Editing IPCC:AR6/SROCC/Chapter-3 (section)

==== 3.2.1.2 Ocean Properties ====

<div id="section-3-2-1-2-ocean-properties-block-1"></div>

<span id="temperature"></span>
===== 3.2.1.2.1 Temperature =====

Ocean temperatures and associated heat fluxes have a primary influence on sea ice (e.g., Carmack et al., 2015; Steele and Dickinson, 2016 <sup>[[#fn:r218|218]]</sup> ). WGI AR5 (their Section 3.2.2) reported that Canada Basin surface waters warmed from 1993 to 2007, and observations over 1950–2010 show the Arctic Ocean water of Atlantic origin (i.e., the Atlantic Water Layer) warming starting in the 1970s. Warming trends have continued: August trends for 1982–2017 reveal summer mixed layer temperatures increasing at about 0.5°C per decade over large sectors of the Arctic basin that are ice-free in summer (Timmermans et al., 2017 <sup>[[#fn:r219|219]]</sup> ) (Figure 3.3). This is primarily the result of increased absorption of solar radiation accompanying sea ice loss (Perovich, 2016 <sup>[[#fn:r220|220]]</sup> ). Between 1979 and 2011, the decrease in Arctic Ocean albedo corresponded to more solar energy input to the ocean ( ''virtually certain'' ) of approximately 6.4 ± 0.9 Wm <sup>-2</sup> (Pistone et al., 2014 <sup>[[#fn:r221|221]]</sup> ), ''likely'' reducing the growth of sea ice by up to 25% in both Eurasian and Canadian basins (Timmermans, 2015 <sup>[[#fn:r222|222]]</sup> ; Ivanov et al., 2016 <sup>[[#fn:r223|223]]</sup> ) (Section 3.2.1.1).

While Atlantic Water Layer temperatures appear to show less variability since 2008, total heat content in this layer continues to increase (Polyakov et al., 2017 <sup>[[#fn:r224|224]]</sup> ). Recent changes have been dubbed the ‘Atlantification’ of the Northern Barents Sea and Eurasian Basin (Arthun et al., 2012 <sup>[[#fn:r225|225]]</sup> ; Lind et al., 2018 <sup>[[#fn:r226|226]]</sup> ), characterised by weaker stratification and enhanced Atlantic Water Layer heat fluxes further northeast ( ''medium confidence'' ). Polyakov et al. (2017) <sup>[[#fn:r227|227]]</sup> estimate 2–4 times larger heat fluxes in 2014–2015 compared with 2007–2008. In the Canadian Basin, the maximum temperature of the Pacific Water Layer increased by ~0.5°C between 2009 and 2013 (Timmermans et al., 2014 <sup>[[#fn:r228|228]]</sup> ), with a doubling in integrated heat content over 1987–2017 (Timmermans et al., 2018 <sup>[[#fn:r229|229]]</sup> ). Over 2001–2014, heat transport associated with Bering Strait inflow increased by 60%, from around 10 TW in 2001 to 16 TW in 2014, due to increases in both volume flux and temperature (Woodgate et al., 2015 <sup>[[#fn:r230|230]]</sup> ; Woodgate, 2018 <sup>[[#fn:r231|231]]</sup> ) ( ''low confidence'' ).

The Southern Ocean is important for the transfer of heat from the atmosphere to the global ocean, including heat from anthropogenic warming (Frölicher et al., 2015 <sup>[[#fn:r232|232]]</sup> ; Shi et al., 2018 <sup>[[#fn:r233|233]]</sup> ). The Southern Ocean accounted for ~75% of the global ocean uptake of excess heat during 1870–1995 (Figure SM3.2; Frölicher et al., 2015 <sup>[[#fn:r234|234]]</sup> ), of which ~43% resided in the Southern Ocean with the remainder redistributed to lower latitudes. Over 1970–2017, observations show that the upper 2000 m of the ocean south of 30°S was responsible for 35–43% of the increase in global ocean heat content (Table 3.1). Both models and observations show that, relative to its size (Table SM3.1), the Southern Ocean is disproportionately important in the increase in global upper ocean heat content ( ''high confidence'' ). Multi-decadal warming of the Southern Ocean has been attributed to anthropogenic factors, especially the role of greenhouse gases but also ozone depletion (Armour et al., 2016 <sup>[[#fn:r235|235]]</sup> ; Shi et al., 2018 <sup>[[#fn:r236|236]]</sup> ; Swart et al., 2018 <sup>[[#fn:r237|237]]</sup> ; Irving et al., 2019 <sup>[[#fn:r238|238]]</sup> ) ( ''medium confidence'' ).

<div id="section-3-2-1-2-ocean-properties-block-2"></div>

<span id="table-3.1."></span>
<!-- START IMG -->
<!-- TABLE IMG -->
<!-- IMG TITLE -->
'''Table 3.1.'''
<!-- IMG CAPTION -->
Ocean heat content trend (0–2000 m depth) during 2005–2017 and 1970–2017 for the global ocean and Southern Ocean. Ordinary Least Square (OLS) method is used; units are 10 21 J yr -1 . Uncertainties denote the 90% confidence interval accounting for the reduction in the degrees of freedom implied by temporal correlations of residuals, as per Section 5.2. Values in curved brackets are percentages of heat gain by the Southern Ocean relative to the global ocean. Data sources are as per Table SM3.1. The mean proportion and its 5–95% confidence interval (1.65 times standard deviation of individual estimates) are in the last column.
<!-- IMG FILE -->
[[File:0d2d12aa9b463e5f579b6e869ce36a80 table3.1.png]]

Surface warming during 1982–2016 was strongest along the northern flank of the ACC '','' contrasting with cooling further south (Figure 3.3). Interior warming was strongest in the upper 2000 m, peaking around 40°S–50°S (Armour et al., 2016 <sup>[[#fn:r239|239]]</sup> ) (SM3.2.1; Figures SM3.2 and SM3.3). There is ''high confidence'' that this pattern of change is driven by upper-ocean overturning circulation and mixing (Cross-Chapter Box 7 in Chapter 3), whereby heat uptake at the surface by newly upwelled waters is transmitted to the ocean interior in intermediate depth layers (Armour et al., 2016 <sup>[[#fn:r240|240]]</sup> ). Whilst temperature trends in the ACC itself are driven predominantly by air-sea flux changes (Swart et al., 2018 <sup>[[#fn:r241|241]]</sup> ), the warming on its northern side appears strongly influenced by wind-forced changes in the thickness and depth of the mode water layer (Desbruyeres et al., 2017 <sup>[[#fn:r242|242]]</sup> ; Gao et al., 2018 <sup>[[#fn:r243|243]]</sup> ) ( ''medium confidence)'' . Below the surface south of the ACC, warming extends close to Antarctica, intruding onto the continental shelf in the Amundsen-Bellingshausen Sea where temperature increases of 0.1°C–0.3°C per decade have been observed over 1983–2012 (Schmidtko et al., 2014 <sup>[[#fn:r244|244]]</sup> ) (Section 3.3.1.5). This latter warming may be driven by changes in wind forcing (Spence et al., 2014 <sup>[[#fn:r245|245]]</sup> ), and exhibits significant decadal variability (Jenkins et al., 2018 <sup>[[#fn:r246|246]]</sup> ).

After around 2005, improved upper ocean heat content estimates became available via Argo profiling floats (Section 1.8.1; Section 5.2). For 2005–2017, multiple datasets show that the heat gained by the Southern Ocean south of 30°S was 45–62% of the global ocean heat gain (Table 3.1) (equivalent figures for other indicative Southern Ocean extents are in Table SM3.2). This accords with Roemmich et al. (2015) <sup>[[#fn:r247|247]]</sup> , who found that during 2006–2013 the ocean south of 20°S accounted for 67–98% of total heat gain in the upper 2000 m of the global ocean. (The smaller proportion for 2005–2017 c.f. 2006–2013 is due to comparatively greater warming in the earlier part of the common period). The recent Southern Ocean heat gain is thus larger than its long-term trend over either the preceding several decades (1970–2004, 30–51%, Table SM3.3) or the full period 1970–2017 (35–43%; Table 3.1 and above). There is ''high confidence'' that the Southern Ocean has increased its role in global ocean heat content in recent years compared with the past several decades. Attribution of this increased role is currently lacking.

The ocean below 2000 m globally stores ~19% of the excess anthropogenic heat in the Earth system, with a large fraction (6% of global total heat excess) located in the deep Southern Ocean south of 30°S (Frölicher et al., 2015 <sup>[[#fn:r248|248]]</sup> ; Talley et al., 2016 <sup>[[#fn:r249|249]]</sup> ) ( ''medium confidence)'' . The WGI AR5-quantified warming of these waters was recently updated (Desbruyeres et al., 2017 <sup>[[#fn:r250|250]]</sup> ) to an equivalent heat uptake of 0.07 ± 0.06 W m −2 below 2000 m since the beginning of the century, resulting in an extra 34 ± 14 TW south of 30°S from 1980 to 2012 (Purkey and Johnson, 2013 <sup>[[#fn:r251|251]]</sup> ) ''.'' Antarctic Bottom Water volume is decreasing (Purkey and Johnson, 2012 <sup>[[#fn:r252|252]]</sup> ), resulting in a deepening of density surfaces and driving much of the warming on depth surfaces below 2000 m (Desbruyeres et al., 2017 <sup>[[#fn:r253|253]]</sup> ). This reduction in bottom water volume is suggestive of a decrease in its production (Purkey and Johnson, 2013 <sup>[[#fn:r254|254]]</sup> ). In the Indian and Pacific basins close to Antarctica, bottom water is freshening (Purkey and Johnson, 2013 <sup>[[#fn:r255|255]]</sup> ; Menezes et al., 2017 <sup>[[#fn:r256|256]]</sup> ) consistent with the uptake of enhanced Antarctic ice shelf and glacial melt (Purkey and Johnson, 2013 <sup>[[#fn:r257|257]]</sup> ).

<!-- END IMG -->
<div id="section-3-2-1-2-ocean-properties-block-3"></div>

<span id="salinity"></span>
===== 3.2.1.2.2 Salinity =====

Salinity is the dominant determinant of polar ocean density, and exerts major controls on stratification, circulation and mixing. Salinity changes are induced by freshwater runoff to the ocean (rivers and land ice), net precipitation, sea ice, and advection of mid-latitude waters, with the potential to impact water mass formation and circulation (e.g., Thornalley et al., 2018; see also Section 6.7.1).

Updating WGI AR5 (their Section 3.3.3.3), recent Arctic-wide estimates yield a freshwater increase (relative to salinity of 34.8 on the Practical Salinity Scale, used throughout this chapter) of 600 ± 300 km 3 yr –1 over 1992–2012, with about two-thirds concomitant with decreasing salinity, and the remainder with a thickening of the freshwater layer ( ''medium confidence'' ) (Rabe et al., 2014 <sup>[[#fn:r258|258]]</sup> ; Haine et al., 2015 <sup>[[#fn:r259|259]]</sup> ; Carmack et al., 2016 <sup>[[#fn:r260|260]]</sup> ). The Beaufort Gyre region has increased its freshwater by ~40% (6600 km 3 ) over 2003–2017; this, and the Gyre’s strengthening, have been attributed to dominance of clockwise wind patterns over the Canadian Basin over 1997–2016 and freshwater accumulation from sea ice-melt (Krishfield et al., 2014 <sup>[[#fn:r261|261]]</sup> ; Proshutinsky et al., 2015 <sup>[[#fn:r262|262]]</sup> ). Freshwater decreases in the East Siberian, Laptev, Chukchi and Kara seas are estimated to be ~180 km 3 over 2003–2014 (Armitage et al., 2016 <sup>[[#fn:r263|263]]</sup> ). During the 2000s, freshwater content in the upper 100 m of the northern Barents Sea declined by about 32%, from a mean of ~2.5 m (relative to a salinity of 35) in 1970–1999, to 1.7 m in 2010–2016 (Lind et al., 2018 <sup>[[#fn:r264|264]]</sup> ). An increasing trend of 30 ± 20 km 3 yr –1 in freshwater flux through Bering Strait, primarily due to increased volume flux, was measured from 1991 to 2015, with record maximum freshwater influx in 2014 of around 3500 km 3 in that year (Woodgate, 2018 <sup>[[#fn:r265|265]]</sup> ). Freshwater fluxes from rivers are also increasing (Section 3.4.1.2.2), and there have been observed increases in discharge of glacial ice from Greenland (Section 3.3.1.3).

Observed Southern Ocean freshening trends are consistent with WGI AR5; subsequent studies have increased confidence in their magnitude and sign, and also attributed them to anthropogenic influences (Swart et al., 2018 <sup>[[#fn:r266|266]]</sup> ). Changes over 1950–2010 show persistent surface water freshening over the whole Southern Ocean, with subducted mode/intermediate waters carrying trends of 0.0002–0.0008 yr –1 to below 1500 m (Skliris et al., 2014 <sup>[[#fn:r267|267]]</sup> ), whilst de Lavergne et al. (2014) observe a circumpolar freshening south of the ACC of 0.0011 ± 0.0004 yr –1 in the upper 100 m since the 1960s ( ''medium confidence'' ). This intensifies over the Antarctic continental shelves (except along the Western Antarctic Peninsula), where freshening of up to 0.01 yr –1 is observed (Schmidtko et al., 2014 <sup>[[#fn:r269|269]]</sup> ). Freshening may be driven by increases in precipitation, but while models (Pauling et al., 2016 <sup>[[#fn:r270|270]]</sup> ) and observations suggest an increase may have occurred over the last 60 years, uncertainty is presently too high to quantify its net impact (Skliris et al., 2014 <sup>[[#fn:r271|271]]</sup> ). Recently, there has been increased recognition of the importance of sea ice in driving Southern Ocean salinity changes, with Haumann et al. (2016) <sup>[[#fn:r272|272]]</sup> demonstrating that wind driven sea ice export has increased by 20 ± 10% from 1982 to 2008, and that this may have driven freshening of 0.002 ± 0.001 yr –1 in the surface and intermediate waters. Separately, the central role of sea ice in driving water mass transformations in the Southern Ocean has been highlighted (Abernathey et al., 2016 <sup>[[#fn:r273|273]]</sup> ; Pellichero et al., 2018 <sup>[[#fn:r274|274]]</sup> ; Swart et al., 2018 <sup>[[#fn:r275|275]]</sup> ), hence such changes have the potential to affect overturning circulation (Cross-Chapter Box 7 in Chapter 3). Freshwater input to the ocean from the Antarctic Ice Sheet also has the potential to affect the properties and circulation of Southern Ocean water masses; see Section 3.3.3.

<div id="section-3-2-1-2-ocean-properties-block-4"></div>

<span id="stratification"></span>
===== 3.2.1.2.3 Stratification =====

See Supplementary Material (SM3.2.2).

<div id="section-3-2-1-2-ocean-properties-block-5"></div>

<span id="carbon-and-ocean-acidification"></span>
===== 3.2.1.2.4 Carbon and ocean acidification =====

Various elements of marine biogeochemistry and geochemistry in the polar regions are of global importance. Here we focus on aspects relevant to carbon and ocean acidification; others (e.g., changes in dissolved oxygen) are assessed in Section 5.2.2. Compiled datasets on observed trends in ocean acidification from different observational platforms can be found in Table SM5.3.

About a quarter of carbon dioxide (CO 2 ) released by human activities is taken up by the ocean (WGI AR5, their Section 3.8). This dissolves in surface water to form carbonic acid, which, upon dissociation, causes a decrease in pH (acidification) and carbonate ion (CO 3 2– ) concentration. This can affect organisms that form shells and skeletons using calcium carbonate (CaCO 3 , aragonite and calcite as dominant mineral forms). Since AR5, new observations have demonstrated the spatial and temporal variability of ocean acidification and controlling mechanisms of carbon systems in different regions (Bellerby et al., 2018 <sup>[[#fn:r276|276]]</sup> ).

Robbins et al. (2013) <sup>[[#fn:r277|277]]</sup> showed aragonite undersaturation for about 20% of surface waters in the Canada and Makarov Basins, where substantial sea ice melt occurred. Qi et al. (2017) <sup>[[#fn:r278|278]]</sup> reported that aragonite undersaturation has expanded northward by at least 5° of latitude, and deepened by ~100 m between the 1990s and 2010 primarily due to increased Pacific Winter Water transport. In the East Siberian Arctic Shelf, extreme aragonite undersaturation was driven by the degradation of terrestrial organic matter and runoff of Arctic river water with elevated CO 2 concentrations, reflecting pH changes in excess of those projected in this region for 2100 (Semiletov et al., 2016 <sup>[[#fn:r279|279]]</sup> ) ( ''high confidence'' ); this was also observed along the continental margin and traced in the deep Makarov and Canada Basins (Anderson et al., 2017a <sup>[[#fn:r280|280]]</sup> ). The variable buffering capacities of rivers flowing through watersheds with different bedrock geology also influenced the state of ocean acidification in coastal regions (Tank et al., 2012 <sup>[[#fn:r281|281]]</sup> ; Azetsu-Scott et al., 2014 <sup>[[#fn:r282|282]]</sup> ).

The dissolved inorganic carbon (DIC) concentration increased in subsurface waters (150–1400 m) in the central Arctic between 1991 and 2011 (Ericson et al., 2014 <sup>[[#fn:r283|283]]</sup> ). The rate of increase was 0.6–0.9 µmol kg –1 yr –1 in the Arctic Atlantic Water and 0.4–0.6 µmol kg –1 yr –1 in the upper Polar Deep Water due to anthropogenic CO 2 , while no trend was observed in nutrient concentrations. In waters below 2000 m, no significant trend was observed for DIC and nutrient concentrations. Observation-based estimates (MacGilchrist et al., 2014 <sup>[[#fn:r284|284]]</sup> ) revealed a net summertime pan-Arctic export of 231 ± 49 TgC yr –1 of DIC across the Arctic Ocean gateways to the North Atlantic; at least 166 ± 60 TgC yr –1 of this was sequestered from the atmosphere ( ''medium confidence)'' . Similar to other regions (Table SM5.3), observed changes in the carbonate chemistry of the Arctic are indicative of ongoing ocean acidification ( ''high confidence'' ).

Studies covering seasonal-to-decadal variability in the Arctic are limited, with most conducted in ice-free or low ice periods during summer to autumn. However, it has been demonstrated that biological processes, respiration and photosynthesis, control the CaCO 3 saturation states in Chukchi Sea bottom water (Yamamoto-Kawai et al., 2016 <sup>[[#fn:r285|285]]</sup> ). Sea ice formation and melt influence the dynamics of ikaite (CaCO 3 precipitation trapped in sea ice during brine rejection), and therefore local carbonate chemistry (Rysgaard et al., 2013 <sup>[[#fn:r286|286]]</sup> ; Bates et al., 2014 <sup>[[#fn:r287|287]]</sup> ; Geilfus et al., 2016 <sup>[[#fn:r288|288]]</sup> ; Fransson et al., 2017 <sup>[[#fn:r289|289]]</sup> ). Although the increase of pH and saturation states by biological carbon fixation that consumes DIC in surface water is well documented (Azetsu-Scott et al., 2014 <sup>[[#fn:r290|290]]</sup> ; Yamamoto-Kawai et al., 2016 <sup>[[#fn:r291|291]]</sup> ) ( ''high confidence)'' , it has been shown that long photoperiods in Arctic summers sustain high pH in kelp forests, slowing ocean acidification (Krause-Jensen et al., 2016 <sup>[[#fn:r292|292]]</sup> ).

Since AR5, there are new constraints on the seasonal-to-decadal variability in the Southern Ocean CO 2 flux (McNeil and Matear, 2013 <sup>[[#fn:r293|293]]</sup> ; Landschützer et al., 2014 <sup>[[#fn:r294|294]]</sup> ; Landschützer et al., 2015 <sup>[[#fn:r295|295]]</sup> ; Gregor et al., 2017 <sup>[[#fn:r296|296]]</sup> ; Ritter et al., 2017 <sup>[[#fn:r297|297]]</sup> ; Keppler and Landschutzer, 2019 <sup>[[#fn:r298|298]]</sup> ) (Figure SM3.4), with mean annual flux anomalies varying from 0.3 ± 0.1 Pg C yr –1 in 2001–2002 to –0.4 Pg C yr –1 in 2012 (Landschützer et al., 2015 <sup>[[#fn:r299|299]]</sup> ); this can affect the magnitude of the global CO 2 sink (Section 5.2.2). A weakening CO 2 sink during the 1990s (Le Quéré et al., 2007) reversed in the 2000s as part of a decadal cycle (Landschützer et al., 2015 <sup>[[#fn:r300|300]]</sup> ; Munro et al., 2015 <sup>[[#fn:r3+1|3+1]]</sup> ; Williams et al., 2017 <sup>[[#fn:r302|302]]</sup> ) (SM3.2.3; Figure SM3.4), with a weakening again since 2011 (Keppler and Landschutzer, 2019 <sup>[[#fn:r303|303]]</sup> ). While the weakening sink during the 1990s was explained as a response to changes in the circumpolar winds over the Southern Ocean enhancing the outgassing of natural CO 2 , the subsequent changes appear due to a combination of changes in regional winds, temperature and circulation (Landschützer et al., 2015 <sup>[[#fn:r304|304]]</sup> ; Gregor et al., 2017 <sup>[[#fn:r305|305]]</sup> ; Keppler and Landschutzer, 2019 <sup>[[#fn:r306|306]]</sup> ). Data scarcity, especially in winter, remains a challenge (Ritter et al., 2017 <sup>[[#fn:r307|307]]</sup> ; Fay et al., 2018 <sup>[[#fn:r308|308]]</sup> ; Gruber et al., 2019b <sup>[[#fn:r309|309]]</sup> ); recent data from pH-enabled floats highlighted the potential role for winter outgassing south of the Polar Front (Williams et al., 2017 <sup>[[#fn:r310|310]]</sup> ; Gray et al., 2018 <sup>[[#fn:r311|311]]</sup> ). Overall, there is ''medium confidence'' that the Southern Ocean CO 2 sink has experienced significant decadal variations since the 1980s.

Southern Ocean carbon storage is affected by changes in overturning circulation (Cross-Chapter Box 7 in Chapter 3), with the storage of anthropogenic and natural carbon being both variable and out of phase on decadal timescales (DeVries et al., 2017 <sup>[[#fn:r312|312]]</sup> ; Tanhua et al., 2017 <sup>[[#fn:r313|313]]</sup> ) (Table SM3.4). Mode and intermediate waters are strongly involved in changing storage, also showing high sensitivity to shifts in winds (Swart et al., 2014 <sup>[[#fn:r314|314]]</sup> ; Swart et al., 2015a <sup>[[#fn:r315|315]]</sup> ; Tanhua et al., 2017 <sup>[[#fn:r316|316]]</sup> ; Gruber et al., 2019a <sup>[[#fn:r317|317]]</sup> ). Zonal basin differences in the uptake and storage of anthropogenic carbon are not well resolved and there is weak agreement between reanalysis products and Coupled Model Intercomparison Project Phase 5 (CMIP5) models (Swart et al., 2014 <sup>[[#fn:r318|318]]</sup> ). The presence of subduction hotspots suggest that basin-wide studies may be underestimating the importance of mode water subduction as a principal storage mechanism (Langlais et al., 2017 <sup>[[#fn:r319|319]]</sup> ).

Strengthening impacts of Southern Ocean acidification are illustrated by the 3.9 ± 1.3% decrease in derived calcification rates (1998–2014) (Freeman and Lovenduski, 2015 <sup>[[#fn:r320|320]]</sup> ). These have strong regional character, with decreases in the Indian and Pacific sectors (7.5–11.6%) and increases in the Atlantic (14.3 ± 5.1%). There have also been changes in the seasonality of pCO 2 linked to decreasing buffer capacity (McNeil and Sasse, 2016 <sup>[[#fn:r321|321]]</sup> ) (SM3.2.4) or adjustments to primary production (Conrad and Lovenduski, 2015 <sup>[[#fn:r322|322]]</sup> ); seasonal changes are discussed further in Section 5.2.2.

<div id="section-3-2-1-3-ocean-circulation"></div>

<span id="ocean-circulation"></span>