283.
de Boyer Montégut, C., G. A. Madec, S. Fischer, A. Lazar, and D. Iudicone (2004), Mixed layer depth over the global ocean: An examination of
profile data and a profile-based climatology, J. Geophys. Res.,109, C12003, doi:10.1029/2004JC002378.
de Jong, M. F., and L. de Steur (2016), Strong winter cooling of the Irminger Sea in winter 2014–15, exceptional deep convection, and the
emergence of anomalously low SST, Geophys. Res. Lett.,43, 1717–1734, doi:10.1002/2016GL069596.
de Jong, M. F., H. M. van Aken, K. Våge, and R. S. Pickart (2012), Convective mixing in the central Irminger Sea: 2002–2010, Deep Sea Res., I,
63(1), 36–51.
Dee, D. P., et al. (2011), The ERA-Interim reanalysis: Configuration and performance of the data assimilation system, Q. J. R. Meteorol. Soc. ,137,
553–597, doi:10.1002/qj.828.
Duchez, A., E. Frajka-Williams, S. A. Josey, D. G. Evans, J. P. Grist, R. Marsh, G. D. McCarthy, B. Sinha, D. I. Berry, and J. J.-M. Hirschi (2016), Drivers
of exceptionally cold North Atlantic Ocean temperatures and their link to the 2015 European heat wave, Environ. Res. Lett.,11,7,
doi:10.1088/1748-9326/11/7/074004.
Fetterer, F., K. Knowles, W. Meier, and M. Savoie. 2016, updated daily. Sea Ice Index, Version 2, Natl. Snow and Ice Data Cent., Boulder, Colo.,
doi:10.7265/N5736NV7.
Fröb, F., A. Olsen, K. Våge, G. Moore, I. Yashayaev, E. Jeansson, and B. Rajasakaren (2016), Irminger Sea deep convection injects oxygen and
anthropogenic carbon to the ocean interior, Nat. Commun.,7, 13,244, doi:10.1038/ncomms13244.
Gaillard, F., T. Reynaud, V. Thierry, N. Kolodziejczyk, and K. von Schuckmann (2016), In-situ based reanalysis of the global ocean temperature
and salinity with ISAS: Variability of the heat content and steric height, J. Clim.,29, 1305–1323, doi:10.1175/JCLI-D-15-0028.1.
Garcia, H. E., R. A. Locarnini, T. P. Boyer, J. I. Antonov, O. K. Baranova, M. M. Zweng, and D. R. Johnson (2010), in World Ocean Atlas 2009, Volume
3: Dissolved Oxygen, Apparent Oxygen Utilization, and Oxygen Saturation, NOAA Atlas NESDIS 70, edited by S. Levitus, U.S. Gov. Print. Off.,
Washington, D. C.
Hurrell, J. W. (1995), Decadal trends in the North Atlantic Oscillation: Regional temperatures and precipitation, Science,269, 676–679,
doi:10.1126/science.269.5224.676.
Hurrell, J. W., and C. Deser (2009), North Atlantic climate variability: The role of the North Atlantic Oscillation, J. Mar. Syst.,78,28–41,
doi:10.1016/j.jmarsys.2008.11.026.//000268701200004.
Kieke, D., and I. Yashayaev (2015), Studies of Labrador Sea Water formation and variability in the subpolar North Atlantic in the light of
international partnership and collaboration, Prog. Oceanogr.,132, 220–232, doi:10.1016/j.pocea n.2014.12.010.
Lazier, J., R. Hendry, A. Clarke, I. Yashayaev, and P. Rhines (2002), Convection and restratification in the Labrador Sea, 1990–2000, Deep Sea
Res., I,49, 1819–1835.
Geophysical Research Letters 10.1002/2016GL071895
PIRON ET AL. NORTH ATLANTIC 2014–2015 DEEP CONVECTION 1446
Acknowledgments
We thank the reviewers for their con-
structive comments. Anne Piron and
Virginie Thierry are funded by IFREMER
(Institut Français de Recherche pour
l’Exploitation de le Mer), Herlé Mercier is
funded by CNRS (the French Centre
National de la Recherche Scientifique),
and Guy Caniaux is funded by
Météo-France. This paper is a contribu-
tion to the EQUIPEX NAOS project
funded by the French National Research
Agency (ANR) under reference ANR-10-
EQPX-40, to the AtlantOS project
(H2020), to the OVIDE project supported
by IFREMER, CNRS and INSU (Institut
National des Sciences de l’Univers), and
by French national programs (INSU/
LEFE). OVIDE is a contribution to CLIVAR.
This study is also a contribution to the
project BOCATS (CTM2013-41048-P)
supported by the Spanish Ministry of
Economy and Competitiveness (BES-
2014-070449) and cofunded by the
Fondo Europeo de Desarrollo Regional
2007–2012 (FEDER) This work bene-
fitted from the French research infra-
structures Euro-Argo and flotte (TGIR
EuroArgo and TGIR flotte) and more
specifically from the OVIDE cruises rea-
lized on board the French RV Thalassa
and during which many Argo floats
used in that study were deployed. Argo
data were collected and made freely
available by the International Argo
Program and the national programs
that contribute to it (http://www.argo.
ucsd.edu, http://argo.jcommops.org).
Argo data used in that study were
downloaded from the Coriolis Data
Center (http://www.coriolis.eu.org) and
are listed in the references. Surface wind
stress, air-sea heat flux, and SST were
downloaded from the ERA-Interim
website (http://www.ecmwf.int/en/
research/climate-reanalysis/era-
interim). Absolute Dynamic Topography
data were downloaded from AVISO
(Archiving, Validation, and
Interpretation of Satellite
Oceanographic data) web site (http://
www.aviso.oceanobs.com/duacs/). Sea
ice data were obtained from the Centre
de Recherche et d’Exploitation
Satellitaire (CERSAT), at IFREMER,
Plouzané (France) (http://cersat.ifremer.
fr/data/products/catalogue). ISAS maps
[Gaillard et al., 2016] were downloaded
from the ISAS viewer website (http://
www.umr-lops.fr/SO-Argo/Products/
ISAS-T-S-fields/ISAS-viewer).
Lilly, J. M., P. B. Rhines, M. Visbeck, R. Davis, J. R. Lazier, F. Schott, and D. Farmer (1999), Observing deep convection in the Labrador Sea during
winter 1994–1995, J. Phys. Oceanogr.,29, 2065–2098.
Mercier, H., P. Lherminier, A. Sarafanov, F. Gaillard, N. Daniault, D. Desbruyères, A. Falina, B. Ferron , T. Huck, and V. Thierry (2015), Variability of
the meridional overturning circulation at the Greenland–Portugal OVIDE section from 1993 to 2010, Prog. Oceanogr.,132, 250–261,
doi:10.1016/j.pocean.2013.11.001.
Pérez, F. F., H. Mercier, M. Vazquez-Rodriguez, P. Lherminier, A. Velo, P. Pardo, G. Roson, and A. Rios (2013), Reconciling air-sea CO
2
fluxes and
anthropogenic CO
2
budgets in a changing North Atlantic, Nat. Geosci.,6, 146–152, doi:10.1038/ngeo1680.
Pickart, R. S., D. J. Torres, and R. A. Clarke (2002), Hydrography of the Labrador Sea during active convection, J. Phys. Oceanogr.,32, 428–457.
Pickart, R. S., F. Straneo, and G. W. K. Moore (2003a), Is Labrador Sea Water formed in the Irminger basin?, Deep Sea Res., I,50,23–52,
doi:10.1016/S0967-0637(02)00134-6.
Pickart, R. S., M. A. M. Spall, H. Ribergaard, G. W. K. Moore, and R. F. Milliff (2003b), Deep convection in the Irminger sea forced by the
Greenland tip jet, Nature,424, 152–156, doi:10.1038/nature01729.
Piron, A., V. Thierry, H. Mercier, and G. Caniaux (2016), Argo float observations of basin-scale deep convection in the Irminger sea during
winter 2011–2012, Deep Sea Res., I,109,76–90, doi:10.1016/j.dsr.2015.12.012.
Rahmstorf, S., J. E. Box, G. Feulner, M. E. Mann, A. Robinson, S. Rutherford, and E. J. Schaffernicht (2015), Exceptional twentieth-century
slowdown in Atlantic Ocean overturning circulation, Nat. Clim. Change,5, 475–480, doi:10.1038/nclimate2554.
Scaife, A. A., et al. (2014), Skillful long-range prediction of European and North American winters, Geophys. Res. Lett.,41, 2514–2519,
doi:10.1002/2014GL059637.
Takeshita, Y., T. R. Martz, K. S. Johnson, J. N. Plant, D. Gilbert, S. C. Riser, C. Neill, and B. Tilbrook (2013), A climatology-based quality control
procedure for profiling float oxygen data, J. Geophys. Res. Oceans,118, 5640–5650, doi:10.1002/jgrc.20399.
Thomson, R. E., and I. V. Fine (2003), Estimating mixed layer depth from oceanic profile data, J. Atmos. Oceanic Tech.,20, 319–329,
doi:10.1175/1520-0426(2003)020<0319:EMLDFO>2.0.CO;2.
Våge, K., R. S. Pickart, G. W. K. Moore, and M. H. Ribergaard (2008), Winter mixed layer development in the central Irminger Sea: The effect of
strong, intermittent wind events, J. Phys. Oceanogr.,38, 541–565, doi:10.1175/2007JPO3678.1.
Våge, K., R. S. Pickart, V. Thierry, G. Reverdin, C. M. Lee, B. Petrie, T. A. Agnew, A. Wong, and M. H. Ribergaard (2009), Surpri sing return of deep
convection to the subpolar North Atlantic Ocean in winter 2007–2008, Nat. Geosci.,2,67–72, doi:10.1038/ngeo382.
Våge, K., R. S. Pickart, A. Sarafanov, O. Knutsen, H. Mercier, P. Lherminier, H. M. van Aken, J. Meincke, D. Quadfasel, and S. Bacon (2011), The
Irminger Gyre: Circulation, convection, and interannual variability, Deep Sea Res., I,58(5), 590–614, doi:10.1016/j.dsr.2011.03.001.
Yang, Q., T. H. Dixon, P. G. Myers, J. Bonin, D. Chambers, and M. R. van den Broeke (2016), Recent increases in Arctic freshwater flux affects
Labrador Sea convection and Atlantic overturning circulation, Nat. Commun.,7(10525), 2016, doi:10.1038/ncomms10525.
Yashayaev, I. (2007), Hydrographic changes in the Labrador Sea, 1960–2005, Prog. Oceanogr.,73, 242–276.
Yashayaev, I., and J. W. Loder (2009), Enhanced production of Labrador Sea Water in 2008, Geophys. Res. Lett.,36, L01606, doi:10.1029/
2008GL036162.
Geophysical Research Letters 10.1002/2016GL071895
PIRON ET AL. NORTH ATLANTIC 2014–2015 DEEP CONVECTION 1447
