This chapter examines the ocean environments of both polar regions – their hydrography, water masses, and currents. Global sea level changes are also described. Each polar sea is discussed, and then sea ice conditions and polynyas are examined.
7.1 Southern Ocean
The Southern Ocean was defined and its major characteristics described in Section 4.2. Here we look in more detail at its seasonal and geographical variability.
Some 99.7 percent of the area between latitudes 60° and 65° S is ocean. The greatest constriction in the Southern Ocean is at the Drake Passage, where the tip of South America is 1,100 km from the tip of the Antarctic Peninsula (Ostapoff Reference Ostapoff and Mieghem1965). The zonally averaged sea surface temperature (SST) is 0 °C at 61° S, while the equivalent temperature at 61° N is 4 °C higher. There are considerable departures from zonality, however. The Atlantic and Indian Ocean sectors are up to 2–3 °C warmer than the Pacific sector. The diurnal variation of SST in high latitudes is much less than 0.5 °C as a result of the large amount of cloud cover and persistent strong winds.
The surface salinity in the Southern Ocean south of the ocean Polar Front is generally constant around 33.9 psu, but local variations between 31 and 34.5 psu have been observed (Ostapoff Reference Ostapoff and Mieghem1965). The low value is caused by ice melt. The Polar Front (or Antarctic Convergence) generally coincides with a salinity minimum at 200 m depth of approximately 34.0–34.2 psu, while to the south there is a zonal ring of maximum values greater than 34.6 psu.
Tomczak and Godfrey (Reference Tomczak and Godfrey2003) provide a detailed account of Antarctic oceanography. Except near the continent, the waters of the Southern Ocean move eastward, forced by the southern westerlies. This is the only region of the planet where there is essentially unhindered circumglobal flow of ocean water. Maximum velocities occur just north of the Polar Front, and in the Drake Passage the average maximum geostrophic velocity over 100 km distance is approximately 27 cm s−1, with an estimated core velocity of 40–60 cm s−1 (Ostapoff Reference Ostapoff and Mieghem1965). At the bottom, velocities decrease to 5–10 cm s−1.
The Antarctic Circumpolar Current (ACC) flowing continuously around the Antarctic continent is the strongest current in the world’s oceans (Figure 7.1). It has a length of about 20,000 km (Rintoul Reference Rintoul2000; Wyrtki Reference Wyrtki1960) and moves eastward throughout its depth from the surface down. Using three years of in situ velocity data (2006–2009) in the Drake Passage and twenty years of satellite altimetric observations of surface currents, Koenig et al. (Reference Koenig2016) estimated the volume transport at 140 Sv (106 m3 s−1). However, Donahue et al. (Reference Donahue2016) found a larger value of 173 Sv from moorings between 2007 and 2011 (see Section 4.2) This total is primarily baroclinic flow (127.7 Sv), with 45.6 Sv of barotropic origin. A barotropic flow is the same from the ocean surface to the bottom – that is, the flow is uniform with depth. Barotropic flow, by comparison, is in dynamical balance with the sea surface slope. The term “baroclinic” denotes the depth-dependent part of the flow. The baroclinic component of the flow results from the density distribution in the fluid, which varies due to differences in temperature and salinity. Hence, it is a function of the vertical shear. The Drake Passage is open to 2,000 m depth, which includes the Upper Circumpolar Deep Water (UCDW) layers.
Meridional sections across the ACC indicate an equatorward flow of surface water due to the westerly wind stress and poleward transport of subsurface and deep water (Figure 7.2). Mesoscale eddies between the surface Ekman layer and approximately 2,000 m depth transfer most of the ocean heat toward the Antarctic continent. The Upper Circumpolar Deep Water (UCDW) transports about 2 Sv southward, according to Wyrtki (Reference Wyrtki1960). North Atlantic Deep Water (NADW) and Lower Circumpolar Deep Water (LCDW) move southward below the UCDW. The bottom water flows northward (Figure 7.2). Observations show that winds and current almost coincide and that both the belt of maximum westerlies and the current show a shift to the south from about 40–45° S in the central South Atlantic to about 55–60° S in the eastern South Pacific.
The Antarctic Polar Front is located south of the ACC axis and south of the maximum westerlies. According to Wyrtki (Reference Wyrtki1960), it can be either a divergent or a convergent phenomenon, depending on the strength and position of the westerly wind maximum. With weak westerlies, there is divergence to the north and convergence to the south of the main current that intensifies the Polar Front. If the westerlies are strong and displaced southward, there is convergence to the north and divergence to the south of the maximum westerlies, extending up to the ice edge. With the westerly wind belt in a northerly position, the Polar Front is in the range of divergent motion. Easterly winds develop near the continental margin and Antarctic divergence is displaced from the ice edge with poleward motion of surface water.
Kort (Reference Kort and Odishaw1964) notes that Antarctic Bottom Water (AABW) forms in autumn–winter when shelf water begins to cool and ice growth increases its salinity by brine drainage. Fofonoff (Reference 331Fofonoff1956) demonstrated that shelf water sinks down the continental slope when the salinity is at least 34.51 psu. The continental shelf edge around Antarctica has a depth of up to 800 m. The regions where bottom water forms are the southwest and western Weddell Sea, the Ross Sea, the Shackleton shelf glacier, and near-shore areas of Princess Martha Coast (5° E to 20° W). Orsi et al. (Reference Orsi, Johnson and Bullister1999) calculated that the rate of newly formed AABW sinking down the slope around Antarctica is 8–10 Sv.
The Southern Ocean is the only place where there is extensive direct upwelling of deep water to the sea surface. Dynamically, this movement is attributable to the open latitude band (56–63° S) of the Drake Passage. Antarctic Bottom Water (AABW) at 67° S, 30° W has a potential temperature of −0.88 °C and salinity of 34.64 psu (Johnson Reference Johnson2006). The net buoyancy gain shown in Figure 7.2 over much of the Southern Ocean is a result of both heat input and net precipitation.
The permanent thermocline that characterizes most of the world ocean is absent around Antarctica (Tomczak and Godfrey Reference Tomczak and Godfrey2003). Density variations with depth are small, and the pressure gradient force is distributed more evenly through the water column. As a result, currents can extend to great depth.
Available hydrographic observations for the Southern Ocean were compiled by Orsi et al. (Reference Orsi, Whitworth and Nowlin1995). The northern boundary of Subantarctic Surface Water delimits the Subantarctic Front (SAF). The sharp termination of the poleward extent of UCDW characteristics coincides with a frontal feature that separates the ACC from the Weddell Gyre. The traditional recognition of only three fronts is attributed by Sokolow and Rintoul (Reference 336Sokolow and Rintoul2009a, Reference Sokolow and Rintoul2009b) to the emphasis placed on the Drake Passage. These researchers used fifteen years of sea surface height (SSH) data to map circumpolar ACC fronts. In twelve longitudinal sectors of 30°, regions of high gradient were defined as those that exceeded a threshold of 0.25 m per 100 km. Sokolow and Rintoul then found the SSH contours that most efficiently represented the high gradient regions. They showed that the ACC consists of multiple frontal filaments or jets that are aligned along particular streamlines throughout the circumpolar path of the current. The SSH value (approximate streamline) associated with each frontal branch was found to be nearly constant, both in time and around the circumpolar path. The frontal branches inferred by fitting SSH contours to fifteen years of SSH gradient maps agree very well with the front positions inferred from synoptic sections of water mass properties; see Figure 7.3.
Emery and Meincke (Reference Emery and Meincke1986) provided a summary of global water masses. For the upper 500 m around Antarctica, they describe two components:
Below 1,500 m, there are three components:
North Atlantic Deep Water (NADW): 1.5–4.0 °C, 34.8–35.0 psu
Antarctic Bottom Water (AABW): −0.9 to 1.7 °C, 34.64–34.72 psu
Giglio and Johnson (Reference Giglio and Johnson2016) used Argo float data north of 60° S for 2006–2013 to identify the Subantarctic and Polar Fronts in the Antarctic Circumpolar Current; there are two Subantarctic Fronts. Dynamic height contours can be used to identify all three fronts, which correspond to local maxima in vertical shear. The ACC fronts are associated with strong gradients in temperature and salinity. Based on potential temperature criteria, the Polar Front is farthest south, between 75° and 110° W at around 62° S, and around 61° S at 180° longitude. It is farthest north (48° S) at 75° E and at about 50° S from 30° W to 30° E. During 2006–2013, in the upper 2,000 m of the global ocean sampled by Argo floats, the southern hemisphere ocean was the main recipient of heat from global warming. Roemmich et al. (Reference Roemmich2015) have shown that ocean heat gain over the 0–2,000 m layer took place at a rate of 0.4–0.6 W m−2 during 2006–2013; 67–98 percent of this gain occurred in the southern hemisphere extratropical ocean.
Graham et al. (Reference Graham2012) use one hundred year simulations to show that the number and intensity of fronts is set largely by the bottom topography of the Southern Ocean. The number of fronts is reduced in regions where the path of the ACC is constricted or blocked by topography, as in the Drake Passage. Fronts within the ACC are more barotropic and extend down to the ocean floor making them sensitive to the topography.
7.2 Ross Sea
The Ross Sea is a deep embayment in the Southern Ocean between Victoria Land in the west and Marie Byrd Land in the east. Its northern limit is the edge of the continental shelf. It covers about 960,000 km2. The southern part is covered by the Ross Ice Shelf (see Section 6.3) and the sea is covered by sea ice for most of the year. The sea is generally less than 900 m depth; in the west, it is less than 300 m deep over wide areas. The ocean surface circulation is dominated by a wind-driven cyclonic (clockwise) gyre, which is forced partly by the east wind drift along the coastline that turns northward along the coast of Victoria Land. The gyre is accompanied by upwelling of deep water. The bottom water characteristics at 72° S, 165° W are potential temperature of −0.24 °C and salinity of 34.70 psu (Johnson Reference Johnson2006). The slow-moving Circumpolar Deep Water is a relatively warm, salty water mass that flows onto the continental shelf at certain locations in the eastern Ross Sea, influenced by the bottom topography. It has a temperature of 1–2 °C and a salinity between 34.62 and 34.73 psu.
Jacobs et al. (Reference Jacobs, Amos and Bruchhausen1970) report that the westerly current of CDW over the Ross Sea continental slope provides a dynamic barrier to northward thermohaline flow of dense Ross Sea Shelf Water (RSSW). Insufficient brine may be released by the freezing of sea ice to produce the major portion of RSSW. Ice Shelf Water (ISW), identified by a pronounced temperature minimum, has temperatures as low as −2 to −1 °C near its source at the base of the Ross Ice Shelf. Low-salinity Antarctic Bottom Water (AABW) forms during summer over the continental slope in the eastern Ross Sea from a mixture of CDW and ISW. Higher-salinity AABW is produced in summer in the western Ross Sea from a combination of RSSW and CDW.
7.3 Weddell Sea
The Weddell Sea is an embayment in the Antarctic continent that is delimited by the Antarctic Peninsula to the west and the coast of Coats Land to the east. It has an area of about 2.8 million km2. The southern part is covered by the thick Ronne–Filchner Ice Shelf (see Section 6.3). The Antarctic continental shelf widens to 250 km along the Antarctic Peninsula and up to about 480 km along the southern edge of the Weddell Sea. Marking the edge of the continent, the break lies at an unusually great depth of about 500 m. The Weddell Sea is one of the few locations in the global ocean where deep bottom water forms and contributes to the global ocean thermohaline circulation (see Box 7.1). The wind-driven, clockwise, cyclonic Weddell Gyre gives rise to an outflow in the western part of the sea that transports sea ice and icebergs northward (Muench and Gordon Reference Muench and Gordon1995). It has a cold, low-salinity surface layer; at 68° S, 53° W the surface potential temperature is −1.85 °C and the salinity is 34.25 psu (Johnson Reference Johnson2006). This is separated by a thin, weak pycnocline from a thick layer of relatively warm and salty water, referred to as Weddell Deep Water, and a cold bottom layer. Based on data obtained from the western Weddell Sea during the austral winter 1992 US–Russian drifting ice station experiment, Muench and Gordon (Reference Muench and Gordon1995) reported that the northward flow increased more than twice as much from south to north. A significant fraction of the northward transport was contained in a 300–500 m thick bottom layer of cold water.
The thermohaline circulation (THC) is the part of the large-scale ocean circulation that is driven by global density gradients created by fluxes of surface heat and freshwater (Figure 7.4). Density gradients are determined by gradients of water temperature and salt content. The THC is also known as the meridional overturning circulation (MOC) or the global conveyor belt, in recognition that these currents are responsible for the large-scale exchange of water masses in the ocean.
The thermohaline circulation is mainly driven by the formation of deep water masses in the northern North Atlantic and the Southern Ocean near Antarctica as a result of differences in water temperature and salinity. The absence of deep water formation in the Pacific Ocean may reflect a higher freshwater flux, making it more stable than the Atlantic Ocean. Atlantic Deep Water (ADW) flows southward from the Greenland Sea and eventually surfaces in the North Pacific, at which point some of it becomes the Indonesian through-flow. ADW intermingles with Antarctic Intermediate Water (AIW) and flows into the South Atlantic via Drake Passage. The Indonesian through-flow traverses the South Indian Ocean and mixes with returning AIW to form a northward flow to complete the ocean circulation. The entire circulation takes approximately 2,000 years.
A significant phenomenon in the Weddell Sea is the sinking of water mass by the process of cabbeling. Cabbeling involves the formation of denser water by the mixing of two water masses with different temperatures and salinity; warm, salty water from mid-latitudes encounters cold, freshwater from melting sea ice and the effect of wind cooling near Antarctica. Offshore winds drive sea ice away from the coast, thereby leading to continual ice growth in coastal polynyas (see Section 7.9). An important associated process is brine extrusion during ice growth, which increases water salinity.
Launiainen and Vihma (Reference Launiainen, Vihma, Johannessen, Muench and Overland1994) analyzed surface energy fluxes in the Weddell Sea from five drifting buoys during 1990–1992. The buoys drifted from about 73° S in the Weddell Sea northward to about 63° S, and then east–northeastward toward 25° to 0° W. Table 7.1 summarizes some of these researchers’ results. Sensible heat fluxes were largest, 100–300 W m−2, over leads and coastal polynyas that represented 5–7 percent of the area in winter. These fluxes approximately balanced the downward flux over the sea ice. Summer values of sensible heat there were small due to the small difference between air and sea surface temperatures. The mean annual heat loss from the Weddell Sea is estimated to be 20–30 W m−2.
|Surface||Sn||L↑||H||LE||Flux through ice|
|Sea ice: winter||0||20–40||−17||3||10|
|Sea ice: summer||−40||30–50||−5||3||–|
|Ice shelf: winter||0||30–50||16||0||1|
|Ice shelf: summer||−60||50–60||−12||2||1|
|Coastal polynya: winter||−2||100–130||240||80|
|Coastal polynya: summer||−240||60–70||(30)||(20)|
Winter is June to end of August. Summer is December to end of February.
7.4 Arctic Ocean
The Arctic Ocean has limited contact with the Atlantic and Pacific oceans due to the presence of narrow passageways and shallow sills in this body of water. It is, therefore, designated as a Mediterranean sea (Tomczak and Godfrey Reference Tomczak and Godfrey2003). The circulation of the Arctic Ocean, unlike that of the Southern Ocean, is driven by thermohaline forcing.
The seas that make up the Arctic Ocean are the Barents, Kara, Laptev, East Siberian, Chukchi, Beaufort, and Lincoln (Figure 7.5). The Barents Sea is the largest, with an area of 1.405 million km2. Important seas adjacent to the Arctic Ocean are the Bering Sea and the Greenland Sea, as well as Baffin Bay. The Arctic Ocean has continental shelves covering 2.5 million km2. About 40 percent of this area occurs in the interior shelves of the Kara Sea, Laptev Sea, East Siberian Sea, and Beaufort Sea (Williams and Carmack Reference Williams and Carmack2015), which are distinguished from the inflow and outflow shelves by their massive amounts of river runoff. In the mid-shelf region, wind and ice motion-induced surface stresses dominate mixing and circulation, resulting in high variability. Along the northern boundary, water is forced by upwelling- and downwelling-favorable surface stresses, which drive shelf-basin exchanges with cyclonic boundary currents of Atlantic and Pacific origin over the upper slope. Shelf-basin exchange is further modified by shelf-break morphology.
The Arctic Ocean is divided by the Lomonosov Ridge into two deep basins, the Canada Basin and the Eurasian Basin, which reach depths of 5449 m (Figure 7.6). The ridge extends across the Arctic from north of Greenland to the New Siberian Islands and in places is less than 400 m deep. A 1,700-km-wide opening exists to the North Atlantic along a large sill that runs from Greenland to Iceland, the Faroe Islands, and the Scotia bathymetric features. Approximate sill depths are 600 m in Denmark Strait (between Greenland and Iceland), 400 m between Iceland and the Faroe Islands, and 800 m between the Faroe Islands and Scotland.
Temperature and salinity data for the Arctic basin were assembled for a joint US–Russian atlas of Arctic oceanography for both summer and winter (Tanis and Timokhov Reference Tanis and Timokhov1997, Reference 337Tanis and Timokhov1998). Mean values are shown versus depth for the Canadian, Amundsen, and Nansen basins in Table 7.2.
|Depth (m)||(a) T||S||(b) T||S||(c) T||S|
Atlantic water penetrates into the Nansen basin at 300 m, where the average temperature is 1.7 °C and salinity shows a sharp increase. The salinity values at and below 2,000 m are remarkably uniform. The lowest water temperatures of −1.8 °C or less, and the highest salinities of approximately 34.9 psu, are found in winter in the northern Barents and Kara seas, the western Laptev Sea, and adjoining areas of the Nansen basin (see Table 7.2).
The Environmental Working Group (EWG) data set has enabled the study of long-term variability of the Arctic Ocean (Polyakov et al. Reference Polyakov2004). Swift et al. (Reference Swift2005) used a special version of the EWG data, which allowed them to examine the variability in the upper layer, the Atlantic layer, and the Pacific waters (Rudels Reference Rudels2015). Similar strong and sudden changes as in the 1990s could be detected in the 1960s, which showed a strong increase in Atlantic water temperatures, and in the 1970s, which featured a salinity increase in the upper layer. The weakening and disappearance of the Pacific water in the central Arctic Ocean was found to occur in the mid-1980s, before the inflow of warm Atlantic water and the redistribution of the low-salinity shelf water into the Canadian Basin arose in the 1990s.
A cross-section of potential temperature and salinity in the Arctic Ocean is shown in Figure 7.7; the figure also highlights the bathymetric features. Potential temperature (θ) is used in oceanography to compare the temperature of water parcels at different pressure levels in the ocean where compressibility of the water plays a small role – namely, compression (expansion) causes a rise (fall) of temperature. Potential temperature is defined as the temperature of a water parcel that is moved adiabatically to a different pressure level. The adiabatic lapse rate in the ocean is approximately 0.1–0.2 °C km−1 (in contrast to the 9.8 °C km−1 rate for unsaturated air). In the ocean, θ is defined in reference to the sea surface, so it is always slightly lower than the actual temperature. It is a conservative property. Figure 7.7 illustrates the three main layers in the Arctic: (1) Polar Surface Water down to about 200 m, (2) intermediate water from about 200 to 800 m, and (3) deep and bottom waters below 800 m.
The surface circulation in the Arctic Ocean is driven mainly by the clockwise gyre that is a response to the winter ridge of high pressure over the western Arctic and the Beaufort Sea high in spring (Serreze and Barrett Reference Serreze and Barrett2011). Ice circulates slowly around the gyre. On the Eurasian side, the Transpolar Drift Stream takes water and sea ice from the Eurasian coast to Fram Strait, where it forms the East Greenland Current. The Atlantic water that enters the Arctic via the Norwegian and Barents seas sinks beneath the surface polar water and flows counterclockwise around the basin below 250–300 m depth toward the Canadian Arctic archipelago. It has much higher temperature (0–0.5 °C) and salinity (34.9 psu) than the overlying Arctic Water (see Figure 4.10).
The Arctic Ocean has low surface salinity, which is approximately 27–30 psu over the Eurasian continental shelf and in the Beaufort Sea, for example. This is primarily caused by the immense quantity of river discharge in Eurasia (the Ob, Lena, and Yenisei rivers) and western North America (the Mackenzie River). The runoff during the October–September water year amounts to 404 km3 yr−1 for the Ob, 603 for the Yenisei, 525 for the Lena, and 333 for the Mackenzie (Serreze et al. Reference Serreze2003). Salinity values for water coming off the Eurasian river mouths drop to 20 psu or less over the coastal shelves. This decrease is important because in polar oceans density is primarily determined by salinity rather than by temperature. The low salinity of the Arctic is also partly a result of inflow of relatively freshwater from the North Pacific via Bering Strait, with this flow being equivalent to almost half of the river runoff. Halocline stratification is much stronger in the Amerasian Basin than in the Eurasian Basin, owing to additional inputs of low-salinity water from the Pacific via Bering Strait, and surface flow convergence under the atmospheric Beaufort High (Carmack et al. Reference Carmack2016). Pacific water enters the Arctic in the depth range 60–220 m.
A major feature of the Arctic Ocean since the 1990s has been a general warming trend (Seidov et al. Reference Seidov2015). Based on the World Ocean Database (WOD) that has been assembled by the National Oceanic and Atmospheric Administration (NOAA), differences of temperature between 2005–2010 and the coldest decade of 1975–1984 show that upper-ocean warming has occurred nearly everywhere, with the amplitudes being consistently high in the Greenland–Iceland–Norwegian Sea (GINS), Barents Sea, Eurasian Arctic, and western Arctic.
Levitus et al. (Reference Levitus2000) found that heat content of the surface 0–3,000 m layer in the world ocean increased by approximately 2 × 1023 J between the mid-1950s and mid-1990s, which corresponds to a warming of 0.06 °C. Seidov et al. (Reference Seidov2015) have shown that the heat content of the layer between 50 and 300 m depth in the GINS and the Arctic Ocean turned sharply positive, relative to the base climate of 1955–2006, after 2000. The temperature of the top 100–150 m layer in the Barents Sea has increased by approximately 4 °C since the late 1970s.
Rudels (Reference Rudels2015) reports that north of the Laptev Sea, the temperature of the Atlantic water observed on the Nansen and Amundsen Basin Observational System (NABOS) cruise in 2002 was reduced considerably from the high values observed by Polarstern in the 1990s. However, the temperature measured at the NABOS moorings on the Laptev Sea slope began to rise in 2004, when a sudden increase in both the temperature and the thickness of the Atlantic layer was observed. The temperature has remained high ever since (Dmitrenko et al. Reference Dmitrenko2008; Polyakov et al. Reference 335Polyakov2005). Wells et al. (Reference Wells, Couldrey and Ivchenko2013) reported, based on Argos data, that the heat storage in the northern North Atlantic increased during 1999–2010, especially between 60° and 70° N. Cold anomalies were observed during the same interval between 20° and 50° N.
7.5 Arctic Seas
The distribution of the Arctic seas around the basin is shown in Figure 7.5. These seas are discussed in turn in this section.
7.5.1 Barents Sea
The Barents Sea between Svalbard and Novaya Zemlya has an area of 1,420,000 km2 and is rather shallow, with an average depth of about 225 m. It is linked to both the Norwegian Sea/North Atlantic and the Arctic Ocean, and extends to about 80° N. The northeastward Atlantic Current divides into three branches: the West Spitsbergen Current in eastern Fram Strait, flow into the Barents Sea, and the Norwegian Coastal Current (Pfirman et al. Reference Pfirman, Bauch, Gammelsrod, Johannessen, Muench and Overland1994). The warm, salty waters of the North Atlantic Current form a Polar Front, along with cold water from the north, that extends generally southeastward from Svalbard toward Novaya Zemlya (Loeng Reference Loeng1991). The southern part of the sea remains ice-free year-round to about latitude 75° N. The loss of heat to the atmosphere, while Atlantic water traverses the Barents Sea, greatly modifies this water mass prior to its entry into the Arctic Ocean proper. When passing through the Barents Sea, the Norwegian Atlantic Current is strongly modified by cooling, mixing, and freezing during winter, and all the Atlantic Water (AW) entering in the west is modified and leaves the shelf toward the Arctic Ocean mostly with temperatures below 0 °C (Ingvaldsen et al. Reference Ingvaldsen, Asplin and Loeng2004).
Loeng et al. (Reference Loeng, Ozhigin and Adlandsvick1997) and Ingvaldsen et al. (Reference Ingvaldsen, Asplin and Loeng2004) have determined the transports of water through the Barents Sea. There is an average inflow and outflow transport of approximately 4.6 Sv in winter and 3.1 Sv in summer, of which the through-flow of Atlantic water contributes 1.7 Sv in winter and 1.3 Sv in summer across a section from 71.5° to 73.5° N (Ingvaldsen et al. Reference Ingvaldsen, Asplin and Loeng2004). A spring minimum, which is sometimes an outflow, is associated with northerly winds. Based on data from moorings during October 1991 to September 1992, at 77.3° N, 62.9° E to 78.8° N, 58.6° E, Schauer et al. (Reference Schauer2002) have shown that the majority of the Atlantic inflow entering from the Norwegian Sea leaves the northeastern Barents Sea as cold, dense bottom water as a result of cooling and freezing. The flow toward the Kara Sea was between 0.6 and 2.6 Sv during the period covered by their analysis.
There is an outflow of cold Arctic Water just south of Bear Island – the Bear Island Current – but its magnitude is unknown. The Atlantic domain shifts northward in winter. This phenomenon is connected to coastal downwelling off Norway, which moves the AW away from the coast. Arctic Water, northeast of the polar front, appears to originate from convection due to sea ice formation, but some may be advected from the northern Kara Sea and the Arctic Ocean (Pfirman et al. Reference Pfirman, Bauch, Gammelsrod, Johannessen, Muench and Overland1994). It comprises three main components: the westward-flowing Persey Current near 78° N, the East Spitsbergen Current flowing southwestward out of the Arctic Ocean between 45° and 30° E, and the strong Hopen–Bjørnaya Current flowing southward along the eastern flank of the Spitsbergen Banks to about 74.5° N.
From autumn to early summer, the region north of the Polar Front is ice covered. The ice melts as summer progresses and the margin shifts northward. In mid- to late summer, this melt forms a 20-m-thick surface layer of relatively warm (due to radiative heating) freshwater above the cold AW. Here, warm, saline Atlantic-derived water is found at depths less than 75 m.
7.5.2 Kara Sea
The Kara Sea is located north of Russia between Novaya Zemlya and Franz Josef Land in the west and Severnaya Zemlya in the east. It has an approximate area of 880,000 km2 and an average depth of only 110 m. The central regions have a large number of island groups. About 82 percent of the Kara Sea occupies part of the Siberian shelf; accordingly, about 40 percent of it is less than 50 m deep. The water is very cold (−1.4 °C) and typically remains frozen for nine months of the year.
The Kara Sea receives large amounts of discharge from the Ob (approximately 400 km3) and Yenisei (630 km3) rivers, so the salinity in summer off the mouths of the Ob and Yenisei is only 10–12 psu. Over the rest of the sea, the salinity is 25 psu in winter and 22 psu in summer. The northern Kara Sea is influenced by deep Atlantic water, which penetrates the Kara Sea from the Arctic Basin via the deep St. Anna and Voronin troughs. According to Schauer et al. (Reference Schauer2002), in 1996, warm Atlantic water, with temperatures up to 3 °C, was located in the western St. Anna Trough at 200 m depth.
7.5.3 Laptev Sea
The Laptev Sea off northern Siberia is bounded in the west by the Taimyr Peninsula and Severnaya Zemlya and in the east by the New Siberian Islands and Kotelny Island. The Siberian mainland is dissected by several large gulfs and bays. The sea has an area of 660,000 km2 and a mean depth of about 50 m. The southern and southeastern areas, comprising 45 percent of the total area, have water depths ranging from 10 to 50 m. A dramatic incline, beginning at 100 m depth and ending at the 3,000 m level, divides the sea into the northern and southern parts along the parallel of the Vil’kitsky Strait. Assuming the 200-m isobath to be the shelf boundary, the shelf zones make up 72 percent of the area of the Laptev Sea (Timokhov Reference Timokhov1994).
The runoff to the Laptev Sea from five major rivers – Khatanga, Anabar, Lena, Olenek, and Yana – is about 767 km3. The summer salinity off the mouth of the Lena River is only 5–10 psu, but in the northern parts it rises to 28 psu. Except in August–September, this sea remains largely ice covered. In winter, water temperature varies from −1.4 °C in the eastern sea, up to −0.8 °C in the northwestern sector. In summer, the southwestern upper 15-m layer is warmed to a temperature of 5–7 °C and river runoff near the coast results in surface temperatures of 8–10 °C (Timokhov Reference Timokhov1994). Temperatures increase to 1° in the southeastern part, but remain about −1° in the northern areas. Surface water motion is generally cyclonic, with eastward flow along the coasts.
The Laptev Sea is one of the core areas for ice production in the Arctic Ocean, with a distinct connection to Transpolar Drift characteristics. It transports large quantities of sediment that are incorporated into the ice due to the processes that form frazil ice (a suspension of ice crystals). Ice production in the Laptev Sea was mapped with the Moderate-Resolution Imaging Spectroradiometer (MODIS) by Preusser et al. (Reference Preusser2016), which is an especially valuable source of data in this region, given the narrow and elongated flow leads close to the fast-ice edge. These researchers’ results showed that polynyas in the Laptev Sea contribute at least 7 percent to the total potential sea-ice production in Arctic polynyas.
The ice drift and export in the Laptev Sea has been studied by Alexandrov et al. (Reference Alexandrov2000). According to these researchers, during an “average year,” sea ice is exported from the Laptev Sea through its northern and eastern boundaries, with maximum and minimum export occurring in February and August, respectively. The winter ice outflow from the Laptev Sea has been shown to vary between 251,000 km2 (1984–1985) and 732,000 km2 (1988–1989), with the average being 483,000 km2. Sea ice is exported into the East Siberian Sea mostly in summer, the volume discharged having a mean value of 69,000 km2. Out of the seventeen summers investigated by Alexandrov et al., twelve of them were characterized by sea ice import from the Arctic Ocean into the Laptev Sea through its northern boundary.
7.5.4 East Siberian Sea
The East Siberian Sea lies between the New Siberian Islands in the west and Wrangel Island in the east. Three straits connect it to the Laptev Sea. The East Siberian Sea has an area of 910,000 km2 and a mean depth of about 50 m. It is only 10–20 m deep in the west and central parts and 30–40 m deep in the east. Assuming the 200-m isobath to be the shelf boundary, the shelf zones make up 96 percent of the area of the Siberian Sea (Timokhov Reference Timokhov1994). The 50-m isobath passes almost parallel to the mainland shore at an average distance of 550–650 km from it.
Near the western boundary, there is the large New Siberian Island archipelago, which is divided into three main groups: the Lyakhovsky islands, the Anjou islands, and the De Long islands. For most of the year, the sea is ice covered. The water temperature at the surface in the winter decreases from the southwest to northeast and is near freezing. The salinity in winter has a tendency to increase from the southwest, rising from 17–18 psu in the western part to 32–33 psu in the northeastern part. In summer, the water temperature at the surface decreases northward, from 5 to 7° C in the coastal zone to −1 to −1.5 °C in the northern part.
Based on cruise data from September 2000, Semiletov et al. (Reference Semiletov2005) identified two distinct water masses in the southern part of the sea. West of about 160° E, there is freshwater flux from the Lena River; temperatures average 2.6 °C and salinity 22 psu. In the eastern part, there is Pacific inflow, with average temperatures of 0.6 °C and 30 psu salinity.
7.5.5 Chukchi Sea
The Chukchi Sea lies between Wrangel Island and Point Barrow, Alaska. Its southern boundary is the Arctic Circle. It has an area of close to 600,000 km2 and a mean depth of 70 m. Slightly more than 55 percent of this sea is less than 50 m deep. Pacific water flows into the Chukchi via Bering Strait. Northward transport through Bering Strait is seasonal, averaging 1.0 Sv during April–September and 0.6 Sv during October–March (Woodgate et al. Reference Woodgate, Aagaard and Weingartner2005a, Reference 338Woodgate2005b). The mean inflow is opposed to the prevailing winds and is forced primarily by a slope in sea level from the North Pacific to the Arctic Ocean. The inflow branches into pathways following Herald Canyon, Central Channel (71° N, 175° E, between Herald and Hannah shoals), and the Alaskan coast (Spall Reference Spall2007). The northward flow is at its maximum in summer. More than half of the transport exits the Chukchi Sea via Barrow Canyon. The seasonal cycle of salinity in the southern and central Chukchi Sea is dominated by advection through Bering Strait, while local atmospheric forcing and brine rejection are more important north of Herald and Hanna Shoals and in Barrow Canyon (off Point Barrow).
In late summer, the Chukchi Sea exhibits a two-layer water column with well-mixed surface and bottom layers separated by a strong pycnocline (Ladd et al. Reference Ladd2016). During winter, a combination of surface cooling, wind mixing from storms, and brine rejection from ice formation vertically mix the water column. Atlantic Water upwelling events, attributed to easterly winds, occur frequently in Barrow Canyon. During some upwelling events, AW with potential temperature greater than −1 °C and salinity greater than 33.6 psu has been observed to upwell from deeper than 200 m in the Arctic Basin onto the Chukchi Shelf via Barrow Canyon.
A winter polynya that occurs along the Alaskan coast between Cape Lisburne and Point Barrow is the largest in the western Arctic (see Section 7.9). Observations from 1990–1991 and 1991–1992 show the temperature in the central Chukchi Sea is at the freezing mark from late in the year through early summer. Winter water in the Chukchi has a salinity of about 32.5 psu, and the salinity of summer water is 34 psu or greater (Spall Reference Spall2007). Advection through Bering Strait is important for the large-scale timing of ice melt in the Chukchi Sea.
7.5.6 Beaufort Sea
The Beaufort Sea is located between a line north of Point Barrow in the west and Banks Island and the southwestern edge of Prince Patrick Island in the east. It has an area of about 475,000 km2 and a mean depth of around 1,000 m. The continental shelf is narrow near Point Barrow but is more than 140 km wide off the mouth of the Mackenzie River. The surface circulation over the deeper portions of the Beaufort Sea is dominated by the Beaufort gyre, which moves clockwise over the Canada Basin. The pack ice drifts westward in the Beaufort gyre but the subsurface flow along the shelf slope is eastward. Pickart (Reference Pickart2004) has shown that the eastward current is concentrated in an approximately 20-km-wide jet along the shelf break. This jet has three distinct seasonal configurations: (1) in late spring to late summer, cold, winter-transformed Bering Sea water is advected eastward; (2) from about mid-summer to early autumn, a surface-intensified current advects predominantly Bering summer water (for brief periods in August–September, the local winds come from the west); and (3) from mid-autumn to mid-spring, under easterly winds, the jet transports upwelled Atlantic water. The upper 100 m of this jet has a temperature of about −1.4 °C in summer and −1.8 °C in winter. It overlies Pacific water that enters via Bering Strait. Below is a layer of warm Atlantic water at 0–1 °C that has circulated around the Arctic Ocean. The bottom water has a temperature of −0.4 to −0.8 °C.
The inflow into the Beaufort Sea is about 420 km3. Data on the Mackenzie River inflow are provided by Macdonald et al. (Reference Macdonald1989). About 70 percent (by volume) of the shelf water in autumn 1986 was due to river runoff and 30 percent was due to ice melt. Hence the shelf here is largely estuarine in character. Freshwater is removed by winter ice growth early after freeze-up. Then, as the ice thickens, removal of water from the shelf by flushing takes over and lasts approximately 150 days.
7.5.7 Lincoln Sea
The Lincoln Sea stretches from northern Ellesmere Island in the west to Cape Morris Jessup, northern Greenland, in the east. The northern limit is defined as the great circle between those two headlands. It has an area of 64,000 km2 and a mean depth of approximately 250 m. The sea is ice covered throughout the year, with the thickest sea ice found in the Arctic Ocean, up to 15 m thick in places. Haas et al. (Reference Haas, Hendricks and Doble2006) measured multiyear ice thicknesses of 3.9–4.2 m using airborne electromagnetic induction (EMI) measurements. Although some of the ice is exported via Nares Strait, the majority is transported east of Greenland via Fram Strait.
Water depths in the Lincoln Sea range from 100 to 300 m. The water has three distinct properties. First, in the inner part of the Lincoln Sea shelf, the temperature and salinity increase from the surface, where they are about −1.5 °C and 32.4 psu, to about 400 m, and then remain constant to the seafloor. Second, the water over the outer part of the shelf, including the slope, has attributes similar to those in the Canadian Basin, which are not unlike those of waters from the Pacific. Third, the waters north of the slope have characteristics matching those of the Eurasian Basin.
Based on year-round current measurements between 1989 and 1994, Newton and Sotirin (Reference Newton and Sotirin1997) showed that there is an eastward undercurrent, confined to the continental slope, with a width of about 50 km and speeds of 5–9 cm s−1. Temperature and salinity characteristics of this undercurrent are similar to those of Canadian Basin waters, suggesting the existence of a boundary current system that is continuous along the continental slope north of Alaska and the Canadian Arctic archipelago.
7.5.8 Greenland Sea
The Greenland Sea extends east of Greenland to the Svalbard archipelago, north to Fram Strait and the Arctic Ocean, and southeast and south to the Norwegian Sea, Iceland, and Denmark Strait. It has an area of 1.2 million km2. The average depth is 1,450 m, with the maximum depth being 4,850 m.
The average surface water temperature for the Greenland Sea in winter is −1 °С or less in the north and 1–2 °C in the south; the corresponding summer temperatures are about 0 and 6 °C, respectively. The surface water salinity is 33.0–34.0 psu in the east and less than 32.0 psu in the western parts. Waters of the north-flowing North Atlantic Current (NAC) sink in the Arctic Ocean. Some return south in the cold East Greenland Current (EGC), an important part of the Atlantic conveyor belt that flows along the east coast of Greenland and transports Arctic pack ice southward from October to August. In the eastern part of the Greenland Sea is the warm Spitsbergen Current, an offshoot of the NAC. These two currents form a counterclockwise gyre. The transport of Atlantic water, which has a temperature warmer than 1 °C, accounts for about half of the northward flow (Rudels Reference Rudels2015; Schauer et al. Reference Schauer, Dickson, Meincke and Rhines2008). The northward flow in the two West Spitsbergen Current branches is estimated to be 1.8 Sv, with 1.3 Sv of Atlantic water warmer than 2 °C in the eastern branch, and 4.9 Sv, with 1.7 Sv of Atlantic water in the off-shore branch, giving a total northward flow of 6.6 Sv (Beszczynska-Möller et al. Reference Beszczynska-Möller2011). The southward flow in the East Greenland Current is larger, 8.6 Sv, giving a net outflow of 2.0 Sv. The EGC flows from Fram Strait (79° N) to Cape Farewell, transporting large amounts of Arctic ice (see Section 7.3). From a mooring at 75° N in 1994–1995, Woodgate et al. (Reference Woodgate, Fahrbach and Rohardt1999) found an annual mean transport of 21 Sv (across 14–9° W), with the transport varying from 11 Sv in summer to 37 Sv in winter. The flow comprises a seasonally varying wind-driven component (two thirds) and a more consistent thermohaline component (one third).
Offshore from the East Greenland Current, south of 64° N, is relatively saline (34.9–35.0 psu) and warm (4–6 °C) water in the Irminger Current. This branches from the North Atlantic Current at about 26° W and is located above the western slope of the oceanic Reykjanes Ridge (63.5° N, 23° W), southwest of Iceland (Gyory et al. n.d.). Transport estimates for the Irminger Current are in the range of 8–11 Sv (Tomczak and Godfrey Reference Tomczak and Godfrey2003, 244).
The Greenland Sea, as a result of strong surface cooling and wind stress, undergoes a large amount of mixing. The surface mixed layer has a depth of 40–50 m (Blindheim and Østerhus Reference 330Blindheim, Østerhus and Drange2005). Convective events in the central Greenland Sea in the 1990s extended down to between 1,200 and 2,000 m.
In winter, a large area north of Iceland, between Greenland and Jan Mayen, known as the West Ice, is covered by continuous ice. An ice area, the Odden, covering as much as 330,000 km² in most years, extended eastward from the main East Greenland ice edge in the vicinity of 72–74° N during the winters of 1966–1972 (the Great Salinity Anomaly), and in 1982, 1986, 1989, 1997, and 1998. However, this feature has rarely developed since 2000 (Rogers and Hung Reference Rogers and Hung2008). Its formation is associated with the presence of very cold polar surface water in the Jan Mayen Current, high pressure in high latitudes of the North Atlantic, a negative NAO, and anomalous westerly winds. Air temperature and downward longwave flux anomalies in the preceding autumn are unusually low in advance of a winter Odden ice cover, while heat fluxes are weakly positive.
7.5.9 Canadian Arctic Archipelago
The Canadian Arctic archipelago (CAA) consists of three major straits running from east to west – Nares Strait, Jones Sound, and Lancaster Sound – which all open to Baffin Bay. These straits provide a small net outflow from the Arctic Ocean. The shallowest sill is 125 m in Barrow Strait between Cornwallis Island and Somerset Island.
Steele et al. (Reference Steele1996) computed the geostrophic transports between different parts of the Arctic Ocean, and estimated that the outflow through the CAA was 0.56 Sv. For Nares Strait, a mooring array indicated a southward flow of 0.9 Sv (Münchow et al. Reference Münchow, Melling and Falkner2006), but this value has subsequently been lowered. Rudels (Reference Rudels2015) suggests that the total transport southward through the passages of the CAA is between 1.4 and 1.8 Sv.
Pack ice is prevalent for much of the year in the CAA, especially in the north and west. The channels in the southern and eastern parts usually clear by late summer. In 2016, the Northwest Passage became totally ice free during that season.
7.5.10 Adjacent Seas of the North Atlantic
Baffin Bay and Davis Strait
The passages connecting the Arctic Ocean to the North Atlantic exit into Baffin Bay, Davis Strait, and the Labrador Sea. Baffin Bay lies between Greenland to the east, Baffin Island to the west, and Devon and Ellesmere islands to the northwest. Smith Sound in the northern part is occupied by the North Water polynya (see Section 7.9). Baffin Bay has an area of almost 890,000 km2, an average depth of 760 m, and a maximum depth of 2,136 m. The West Greenland Current flows northward from Cape Farewell and recirculates to the west off northwest Greenland. This flow, together with Arctic water flowing south through Nares Strait and Smith Sound, joined by flow out of Lancaster Sound, drives the southward-flowing Baffin Current. The Baffin Current carries sea ice and Greenland icebergs southward into the Labrador Current and to the Grand Banks off Newfoundland. Davis Strait has a sill depth of only 350–550 m.
The Labrador Sea is located east of the coast of Labrador, south of 60° N to a line between Newfoundland and Cape Farewell. It has an area of 840,000 km2, an average depth of approximately 1,900 m, and a maximum depth of 4,300 m. The surface water temperature varies between −1 °C in winter, when two thirds of the sea is ice covered, and 5–6 °C in summer.
The water formed in the central Labrador Sea produces Labrador Sea Water, which is lighter than the North Atlantic Deep Water below it. The densest NADW is formed by mixing of water from the East Greenland Current in the Denmark Strait (between Iceland and Greenland). The Labrador Sea Water circulates around the subpolar gyre many times and mixes with other polar water masses on its journey. NADW flows at depth from the Labrador Sea to Antarctica (Tomczak and Godfrey Reference Tomczak and Godfrey2003, 257). It forms by deep convection events in about six out of ten winters, giving rise to water with temperatures of 3.0–3.6 °C and salinities of 34.86–34.96 psu.
Hudson Bay is connected to the North Atlantic by Hudson Strait and to the Arctic Ocean by Foxe Basin. It has an area of 1.2 million km2, an average depth of only 100 m, and a maximum depth of 270 m. It is ice covered from mid-December to early June, but surface water temperatures reach 8–9 °C in summer in the western part of the bay. Hudson Bay functions essentially as an estuary. There is a slow cyclonic gyre in the water. River runoff is large – approximately 700 km3 – so the salinity is low; this raises the freezing point and facilitates ice growth.
7.5.11 Marginal Seas of the North Pacific
The Bering Sea is bounded by Bering Strait on the north, Alaska on the east, the Russian Far East on the west, and the Alaska Peninsula and the Aleutian Islands to the south. It has an area of 2.3 million km2. The sea is basically divided into a shallow (less than 200 m) Siberian–Alaskan shelf area in the east and north and a deep (3,500–3,800 m) basin in the south and west. There is a cyclonic gyre in the Bering Sea basin, with the south-flowing Kamchatka Current in the west and the north-flowing Bering Slope Current in the east (Stabeno et al. Reference Stabeno, Schumacher, Otahni, Loughlin and Ohtani1999). The Alaska Stream enters the Bering Sea from the North Pacific through gaps in the Aleutian Island chain. Flow over the east Bering Sea shelf is generally first to the northwest, and then northward toward Bering Strait.
Sea ice begins to form over the shelves and is advected southward from Bering Strait in November. Ice forms in polynyas (see Section 7.9) on the leeward side of islands and coasts. It reaches its maximum extent in mid-March, covering about one third to one half of the Bering Sea, and begins to retreat in April.
Sea of Okhotsk
The Sea of Okhotsk is a marginal sea of the northwest Pacific Ocean. It is located between the Kamchatka Peninsula on the east, the Kuril Islands on the southeast, Hokkaido to the south, the island of Sakhalin along the west, and the eastern Siberian coast along the west and north. It has an area of almost 1.6 million km2, an average depth of 860 m, and a maximum depth of 3,370 m. There is a broad, shallow shelf (less than 200 m) in the north; the deep Kuril Basin is found in the south. A cyclonic gyre is present in the Sea of Okhotsk, with a boundary current flowing southward along the coast of Sakhalin Island.
The Sea of Okhotsk receives a large amount of runoff from the Amur River, so it has low surface salinity, which in turn facilitates freezing. Hence there is sea ice cover from October–November to June, or locally July. Coastal polynyas generate dense water on the shelves, which is transported southward in the Sakhalin Current into the Kuril Basin (Gladyshev et al. Reference Gladyshev2001). The sea surface temperature reaches 8–12 °C in summer, and the salinity drops to 32.5 psu during the same season. The southwestern part is warmed by waters from the Sea of Japan and the eastern part by Pacific Ocean inflow.
7.6 Global Sea Level
As discussed in Chapter 2, global sea level has varied by more than 120 m over the last million years. Ice losses from glaciers and ice sheets are making increasing contributions to global sea level rise. However, to put those contributions in context, we must briefly review the components that are involved.
The first is thermal expansion due to ocean warming. This warming involves a density decrease and, therefore, a volume increase (steric sea level rise). The majority of this warming and expansion currently takes place in the upper 700 m. This contribution has increased to 1.1 mm yr−1 for 1993–2010 (Church and Clark Reference Church, Clark and Stocker2013). The large heat capacity of the ocean means that there is considerable delay before the full effects of surface warming are felt throughout the ocean depth. As a result, the ocean will not achieve equilibrium and global average sea level will continue to rise for centuries after atmospheric greenhouse gas (GHG) concentrations have stabilized.
A second major contribution to global sea level is the mass loss of land ice. There were increasing contributions for 1993–2010 from Greenland (0.33 mm yr−1) and Antarctica (0.27 mm yr−1). Glacier contributions, including those from independent ice bodies around Greenland, increased from 0.69 mm yr−1 for 1901–1990 to 0.86 mm yr−1 for 1993–2008. Terrestrial storage in reservoirs and dams was 0.38 mm yr−1 for 1993–2010. The total global average land ice storage of 2.8 mm yr−1 compares with an altimetric observation of 3.2 mm yr−1.
It has proved difficult to determine sea level trends for the Arctic due to the paucity of tide gauges and the limited extent of open water available for satellite altimetry. Recent analysis of both data sources by Svendsen et al. (Reference Svendsen, Andersen and Nielsen2016), however, has provided a reconstruction that shows that the Arctic mean sea level trend between 68° and 82° N was around 1.5 mm ± 0.3 mm yr−1 for the period 1950–2010, in good agreement with the global mean value of 1.8 ± 0.3 mm yr−1 over the same period (Church and White Reference Church and White2011).
7.7 Sea Ice
Sea ice cover is a key element of polar regions that plays a critical role in global and regional climate and oceanic processes. There are three major types of sea ice: first-year ice (FYI), multiyear ice (MYI) and (land)fast ice, each of which has different physical characteristics, as described in this section. A guide to sea ice information services around the world has recently been issues by the World Meteorological Organiztion (2017) that includes descriptions of ice types.
Ice begins to form in the ocean when the surface cools to about −1.8 °C for average ocean salinity (34.5 psu). Ice floats because it has a density of about 917 kg m−3 at 0 °C compared with a density of 1,000 kg m−3 for liquid water. Freshwater has a maximum density at 3.98 °C, but for every 5 psu increase in salinity, the freezing point decreases by 0.28 °C. The temperature of maximum density disappears when the salinity exceeds 24.7 psu. Cooling makes the surface water denser, setting up convection. However, the whole water column does not have to cool to freezing before ice can form, only the upper layer above the level of density maximum – the pycnocline. In the Arctic, the pycnocline is found at 50–150 m depth. Sea ice has two phases: salt-free ice and liquid brine (Ackley Reference Ackley1996). At a growing ice interface, most of the salts are rejected. Brine, gas, and solid salts are usually trapped at sub-grain boundaries within a lattice of essentially pure ice (Timco and Weeks Reference Timco and Weeks2010). First-year sea ice has a typical salinity in the range of 4–6 psu.
Initially, a suspension of tiny ice crystals, called frazil or grease ice, forms in the surface water. In calm conditions, the frazil crystals freeze together to form a continuous thin sheet of transparent ice, called nilas. Water molecules freeze onto the bottom of the ice in a process known as congelation, and the nilas thickens, turning first gray and then white. Congelation ice has a columnar structure. The ice thickens by the extension of ice platelets below; salt is rejected and descends in salty plumes. Brine inclusions are trapped between the platelets as they thicken.
Waves can maintain a suspension of frazil crystals and lead to the growth of small cakes of slush. These grow by accretion and eventually form pancake ice. At the sea ice margin, the pancakes are only a few centimeters in diameter, but in the interior of the pack ice they can be 3–5 m in diameter and 50–70 cm thick. These coalesce into floes and ultimately form a sheet, although rafting may occur locally, increasing the thickness by two to three times. Young ice is more than 30 cm thick, and first-year ice may reach 1.5–2 m in thickness.
Thickness estimates in the Arctic from the Cryosat 2 altimeter and for thin ice from the L-band (1.4 GHz) sensor on the Soil Moisture and Ocean Salinity (SMOS) satellite show mean thickness increases, respectively, from 1.46 and 0.45 m in November 2015 to 1.90 and 0.47 m in April 2016. SMOS thickness peaked at 0.58 m in December 2015 (Ricker et al. Reference Ricker2017).
7.7.1 Sea Ice Edge Location
The sea ice edge is commonly defined by the 15 percent ice concentration line. A transition zone 100–200 km wide – the marginal ice zone –lies between the ice edge and the boundary of ice having a concentration coverage of more than 80 percent. This zone is typically wider around Antarctica than in the Arctic.
Bitz et al. (Reference Bitz2005) performed a modeling study of the mean position and seasonal range of the ice edges in the North Atlantic, North Pacific, and South Atlantic sector of the Southern Ocean. The departure of the wintertime ice edge (here, the limit of 50 percent concentration in ice coverage) from a symmetrical ring around either pole was shown to depend primarily on coastline shape, ice motion, and the melt rate at the ice–ocean interface. At any location, the principal drivers of the oceanic heat flux that melts sea ice are absorbed solar radiation and the convergence of heat transported by ocean currents. In regions where the ice edge extends relatively far equatorward, absorbed solar radiation is the largest component of the ocean energy budget, and the large seasonal range of insolation causes the ice edge to traverse a large distance. In contrast, at relatively high latitudes, the ocean heat flux convergence is the largest component. It has a small annual range, so the ice edge there traverses a much smaller distance.
Holland and Kimura (Reference Holland and Kimura2016) analyzed the ice concentration budgets at both poles using AMRS-E brightness temperature data for 2003–2010. They derived a climatology of the ice concentration budget to enable observational decomposition of the seasonal dynamic and thermodynamic changes in ice cover. In both hemispheres, the spring ice loss was dominated by ice melting. In other seasons, ice divergence maintained freezing in the inner pack while advection caused melting at the ice edge, as ice was transported beyond the region where it could be sustained by thermodynamic processes. Mechanical redistribution by wind stress provided an important sink of ice concentration in the central Arctic and around the Antarctic coastline.
On synoptic time scales, ice edge retreat in the Barents Sea has been shown to respond to intense intrusions of moist air (maximum at 900 hPa) into the Arctic during autumn and winter through their impact on local temperature and sea ice concentration (Woods and Caballero Reference Woods and Caballero2016). Woods and Caballero found that the vertical structure of the warming associated with moist intrusions is amplified in the lower layers, corresponding to a transition of local conditions from a “cold clear” state with a strong inversion to a “warm opaque” state with a weaker inversion. Composite analyses of winter intrusions indicate surface downward longwave flux anomalies of up to 30 W m−2 and surface air temperature anomalies of up to 4.5 °C. In the marginal sea ice zone (75–80° N, 20–80° E), the passage of an intrusion also causes a reduction of 6 percent in sea ice concentration, which persists for many days after the intrusion has passed. There is a positive trend in the number of intrusion events crossing 70° N during December and January that can explain roughly 45 percent of the surface air temperature and 30 percent of the sea ice concentration trends observed in the Barents Sea during the past two decades.
Tsubouchi et al. (Reference Tsubouchi2012) estimated Arctic Ocean boundary fluxes through the four main gateways (Bering Strait, Davis Strait, Fram Strait, and Barents Sea) for 32 days in summer 2005. They found the following transport-weighted mean properties for water entering the Arctic: potential temperature of 4.49 °C and salinity of 34.50 psu. For water leaving the Arctic, including sea ice, the corresponding properties were 0.25 °C and 33.81 psu. Hence, the net effect was to freshen and cool the inflows by 0.69 psu in salinity and 4.23 °C, respectively. The net heat flux (including sea ice) was 189 ± 37 TW, representing a loss from the ocean to the atmosphere.
The main sources of ocean heating are solar absorption in summer and the ocean heat flux convergence (OFHC) year-round. The relative importance of these terms averaged over the marginal ice zones (MIZ) is shown in Table 7.3. The average OHFC near the ice edge is shown to be 65 W m−2 south of 62° N and 100 W m−2 north of 72° N. It is approximately 200 W m−2 in the ice-free Norwegian Sea as a result of the North Atlantic Current, compared to 40 W m−2 in the mostly ice-covered Bering Sea. Ice is maintained in the East Greenland Sea by the advection of ice and cold water in the East Greenland Current.
|Region||Absorbed solar radiation||OHFC|
|Sea of Okhotsk||67||33|
In the southern hemisphere in the model, autumn and winter growth rates are highest next to the continent in coastal polynya-like features. The ice edge advances rapidly in autumn, as the growth rates away from the coast are much higher then than in winter. In summer, there is no melt at the top of the ice due to snow cover and because melt ponds are virtually absent. Basal and lateral melt is important in that season.
7.7.2 Arctic Sea Ice
Arctic sea ice is currently made up of about 80 percent FYI and 20 percent MYI, a reversal from the proportions that prevailed up until the early 1990s (Arctic Climatology Project Reference Tanis and Smolyanitsky2000, Sea Ice Atlas). FYI has a higher fraction of level ice, whereas MYI has more ridges and hummocks. In the Arctic in winter, ice divergence accounts for half as much ice volume change as ice growth. In summer, basal plus lateral melt exceeds melt at the top surface in the Arctic Basin, except where continental climates have a strong influence on the sea-ice cover, as in the Canadian archipelago. At SHEBA in 1997–1998, Perovich et al. (Reference Perovich1999) reported approximately 100 cm of basal melt and only 30 cm of top melt.
The ice edge pattern in the northern hemisphere in summer is relatively zonal, while in winter it varies greatly with longitude. The location of the wintertime ice edge in the northern hemisphere depends on the ice dynamics, which advect ice toward the periphery of the ice pack, and basal and lateral melt, which is confined to a narrow band near the ice edge.
Arctic sea ice circulates slowly clockwise around the Beaufort gyre and moves from the Asiatic coast across the pole in the Transpolar Drift Stream (TPDS) to exit the Arctic via the Fram Strait in the East Greenland Current. Kwok et al. (Reference Kwok, Cunningham and Pang2004) summarized ice export estimates for 1978–2002 and examined, over a nine-year record, the associated variability in the time-varying upward-looking sonar (ULS) thickness distributions in Fram Strait. The average annual ice area flux over the period was 866,000 km2 yr−1. Between the 1980s and 1990s, the decadal difference in the net exported ice area was approximately 400,000 km2, or roughly half the annual average.
Using thickness estimates from ULS moorings, Kwok et al. (Reference Kwok, Cunningham and Pang2004) estimated the average annual ice volume flux (1991–1999) to be 2,218 km3 yr−1 (0.07 Sv). Over the ULS ice thickness data set, there was an overall decrease of 0.45 m in the mean ice thickness and a decrease of 0.23 m over the winter months (December through March). Correspondingly, the mode of the MY ice thickness exhibited an overall decrease of 0.55 m and a winter decrease of 0.42 m.
Krumpen et al. (Reference Krumpen2016) analyzed a data set of ground-based and airborne electromagnetic ice thickness measurements collected during summers between 2001 and 2012. The primary source of the surveyed sea ice leaving Fram Strait is the Laptev Sea; the age of this ice decreased from three to two years between 1990 and 2012. The thickness data consistently show a general thinning of sea ice over the last decade, with a decrease in modal thickness of second-year and multiyear ice. The mean thickness decreased from 2.58 m in 2001 to 2.17 m in 2012. Overall, Krumpen et al. found a positive trend in the monthly Fram Strait area flux, noting that it has increased by 25 percent since the 1960s.
The most prominent atmospheric driver of anomalous sea-ice motion across Fram Strait is an east–west dipole pattern of sea level pressure (SLP) anomalies with centers of action located over the Barents Sea and Greenland (Tsukernik et al. Reference Tsukernik2009) The association between the SLP dipole pattern and Fram Strait ice motion is maximized at 0-lag, persists year-round, and is strongest on time scales of 10–60 days.
Gudkovich (Reference Gudkovich1961) showed that there are two different oceanic circulation patterns in the Arctic Ocean. First, an anticyclonic regime exists where the area of the Beaufort gyre increases and the area of the cyclonic Laptev gyre shrinks; the TPDS originates from the Laptev, East Siberian, and Chukchi seas and transports ice toward the Greenland Sea. Second, a cyclonic regime is observed, with a contraction of the Beaufort gyre occurring simultaneously with an expansion of the Laptev gyre; the TPDS slows down and its entrance shifts toward the Beaufort Sea. Proshutinsky and Johnson (Reference Proshutinsky and Johnson1997) found that these regimes alternate every five to seven years.
A surface feature characterization of ice in the central Arctic and Beaufort–Chukchi Sea has recently been performed using airborne IceBridge data for 2009–2014 (Petty et al. Reference Petty2016). Elevation threshold values of 20 and 80 cm above a derived level ice surface were used, which means that the former included snow sastrugi features as well as ice deformation features. The results demonstrated predominantly higher surface features (1 m or greater) in the central Arctic region, mainly north of Greenland and the Canadian archipelago, and predominantly lower features (1 m or less) in the Beaufort–Chukchi Sea region. Feature heights were markedly higher (1.5–1.7 m or greater) along the coast of Greenland. Table 7.4 shows the overall results for 2009–2014. Regression of surface feature height with total ice thickness showed a mean correlation of 0.72, based on a square-root relationship.
|Central Arctic (a)||Beaufort–Chukchi||Central Arctic (b)||Beaufort–Chukchi|
|Mean (m)||Mode (m)||Mean (m)||Mode (m)|
Kapsch et al. (Reference Kapsch2016) have analyzed the role of downwelling longwave (LWD) and shortwave (SWD) radiation on summer sea ice in the Arctic. They found that positive LWD anomalies in spring and early summer have significant impact on the September ice extent, whereas winter anomalies produce only a small effect. Positive anomalies in spring and early summer initiate an earlier melt onset, thereby triggering several feedback mechanisms that amplify melt during the succeeding months. Realistic positive SWD anomalies appear to be important only if they occur after the melt has started and the albedo is significantly reduced. Simultaneous positive LWD and negative SWD anomalies during cloudy conditions during spring have a significant impact on summer sea ice, while summer clouds have almost no effect.
The relation between Arctic sea ice and clouds is seasonally dependent. It was examined during 2006–2008 by Kay and Gettelman (Reference Kay and Gettelman2009). No cloud response to sea-ice loss was found in summer, but low clouds did form over newly open water during early autumn. This seasonal variation in the cloud response to sea-ice loss can be explained by near-surface static stability and air–sea temperature gradients. During summer, temperature inversions and weak air–sea temperature gradients limit atmosphere–ocean coupling. In contrast, relatively low static stability and strong air–sea gradients during early autumn permit upward turbulent fluxes of heat and increase low cloud formation over newly open water.
This work was extended by Taylor et al. (Reference Taylor2013) using Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP) data, CALIPSO CloudSat Cloud Profiling Radar (CPR), Clouds and Earth’s Radiant Energy System (CERES), and AQUA MODIS in a cloud property vertical profile merging process for July 2006 through June 2010. The covariances between Arctic low cloud properties and sea ice concentration were quantified. Smaller average cloud fraction and liquid water were found where there was more sea ice. The largest-magnitude cloud–sea ice covariance occurred between 500 and 1,200 m when the lower tropospheric stability (θ850 hPa – θSFC) was between 16 and 24 K. Increased lower tropospheric stability was associated with decreases in low cloud fraction, cloud liquid water, cloud ice water, and cloud total water. The covariance between low cloud properties and sea ice was found to be largest in autumn. Regionally, the Laptev, Chukchi, and Beaufort seas exhibited the largest regional covariance between cloud properties and sea ice in summer and autumn; the Barents and Kara Sea regions exhibited the largest covariance in winter. Cloud properties were found overall to vary more between two atmospheric regimes than with sea ice concentration. However, the covariance between the liquid water path (LWP) and the sea ice concentration in autumn was of similar magnitude to the average LWP differences between the stable and highly stable regimes.
Arctic sea-ice predictability has been shown to involve the effects of both persistence and re-emergence of anomalies. A review of predictability was presented by Guemas et al. (Reference Guemas2016). Persistence of sea ice area (SIA) was shown to have a characteristic e-folding time scale that varies seasonally from two to five months, based on data for 1978–2008 (Blanchard-Wrigglesworth et al. Reference Blanchard-Wrigglesworth2011). July–August SIA was significantly correlated with that in September. Guemas et al. also estimated the persistence time scale of sea ice thickness (SIT) as approximately one year in the central Arctic and a few months in the seasonal ice zone. The ocean provides important predictability. Based on observations for 1864–1998, Vinje (Reference Vinje2001) identified a strong link between the Atlantic water temperatures in the southern Norwegian Sea and the sea ice extent (SIE) two to three years later in the Barents and Kara seas, via the warm advection by the North Atlantic Current.
Re-emergence of SIA anomalies has been shown to be a further factor in persistence. Blanchard-Wrigglesworth et al. (Reference Blanchard-Wrigglesworth2011) highlighted two different mechanisms for the re-emergence of SIA anomalies on time scales from a few months up to one year. One mechanism explains the re-emergence from the melt season to the growth season due to the persistence of SST anomalies. A negative (positive) SIA anomaly in spring is associated with a positive (negative) SST anomaly along the sea ice edge, which favors a negative (positive) SIA anomaly when the sea ice cover returns during the next autumn. A different mechanism explains the re-emergence from the growth season to the melt season as a result of the persistence of SIT anomalies. A positive (negative) SIA anomaly in the growth season is associated with an early (late) date of freeze-up, locally creating a positive (negative) SIT anomaly that slows down (accelerates) the sea-ice retreat during the next spring, and is therefore associated with a local positive (negative) SIA anomaly. Bushuk and Giannakis (Reference Bushuk and Giannakis2017) extended this work by showing the roles of SIT to sea ice concentration (SIC) in growth-to-melt season re-emergence and of SST and SLP to SIC in melt-to-growth season re-emergence.
The Sea Ice Outlook (SIO) has been prepared since 2008 by many research groups. Hamilton and Stroeve (Reference Hamilton and Stroeve2016) analyzed the performance of more than 400 predictions from SIO’s first eight years, testing for differences in ensemble skill across years, months, and five types of methods: heuristic, statistical, mixed, and ice–ocean or ice–ocean–atmosphere models. Their results highlighted a pattern of easy and difficult years, corresponding roughly to the distinction between climate and weather. Difficult years, in which most predictions were far from the observed extent, tended to have large positive or negative excursions from the overall downward trends. In contrast to these large interannual effects, ensemble improvement from June to July and August was modest. Among method types, predictions based on statistics and ice–ocean–atmosphere modeling more closely matched the actual effects. Thinning ice that is sensitive to summer weather, complicating prediction, reflects the current transitional era between a past Arctic cool enough to retain much thick, resistant multiyear ice, and a warmed future Arctic where little ice remains at the end of summer.
7.7.3 Landfast Ice
Landfast ice comprises two main components – bottom fast ice where grounded pressure ridges anchor the ice in shallow water (less than approximately 18 m depth) and attached floating ice over deep water. Off northern Alaska, floating fast ice covers most of the shelf out to the 18-m isobath, but inward of the 2-m isobaths the ice is grounded on the seafloor. This boundary is marked by a grounded pressure ridge (stamukhi). Ridging is caused by wind and wave action in the early winter (Barry Reference Barry, French and Slaymaker1993; Barry et al. Reference Barry, Moritz and Rogers1979; Mahoney et al. Reference Mahoney2007). Druckenmiller et al. (Reference Druckenmiller2012) have reported that landfast ice thickness in spring off Barrow during 2008–2012, as measured along trails constructed by Iñupiat hunters, had modal thickness values of 1.5–1.6 m. Mahoney et al. (Reference Mahoney, Eicken and Hendricks2015) have reported on the landfast ice mass balance station operated off Barrow by the University of Alaska since 2000; this station has measured changes in the growth and melt of landfast ice. The authors conclude that the mean annual maximum ice thickness of 1.5 m from 2000 to 2015 was significantly thinner than the thicknesses of around 1.8 m commonly reported during the 1970s – a difference attributed to shorter, warmer winters.
In Home Bay, Baffin Island, the edge of the fast ice can be over water that is 180 m deep some 70 km offshore (Jacobs et al. Reference Jacobs, Barry and Weaver1975). In the Kara Sea, Volkov et al. (Reference Volkov2002) discerned two basic mechanisms of fast ice formation. First, grounded pressure ridges stabilize the ice, facilitating fast-ice growth in shallow regions (less than 25 m depth). The spatial extent of this ice is limited by the thickness of the pressure ridge and the ocean depth. Second, further fast-ice growth may occur as ice floes drift onshore and attach themselves onto the coast or fast-ice edge. This second mechanism is the main formation mechanism in the northeastern Kara Sea. Olason (Reference Olason2016) notes that offshore islands will prevent the ice drift under offshore winds, allowing fast ice to form over deep water. A crucial element is the formation of static arches in granular materials passing through openings and converging channels. It has been shown that the strength of the arch depends critically on the uniaxial compressive strength of the material. Thus, ice arches will form in channels and narrow passages.
According to 10-day sea ice charts for the region during 1933–2006 (Mahoney et al. Reference Mahoney2008), fast ice forms first off the Taimyr Peninsula (the Severozemelsky region) around December 11. It remains mostly unaltered until the summer breakup, which occurs from July 11 to August 11.
Russian scientists have long recognized “ice massifs” in the Eurasian shelf seas. Massifs, formed by fast ice, appear and persist in regions with shallow depths and irregular coastlines. Timokhov (Reference Timokhov1994) noted that in the eastern parts of the Kara and Laptev seas, as well as to the west of the East Siberian Sea, fast ice extends over hundreds of kilometers from the shore, forming the basis for the Severnaya Zemlya, Yana, and New Siberian ice massifs. Each ice massif has its own respective flaw polynya. For example, in the Kara Sea, the Anderma and Yamal polynyas correspond to the Novaya Zemlya massif, and the Ob-Yenisei and Severnaya Zemlya polynyas correspond to the northern Kara massif. In the Laptev Sea, opposite the Taimyr massif, there is an extensive Siberian polynya.
7.7.4 Sea Ice Leads
Sea ice leads are important climatologically, oceanographically, and geophysically. Leads (open and refrozen) account for about half of the turbulent heat transfer to the atmosphere in winter, even though they represent only about 1 percent of the sea ice area (Maykut Reference Maykut1978).
A five-year climatology of leads in the western Arctic was derived by Miles and Barry (Reference Miles and Barry1998) using Defense Meteorological Satellite Program (DMSP) thermal and visible band imagery with 2.7- and 0.6-km resolution, respectively. In these images, leads of 200–300 m width are detectable. The occurrence (density) and orientation of leads are derived from gridded maps made at 10-day intervals. In the Miles and Barry climatology, relative lead densities are observed to be highest in early winter, decreasing 20 percent from November through April. The highest densities are observed in the central Canada Basin, and the lowest are in the East Siberian Sea. Preferred lead orientations are identified as generally north–south in the Beaufort Sea sector and east–west in the East Siberian Sea sector, with transitional orientations in the intermediate area. The spatial patterns of mean ice divergence and lead density are in general correspondence, with the highest values found in the Beaufort Sea and the lowest in the East Siberian Sea. The strength of the association between the lead density and divergence is indicated by the correlation coefficient (r = 0.68); the r values for November–January and February–April are 0.53 and 0.71, respectively. The circular correlation of lead orientation with the shear is 0.81 for NDJ and 0.75 for FMA. The mean angle of 90.7° for NDJ and 89.1° for FMA indicates a strong association between the two orientations: In theory, fractures in ice are expected to form orthogonally to the direction of the principal stress. The preeminent geometric feature of the lead distributions is a characteristic rectilinear pattern, with an intersection angle of about 30°, in accordance with theory (Erlingsson Reference Erlingsson1991). This angle appears to be consistent throughout the range of scales observable on the images, from kilometers to hundreds of kilometers.
Willmes and Heinemann (Reference Willmes and Heinemann2016) used the thermal infrared (IR) data from MODIS for January to April 2003–2015 over the entire Arctic Ocean to determine lead frequencies and regional characteristics on a daily basis. They found that the marginal ice zones in Fram Strait and the Barents Sea are the primary regions for lead activity. There are also distinct patterns of predominant fracture zones in the Beaufort Sea and along the shelf-breaks, mainly in the Siberian sector along the flaw polynyas in the Kara, Laptev, and East Siberian seas (Figure 7.8). Note that some areas of leads appear to be related to shoals – for example, Hanna Shoal in Figure 7.8. Wang et al. (Reference Wang2016) established that wintertime lead area fraction during the last three decades has not undergone significant trends. However, a substantial positive trend in lead area fraction was found in summer, located where sea ice concentration is already low.
Lindsay and Makshtas (Reference Lindsay, Makshtas, Bobylev, Kondratyev and Johannesen2003) assembled energy balance data over the Arctic pack ice using measurements from Soviet North Pole stations; selected data are shown in Table 7.5. Solar downwelling peaks in June, whereas net radiation reaches its maximum in July. The turbulent energy flux terms are small throughout the year. About 60 W m−2 is available for ice melting in June and July.
7.7.5 Arctic Sea Ice Trends
Records of sea ice conditions are highly variable in their source (ship, aircraft reconnaissance, satellite remote sensing, and submarine sonar), descriptors (extent, concentration, thickness), regional extent, and time interval covered. Recently, Walsh et al. (Reference Walsh2017) have assembled an Arctic database from 1850 to the present. In doing so, they combined information from six sources prior to the availability of satellite passive microwave data in 1979:
1. The W. Dehn collection of sea ice charts for the Alaskan region for 1953–1986
2. The Russian Arctic and Antarctic Research Institute database, spanning 1933–2006 and covering the Eurasian Arctic, including the Chukchi Sea
3. The National Research Council of Canada sea ice data spanning about 1815–2000 and covering the eastern Canadian waters – Baffin Bay, Davis Strait, the Labrador Sea, and the Gulf of St. Lawrence
4. The historical sea ice charts of the Danish Meteorological Institute (DMI), spanning 1894–1956
5. The Arctic Climate System Study (ACSYS) database of North Atlantic ice edge positions, 1750–2002
6. Whaling ship reports for the Alaskan region from the Bockstoce collection, 1849–1914
The coverage of the DMI charts includes all of the Atlantic sector and the Pacific sector as far south as the Bering Sea.
The time series of March and September ice extent from the Walsh et al. (Reference Walsh2017) database is shown in Figure 7.9. These authors have also demonstrated that the decrease of pan-Arctic sea-ice extent from the 1920s to the 1940s is most apparent in the summer months. However, the recent decrease is apparent in all seasons and is the only systematic excursion from the mean that is prominent in all seasons. Winter ice reached a record low of 14.42 million km2 in March 2017, surpassing previous records in March 2015 and 2016.
A long-term perspective on the recent sea ice decline in late summer has been provided by Kinnard et al. (Reference Kinnard2011) using four types of high-resolution terrestrial proxy records from the circum-Arctic region and ocean cores. They conclude that while extensive uncertainties remain, especially before the sixteenth century, both the duration and the magnitude of the current decline in sea ice seem to be unprecedented for the past 1,450 years. Enhanced advection of warm Atlantic water to the Arctic seems to be the main factor driving the decline of sea ice extent on multidecadal time scales. This process is also observed during the recent sea ice decline (Spielhagen et al. Reference Spielhagen2011). Paradoxically, Kinnard et al. also report an interval from the late fifteenth to the early seventeenth century, during the Little Ice Age, when sea ice extent also decreased. They attribute this to enhanced southerly advection of warm air into the Arctic.
Observations in the Bering Strait were made in summers 1778 and 1779 by British navigators James Cook and Charles Clerke. The ice edge was encountered at 70.7° and 70.3° N in the respective summers. Stern (Reference Stern2016) reported that these limits were essentially maintained until the 1990s, when northward retreat of summer ice in the Chukchi Sea became increasingly evident.
Regional trends have also been examined by Walsh et al. (Reference Walsh2017). For example, increased sea ice cover relative to preceding decades was noted in the Greenland Sea from the 1960s to the early 1970s. This expansion of the ice cover, apparent in both September and March, coincided with the Great Salinity Anomaly (see Box 4.3) that migrated through the East Greenland waters in the 1960s and circulated around the North Atlantic over the next decade (Dickson et al. Reference Dickson1988). The outstanding feature of both the Barents Sea and Greenland Sea time series is the decrease in sea ice cover during the last few decades. This decrease took both regions to period-of-record lows in March and, in at least one or two of the years since 2000, to record lows in September.
The Canadian archipelago saw frequent light ice years from the 1880s to the early 1910s, and then from the 1940s through the 1950s. Ice cover in this region increased from about 1960 to 1980, consistent with a cooling in this region during these decades. However, since the 1990s, the region has seen an unprecedented retreat of sea ice during the warm season.
The Beaufort and Chukchi seas are dominated by the recent retreat of summer sea ice. While both seas continue to be completely ice covered in March, the September sea ice coverage has decreased so precipitously that the extent in the past few years has been less than half the extent in the 1970s and 1980s. The Chukchi Sea, in particular, has been nearly ice free in autumn since September 2007. The Bering Sea ice, which is seasonal, does not show a significant reduction over the post-1850 period, in contrast to the Beaufort and Chukchi seas.
The most consistent sea ice concentration record is that provided by passive microwave remote sensing since late 1978 (Meier et al. Reference Meier2014b). This record began with the Scanning Multichannel Microwave Radiometer (SMMR) on NASA’s Nimbus 7 satellite, which was succeeded in 1987 by the Special Sensor Microwave Imager (SSM/I) on DMSP satellites until 2008, and by the Special Sensor Microwave Imager/Sounder (SSMIS) from 2008 to the present. The 1980s showed mostly interannual variability, but this was followed by progressive decline, especially in summer ice extent in the 2000s (Serreze and Stroeve Reference Serreze and Stroeve2015). The average decrease was 13 percent decade−1 in September, compared with less than 3 percent decade−1 in March (see Figure 7.9). The average ice area (1979–2013) for March was 15.5 × 106 km2, and that for September 6.4 × 106 km2 (Meier et al. Reference Meier2014b).
A record September sea ice concentration minimum of 4.3 million km2 was set in 2007 (Stroeve et al. Reference Stroeve2012). This record was broken in September 2012 (Figure 7.10), when the ice shrank to 3.4 million km2 (Figure 7.11), in part as a result of the export of multiyear ice through Fram Strait and in part due to the influence of a North Pacific storm in early August that broke up the already thin ice. September 2017 featured the eighth lowest sea ice extent on record.
The age of the sea ice has decreased greatly since the 1980s. Figure 7.12 shows that the proportion of FYI increased from about 55 percent in 1985 to 70 percent in 2016, while ice aged four years or greater has vanished.
The atmospheric response to sea ice loss during 1979–2009 has been analyzed by Smith et al. (Reference Smith2017) using the Hadley Centre GEM3 coupled model. They found that a weak low sets up over the Arctic in summer and autumn, which also leads to warming in the North Atlantic Ocean. Increased Antarctic sea ice over the same time interval drives a poleward shift of the southern hemisphere midlatitude jet, especially in the cold season.
The relative contributions of greenhouse gas forcing and internal atmospheric variability to summer sea ice loss have recently been studied by Ding et al. (Reference Ding2017). Using model calculations, they showed that internal variability has dominated the Arctic summer circulation trend and may be responsible for about 40 percent of the overall decline in September sea ice since 1979. The tendency for a stronger anticyclonic (clockwise) circulation over Greenland and northeastern Canada in June, July, and August has increased the downward longwave radiation above the ice, with warming and moistening of the lower troposphere also occurring. These changes are followed by negative sea ice anomalies in September. The source of the internal variability is indicated by a relationship between SST variability in the tropical Pacific and annual mean atmospheric circulation over the Arctic, centered over Greenland, that is a driver of the sea ice trend.
Changes in regional sea ice extent in the Arctic during autumn and early winter 1979–2014 have been analyzed by Chen et al. (Reference Chen, Alley and Zhang2016). The largest negative trends (approximately −20 percent decade−1) were found during autumn in the Beaufort Sea, the Barents–Kara Sea, and the Laptev–East Siberian Sea. During early winter, the largest trends in sea ice extent were found in the regions of Hudson Bay and the Barents–Kara Sea, around −10 percent decade−1. Sea ice losses in the Beaufort Sea and the Barents–Kara Sea were associated with a cooling over Eurasia, but in the former region the circulation anomaly was reminiscent of a Rossby wave train across the North Pacific, whereas in the latter area the pattern projected onto the negative phase of the Arctic Oscillation.
Analysis of September open water fraction in the Pacific and Atlantic sectors for 1979–2014 by Goldstein et al. (Reference Goldstein2016) has suggested the development of a statistically significant shift in the mean and an increase in the variance around 1988 and another breakpoint around 2007 in the Pacific sector. Breakpoints in the Atlantic sector record of open water were also evident in 1988 and 2007, but were more weakly significant. The breakpoints appeared to be associated with concomitant shifts in average ice age, and tended to lead to changes in Arctic circulation regimes.
The characteristics of changes in ice conditions during 1979–2014 were very different in the Atlantic and Pacific sectors (divided by the 100° E and 100° W meridians) of the Arctic (Lynch et al. Reference Lynch2016). The trend in September open water fraction was 5.2 percent decade−1 in the Arctic north of 70° N, only 2.1 percent decade−1 in the Atlantic sector, but 9.1 percent decade−1 in the Pacific sector. The per-decade trend reached 17.2 percent in the East Siberian Sea, 16.0 percent in the Chukchi Sea, and 12.0 percent in the Beaufort Sea. Anomalies in the Pacific sector ice cover can be partially compensated for by anomalies of opposite sign in the Atlantic sector. An assessment of linkages between summer atmospheric patterns and sectoral anomalies in the area of maximum open water north of 70° N demonstrates that there is asymmetry in the mechanisms.
Years with low ice extent and high open water fraction are uniformly associated with positive 925-hPa temperature anomalies and southerly flow in both the Atlantic and Pacific sectors. However, years with high ice extent and low open water fraction in both sectors reveal two dominant mechanisms. Some years with anomalously low maximum open water fraction are associated with negative temperature anomalies and southerly transport – a cool summer pattern that allows ice to persist over larger areas. In contrast, other low-open-water years are characterized by a mechanism, whereby – even when melting – ice cover is continually replenished by advection from the north.
According to Alexiev et al. (Reference Alexiev, Glok and Smirnov2016), increase of surface air temperature (SAT) in the marine Arctic (the part covered with sea ice in winter) was closely related to reduction of sea ice extent (SIE) in summer for 1980–2014. Based on this finding, anomalies of Arctic September SIE were reconstructed from the beginning of twentieth century using a linear regression. The reconstructed SIE shows a substantial decrease in the 1930–1940s, with a minimum occurring in 1936, although the decrease was only a half of the decline in the record year of 2012.
Recent increases in Atlantic water inflow into the eastern Eurasian Basin (north of Severnaya Zemlya) have been reported by Polyakov et al (Reference Polyakov2017). Since 2003, enhanced release of oceanic heat has reduced winter sea-ice formation at a rate now comparable to losses from atmospheric thermodynamic forcing. Polyakov et al. note that release of 1 W m−2 over the year causes a sea ice loss of 10 cm; upward flux through the pycnocline averaged 12 W m−2 for winter 2013–2014 and 7.5 W m−2 for winter 2014–2015. These are equivalent to 54- and 40-cm reductions in ice growth, respectively, in the east Eurasian Basin. The researchers term this process the “atlantification” of the Eurasian Basin.
Relations between sea ice and tropical convection as expressed in the Madden Julian Oscillation (MJO) were investigated by Henderson et al. (Reference Henderson, Barrett and LaFleur2014). Anomalies in daily change in sea ice concentration were isolated for all phases of the Real-Time Multivariate MJO index during both summer (May–July) and winter (November–January) months. The relationship between sea ice concentration and the MJO included three aspects. First, the MJO projects onto the Arctic atmosphere in both winter and summer, as shown by statistically significant “wavy patterns” at 500 hPa, with a variety of wave numbers, and consistent anomaly sign changes in composites of surface and mid-tropospheric atmospheric fields for different MJO phases. In November–January, height anomalies in MJO phase 2 (convection over the Indian Ocean) resemble positive AO polarity while height anomalies in phases 6 and 7 (convection over the Western Pacific) resemble negative AO polarity. Second, the MJO modulates Arctic sea ice in both summer and winter seasons, with the region of greatest variability shifting with the migration of the ice margin poleward (equatorward) during the summer (winter) period. This variability is supported by corresponding anomalies in surface wind and temperature. Third, the MJO modulates Arctic sea ice regionally, often resulting in dipole-shaped patterns of variability between anomaly centers in the Barents and Greenland seas in January.
For the southern Beaufort Sea and Amundsen Gulf, Galley et al. (Reference Galley2008) analyzed Canadian Ice Service charts for 1980–2004. In summer, a trend toward increased old sea ice concentration occurred near the mouth of the Amundsen Gulf, with a trend toward decreasing summer first-year sea ice farther west. In winter, increasingly the thick first-year sea ice extent appears to be replacing young sea ice within the flaw lead system in the region. The dynamically driven breakup of sea ice in spring in the Amundsen Gulf is a highly variable event, taking between two and twenty-two weeks to completely remove ice from the gulf. The timing and duration of the open water season depend upon the extent and timing of old ice influx. Freeze-up occurs very quickly, proceeding from west to east with little temporal variability.
Ice in the Eurasian Arctic has generally decreased since 1933, based on data from Soviet and Russia ice charts (Mahoney et al. Reference Mahoney2008). The retreat has not been continuous, however, with the data showing two periods of retreat separated by a partial recovery between the mid-1950s and mid-1980s. The charts, in combination with air temperature records, suggest that the retreat in recent years is pan-Arctic wide and year-round in some regions, whereas the retreat in the early to mid-twentieth century was confined to summer and autumn in the Russian Arctic.
Rothrock et al. (Reference Rothrock, Yu and Maykut1999) showed changes in Arctic ice thickness by comparing submarine sonar ice draft data from 1958 through 1976 to measurements from the 1990s. Their results indicate that there was thinning at every point of comparison between 1993 and 1997 with similar data acquired between 1958 and 1976; the mean ice draft at the end of the melt season decreased by about 1.3 m (40 percent) in most of the deep water portion of the Arctic Ocean, from 3.1 m in 1958–1976 to 1.8 m in the 1990s.
Kwok and Rothrock (Reference Kwok and Rothrock2009; Kwok et al. Reference Kwok and Rothrock2009) have reported that within the data release area of declassified submarine sonar measurements (covering 38 percent of the Arctic Ocean), the overall mean winter ice thickness of 3.64 m in 1980 can be compared to a 1.89 m mean during the last winter (2008–2009) of the ICESat record – an astonishing decrease of 1.75 m in thickness. Between 1975 and 2000, the steepest rate of decrease was 0.08 m yr−1 in 1990, compared to a slightly higher winter/summer rate of 0.10/0.20 m yr−1 in the five-year ICESat record (2003–2008).
There have been two recent expeditions to the Arctic to measure ice and snow cover thickness. Haas et al. (Reference Haas2017) made in situ measurements at ten sampling sites in the Lincoln Sea between Ellesmere Island and 87.1° N in April 2017. Mean and modal total ice thicknesses ranged between 2 and 3.4 m and between 1.8 and 2.9 m, respectively. Coincident snow thicknesses ranged between 0.3 and 0.47 m (mean) and 0.1 and 0.5 m (mode). There was excellent agreement with the snow climatology published by Warren et al. (Reference Warren1999) and with published long-term ice thinning rates.
Merkouriadi et al. (Reference Merkouriadi2017) and Gallet et al. (Reference Gallet2017) investigated the physical properties of first-year (FYI) and second-year (SYI) ice in the Atlantic sector of the Arctic Ocean, during the Norwegian young sea ICE (N-ICE2015) expedition (January–June 2015). Snow depth was 41 ± 19 cm in January and 56 ± 17 cm in February, which is significantly greater than Warren et al. (Reference Warren1999) described for this region. The snow water equivalent was 14.5 cm over FYI and 19 cm over SYI. For April–June overall, the snow thickness was about 20 cm greater than the climatology for SYI, with an average of 55 ± 27 cm and 32 ± 20 cm for FYI.
Sea ice thickness and volume were determined by Kwok and Rothrock (Reference Kwok and Rothrock2009) using ICESat data from 2003 to 2008. They found a greater than 42 percent decrease in MYI coverage since 2005, with a remarkable thinning of approximately 0.6 m in MYI thickness over four years. In contrast, the average thickness of the seasonal ice in midwinter (approximately 2 m), which covered more than two thirds of the Arctic Ocean in 2007, exhibited a negligible trend. Total MYI volume in the winter has experienced a net loss of 6,300 km3 (more than 40 percent) in the four years since 2005, while the FYI cover gained volume owing to increased overall area coverage. The Arctic has a maximum ice volume of 16,400 km3 in the spring.
Landy et al. (Reference Landy2017) used ICESat and Cryosat 2 data to analyze ice thickness in the eastern Canadian Arctic during 2003–2016. The mean seasonal growth rate was 23 cm month−1 from November to April. In Hudson Bay, the ice was 40 cm thicker in the east than in the northwest, whereas in Baffin Bay it was 20 cm thicker in the west than in the east. The April thickness reached 2.12 m in Foxe Basin; it was 1.67 m in eastern Hudson Bay and 1.25 m in the northwest part.
There is extensive and persistent fast ice in the waters of the Canadian Arctic archipelago. Howell et al. (Reference Howell2016) analyzed trends observed at the Cambridge Bay, Resolute, Eureka, and Alert sites during 1957–2014, representing some of the Arctic’s longest records of landfast ice thickness. Observed end-of-winter (maximum) trends of landfast ice thickness were statistically significant at Cambridge Bay (−4.31 ± 1.4 cm decade−1), Eureka (−4.65 ± 1.7 cm decade−1), and Alert (−4.44 ± 1.6 cm decade−1), but not at Resolute. Over the more than fifty-year record, the ice thinned by approximately 0.24–0.26 m at Cambridge Bay, Eureka, and Alert, with essentially negligible change occurring at Resolute. Although statistically significant warming in spring and fall was present at all sites, only low correlations between temperature and maximum ice thickness were present; snow depth was found to be more strongly associated with the negative ice thickness trends.
Russian fast ice thickness measurements were reported by Frolov et al. (Reference Frolov and Johannessen2005). They show a positive trend during 1940–1973 of +0.35 cm yr−1 and a negative trend of −0.52 cm yr−1 during 1973–2000.
When Smedsrud et al. (2016) analyzed a new time series from 1935 to 2014 of Fram Strait sea ice area export, they found that the long-term annual mean export is about 880,000 km2, representing 10 percent of the sea ice–covered area inside the basin. There was large interannual and multidecadal variability, and no long-term trend, over the past eighty years. However, the last decade has witnessed increased ice export, with several years having annual ice exports that exceeded 1 million km2. Since 1979, annual export has increased by about 6 percent decade−1, due to higher southward ice drift speeds caused by stronger southward geostrophic winds, a phenomenon largely explained by increasing surface pressure over Greenland. Spring and summer area export increased by 11 percent decade−1. In contrast, the 1950–1970 period had relatively low export during spring and summer, and mid-September sea ice extent was consistently higher during these decades than both before and after them. Thus export anomalies during spring have a clear influence on the following September sea ice extent in general, and for the recent decade, the export may be partly responsible for the accelerating decline in Arctic sea ice extent. Ice export during winter will generally result in new ice growth and contributes to thinning inside the Arctic Basin. Increased ice export during summer or spring will, in contrast, contribute directly to open water farther north and a reduced summer sea ice extent through the ice–albedo feedback. The relatively low spring and summer export from 1950 to 1970 is, therefore, consistent with a higher mid-September sea ice extent for those years.
7.8 Antarctic Sea Ice
There are distinct differences between sea ice characteristics in the two polar regions. Nearly all ice in the Southern Ocean is seasonal, and its structure differs from that in the Arctic in significant ways. Antarctic sea ice extent retreats to a minimum of 3.1 × 106 km2 in February, which is just 17 percent of the maximum extent of 18.5 × 106 km2 in September (Parkinson Reference Parkinson2014). The extent of pack ice and of the Marginal Ice Zone (MIZ) has been shown to depend on the passive microwave algorithm that is used.
Stroeve et al. (Reference Stroeve2016) compared the results from the NASA Team and Bootstrap algorithms. The annual values (106 km2) for all Antarctica are as follows:
Thus, applying the same thresholds for both sea ice algorithms results in a MIZ from the NASA Team algorithm that is, on average, twice as large as in the Bootstrap algorithm and that contains considerably more broken ice within the consolidated pack ice.
There are rather sparse data available on ice thickness in the Antarctic. Two decades of data compiled by the SCAR Antarctic Sea Ice Processes and Climate (ASPeCt) program, totaling more than 23,000 observations, gave a mean thickness of all ice as 0.87 ± 0.91 m, compared with a level-ice thickness of 0.62 m (Worby et al. Reference Worby2008). Kurtz and Markus (Reference 333Kurtz and Markus2012) used satellite laser altimetry data from NASA’s ICESat combined with passive microwave measurements to analyze basin-wide changes in Antarctic sea ice thickness and volume over a five-year period from 2003 to 2008. The ICESat data for 2003–2005 and the shipboard measurements collected during the ASPeCt program show good agreement in spring:
|ICESat mean (m)||Ships mean (m)|
The thickest ice resides in the western Weddell Sea, the Bellingshausen and Amundsen seas, the western Ross Sea, and surrounding the Antarctic coastline. The thinnest ice is found in the eastern Weddell Sea, the eastern Ross Sea, portions of the Indian and Pacific Oceans, and toward the northern edge of the sea ice. The ICESat record shows that the 2003–2008 mean ice volume reached a minimum of 3,357 km3 in the summer, grew to 8,125 km3 in the autumn, and reached its maximum of 11,111 km3 in the spring.
Strong wind and wave interactions can significantly increase sea ice thickness by rafting and ridging (Lewis et al. Reference Lewis2011) as can heavy snow loading, flooding, and freezing at the top (Massom et al. Reference Massom2001). Nghiem et al. (Reference Nghiem2016) employed QuikSCAT data for 1999–2009 to determine ice trajectories in the Southern Ocean. Their work shows that sea ice, grown earlier in the ice season, drifts northward away from the Antarctic continent, forming a circumpolar frontal ice zone (FIZ) behind the ice edge. In the circumpolar sea ice zone adjacent to the sea ice edge, the scatterometer data exhibit a band of strong radar backscatter, which is consistent with the signature of older, thicker, and rougher sea ice with more snow cover in the FIZ. The formation of this band is attributable to a longer exposure to wind and wave actions, and thickening over time by ice growth and greater snow accumulation. This band of sea ice is up to 1,000 km wide and serves as a “Great Shield,” encapsulating and protecting younger and thinner ice in the interior ice pack. Three age classes of sea ice can be distinguished from the backscatter signatures: rough older ice, older ice, and younger ice. Additionally, permanent ice (ice shelf and fast ice) as well as melt on ice can be identified. Figure 7.13 shows the distributions of these ice classes near the maximum of each season for 1999–2009. In all years, the Antarctic sea ice cover was totally surrounded by a FIZ of rough, older ice. The outer ice edge of the FIZ is close to the coast in regions such as the Somov Sea, D’Urville Sea, and Mawson Sea, or far away from the coast – particularly in the Lazarev Sea, where the FIZ may extend all the way to Bouvet Island (Figure 7.13).
Nghiem et al. (Reference Nghiem2016) have shown that in general, the sea ice edge is determined by the presence of warm water and is typically enclosed by the −1.0 °C SST isotherm in each year. In the interior sea ice region behind the FIZ, persistent katabatic winds force the opening, production, and advection of ice. In the newly opened or ice divergence areas, protected behind the FIZ, the young ice can have a high growth rate. This process is enhanced by the very cold air that is advected off the continent and ice shelves. The isotherm is in the proximity of the southern ACC front, as delineated by Kim and Orsi (Reference Kim and Orsi2014).
7.8.1 Sea Ice Trends in the Antarctic
Recently, sea ice reported in ships’ log books from the Heroic Age of Antarctic exploration (1897–1917) have been analyzed by Edinburgh and Day (Reference Edinburgh and Day2016). They found that, while in most sectors the ice extent was comparable to today, in the Weddell Sea the edge was 1.0–1.7° farther north at that time. A proxy sea ice record from 1702 for the Amundsen–Ross Sea sector indicates opposite trends in winter sea ice extent in the Bellingshausen–Weddell Sea, where this extent has declined over time, and in the Amundsen–Ross Sea, where it expanded northward by 1° latitude during the twentieth century (Thomas and Abram Reference Thomas and Abram2016). Maximum extent in this region was observed during the mid-1990s.
Whaling ship records of inferred summer ice edge location from the 1930s to 1980s were analyzed by de la Mare (Reference de la Mare2009). The data suggest that there was a zonal mean southward migration of the summer (October–March) ice edge of 2.4° latitude in the 1970s–1980s compared to the 1930s–1950s. King and Harangozo (Reference King and Harangozo1998) used coastal station temperature data to infer a southward migration of the autumn and winter sea ice edge along the western Antarctic Peninsula of approximately 1° latitude between 1945–1954 and 1973–1994.
Satellite passive microwave records for the Southern Ocean for 1979–2014 from the NOAA/NSIDC Goddard-merged climate data record of monthly mean sea ice concentration product (Meier et al. Reference 334Meier2014a) show an increase in sea ice cover that is most pronounced in the summer months of December–April (Hobbs et al. Reference Hobbs2016). This trend is dominated by increased sea ice coverage in the western Ross Sea, which is offset by a strong decrease in the Bellingshausen and Amundsen seas. The trends in sea ice areal coverage are accompanied by related trends in yearly duration. The retreat dates in the Amundsen–Bellingshausen Sea were 1.2 days yr−1 earlier and those in the Ross Sea were 1.2 days yr−1 later; the corresponding advance dates were 1.9 days yr−1 later and 1.3 days yr−1 earlier, respectively. November 2016 witnessed a record low ice concentration, with the Amundsen Sea being almost ice free. The lower extents of sea ice coverage have continued and remain to be explained.
Re-examination of Nimbus 1, 2, and 3 satellite data from the 1960s has identified large interannual variations in sea ice extent (Gallagher al. Reference Gallagher, Campbell and Meier2014). The September 1964 ice mean area was a record 19.7 × 106 km2, while in August 1966 the maximum sea ice extent fell to 15.9 × 106 km2.
The atmosphere is thought to be the primary driver of sea ice trends. Cyclonic flow around the Amundsen Sea low drives warm poleward winds into the Antarctic Peninsula–Bellingshausen Sea region, along with a cold equatorward wind over the Ross Sea, with clear implications for the dipole in sea ice trends between these two regions (Hosking et al. Reference Hosking2013; Turner et al. Reference Turner2015). However, the linkages of the Southern Annular Mode (SAM) and the Amundsen Sea low are complex. A positive SAM trend has increased the southern westerlies but it is unclear how this change has affected sea ice extent. The ocean also has an essential role in explaining the seasonality of the trend patterns. Mixed-layer feedback processes between sea ice and ocean have had a role in modulating the sea ice trends, and there appears to be a spatial dependence on where these processes are important. Nevertheless, the record length is short, and Stroeve et al. (Reference Stroeve2016) point out that the sea ice increase over the last 36 years remains within the range of intrinsic internal variability.
Meehl et al. (Reference Meehl2016) have examined the relationship between the sea ice expansion between 2000 and 2014 and tropical Pacific climatic conditions. They found that the Interdecadal Pacific Oscillation, an internally generated mode of climate variability, transitioned from positive to negative in the late 1990s, with an average cooling of tropical Pacific sea surface temperatures. Sea-level pressure and 850-hPa wind changes near Antarctica since 2000 have been conducive to expanding Antarctic sea ice extent, particularly in the Ross Sea region in all seasons, involving a deepening of the Amundsen Sea low. These atmospheric circulation changes are mainly driven by precipitation and convective heating anomalies related to the Interdecadal Pacific Oscillation in the equatorial eastern Pacific, with additional contributions from convective heating anomalies in the Southwest Pacific convergence zone and tropical Atlantic regions.
Cerrone et al. (Reference Cerrone2017) examined the roles of the SAM, the Semiannual Oscillation (SAO), the Pacific–South American (PSA) teleconnection, and the zonal wave number 3 (ZW3) mode on the variability of sea ice concentration for 1982–2013. Most of the sea ice temporal variability was concentrated in the two- to four-year time range associated with the constructive superposition of the PSA and ZW3 patterns. Interannual variations were related to the SAM and SAO patterns. The two-year signal resulted from the superposition of the positive phases of SAM, ZW3, and PSA patterns. The 2.7-year signal combined positive oscillations of the PSA and ZW3, and the four-year signal featured superposed positive oscillations of the SAM and ZW3 patterns and the negative phase of the PSA. This four-year signal is apparent only after 2000, whereas the other two signals are noted throughout the record.
Comiso et al. (Reference Comiso2017) analyzed an improved passive microwave record of sea extent and surface temperature for November 1978–December 2015. They found that there was a strong correlation, with a one-month lag in surface temperature, measured at −0.96 during the growth season and −0.98 during the melt season. There was only a weak relationship between ice extent and the SAM index of circulation. The record ice extent of 20 million km2 in 2014 was also shown to display a high sensitivity to surface temperature. Surprisingly, Antarctic sea ice extent fell to a record minimum in October 2016 and reached an all-time low of 2.28 million km2 on March 3, 2017.
Turner et al. (Reference Turner2017) reported that during the spring months of September to November 2016, the Antarctic sea ice extent decreased by 6.82 million km2. They observed that ice retreat in the Weddell Sea took place rapidly via strong northerly flow with poleward heat fluxes, after an early maximum ice extent in late August. The Amundsen Sea low was at record strength in September. Rapid ice retreat occurred in the Ross Sea in November, when there was record high surface pressure, with the SAM at its most negative for that month since 1968.
Stuecker et al. (Reference Stuecker, Bitz and Armour2017) showed that the extreme El Niño event that peaked in December to February 2015–2016 contributed to pronounced extratropical southern hemisphere SST and sea ice extent anomalies in the eastern Ross, Amundsen, and Bellingshausen seas that persisted in part until the 2016 austral spring. A second factor was internal variability of the SAM, which promoted the exceptional low sea ice extent in November–December 2016.
A polynya (a Russian term) is an ocean area that is largely ice free and that is surrounded by sea ice or sea ice and land. It may form by either of two processes: as a sensible-heat polynya or as a latent-heat polynya. The former, which is thermodynamically driven, typically occurs when warm water upwells keeping the surface water temperature at or above freezing. This reduces ice production and may stop it entirely. A sensible-heat polynya forms in the open ocean and the upwelling is accounted for by the bottom topography. In contrast, a latent-heat polynya is an open water region between a barrier and the ice pack, where the ice is driven away from the coast, an ice shelf, a grounded iceberg, or landfast ice, by offshore winds or ocean currents. New ice forms in the open water, which is then herded downwind toward the first-year pack ice, where the new ice is consolidated onto the pack. Ice growth leads to latent heat release as well as brine expulsion. This process increases ocean salinity, causing the higher-density water to sink. Through this mechanism, latent-heat polynyas in the coastal regions of Antarctica serve as a major source of the world’s bottom waters.
Some polynyas are hybrids of the sensible-heat and latent-heat types. For example, the Barrow Coastal Polynya is considered to be a wind-driven hybrid polynya, with both sensible heat from upwelling Atlantic Water and wind-driven divergence caused by the northeasterly wind (Hirona et al. Reference Hirano2016).
Barber and Massom (Reference Barber, Massom, Smith and Barber2007) provide summary tables of the physical characteristics of many Arctic and Antarctic polynyas, and a map of Arctic polynyas (Figure 7.14). There are large polynyas along the Siberian coast and many smaller ones in the Canadian Arctic archipelago.
Preusser et al. (Reference Preusser2016) have analyzed MODIS data for sixteen circum-Arctic polynyas for November to March from 2002–2003 to 2014–2015. All polynya regions combined covered an average thin-ice area of 184 × 103 km2 and created an average total wintertime accumulated ice production of about 1,444 km3. The main contributors (53 percent) were the Kara Sea region, the North Water polynya, and scattered smaller polynyas in the Canadian Arctic archipelago. The mean thin-ice thickness was estimated to be 13 cm.
Tamura et al. (Reference Tamura, Ohshina and Nihashimm2008) estimated that about 10 percent of sea ice production in the southern hemisphere is accounted for by coastal polynyas. Mean values of annual cumulative sea-ice production for the four major Antarctic coastal polynyas for 1992–2001 were as follows: Ross Sea, 390 km3; Darnley, 181 km3; Mertz, 120 km3; and Shackleton, 110 km3. The total amount for thirteen polynyas was 1,410 km3.
The coastal polynya area around Antarctica during June, July, August, and September (wintertime) 1992–2008 was estimated from SSM/I data to be 245,000 km2 (Kern Reference Kern2009). The polynyas along East Antarctica (60–160° E) accounted for about 40 percent of the total; the most persistent were located along the Lars–Christensen Coast (LCC), Prydz Bay, the western Davis Sea, Mertz Glacier, and in the Ross Sea along the Ross Ice Shelf and in Terra Nova Bay. The polynya at the LCC was observed on 110 ± 5 days during winters 1992–2008 and covered an average area of 2,400 km2 on more than 90 days.
Dale et al. (Reference Dale2016) examined the relationship between wind strength and sea ice concentration anomalies in the Ross Sea polynya in winter (April–October) 2001–2014. Persistent weak winds near the edge of the Ross Ice Shelf were generally associated with positive SIC anomalies in the Ross Sea polynya. Conversely, negative SIC anomalies in this area occurred during persistent strong southerly winds. Strong winds caused significant advection of sea ice in the region. The sea motion anomalies indicated the production of new ice by thermodynamic growth.
According to Barber and Massom (Reference Barber, Massom, Smith and Barber2007), the Ross Sea Polynya (RSP) is the largest in the Antarctic, with a winter area of around 20,000 km2. Two smaller polynyas are located in the western part of the Ross Sea: the Terra Nova Bay Polynya (TNBP), with a mean area of 1,300 km2 and maxima up to 5,000 km2, and the McMurdo Sound Polynya (MSP), with an area about two thirds of the TNBP (Hollands and Dierking Reference Hollands and Dierking2016). The TNBP, which is oriented east–west, is bounded by the Drygalski Ice Tongue in the south and by the Campbell Ice Tongue in the north. It is maintained during the winter season by 25–40 m s−1 katabatic winds that are channeled by glacial valleys and flow off the ice sheet over the adjacent sea ice (Bromwich and Kurtz Reference Bromwich and Kurtz1984). There is a period of maximum efficiency in sea ice production from July to November.
Estimates of ice production and dense water formation in global polynyas during nine winters have been made by Oshima et al. (Reference Oshima, Nihashi and Iwamoto2016) based on Advanced Microwave Scanning Radiometer for EOS (AMSR-E) passive microwave data. They found that ice production rate is high in Antarctic coastal polynyas, in contrast to Arctic coastal polynyas. This is consistent with the formation of dense Antarctic Bottom Water (AABW). The Ross Ice Shelf polynya has by far the highest ice production (253 km3) in the southern hemisphere. The Cape Darnley polynya (65–69° E) is the second highest production area (127 km3). Ten other polynyas have ice productions ranging from 27 to 83 km3.
The Okhotsk Northwestern polynya exhibits the highest ice production (400 km3) in the northern hemisphere, and the resultant dense water formation leads to overturning in the North Pacific. The next largest polynyas are the North Water polynya, with 152 km3, and the Anadye–St. Lawrence Island polynya, with 140 km3; another six polynyas range in size from 15 to 71 km3. Most of the ice production in northern hemisphere polynyas occurs in autumn.
Like leads, polynyas are a source of heat and moisture to the atmosphere. Thus, they modify the weather in surrounding areas.
The North Water (NOW) in northern Baffin Bay, between Greenland and Ellesmere Island, covers an area of 85,000 km2 in spring. It forms south of an ice bridge across northern Smith Sound. Three recurring polynyas are recognized within the NOW region: the Smith Sound, Lady Ann Strait, and Lancaster Sound polynyas (Steffen Reference Steffen1985). Eventually, the three separate polynyas become contiguous in the early spring, forming the NOW polynya. Barber et al. (Reference Barber2001a, Reference Barber2001b) showed that the North Water is maintained by both latent-heat and sensible-heat processes. Airborne remote sensing was carried out over the NOW in winters 1978–1979 and 1980–1981 by Steffen and Ohmura (Reference Steffen and Ohmura1985) and Steffen and Lewis (Reference Steffen and Lewis1988). Gray–white ice with estimated thickness of 15–30 cm was the dominant surface type. Sea surface temperatures increased from west to east by 10–15 °C across northern Baffin Bay. The highest SSTs were off Cape Alexander in Smith Sound, where the range of −1 to −15 °C was 20 °C higher than the SSTs over the fast ice. Cells of upwelling warm (Atlantic) water were common off West Greenland and near Wolstenholme Island, the Carey Islands, and Smith Sound. Temperatures in these cells in December–January were most frequently in the range of −1.0 to −0.8 °C. The energy budget of the NOW in January is −56 W m−2 for Rn, −28 W m−2 for H, and −23 W m−2 for LE, with a residual of 222 W m−2 (Steffen and Ohmura Reference Steffen and Ohmura1985). For the six winter months, the subsurface heat supply from the water is 173.5 W m−2. The supply of heat by refreezing is no more than 35 W m−2, so the sensible heat extracted from the water is about 139 W m−2.
For the Kara Sea, Kern et al. (Reference Kern2005) estimated the average polynya area was 21.2 × 103 km2 for the winters (January–April) of 1996–1997 to 2000–2001, being as large as 32.0 × 103 km2 in 1999–2000 and smaller than 12 × 103 km2 in 1998–1999. The modeled cumulative winter ice-volume flux out of the Kara Sea varied between 100 and 350 km3 yr−1. Bareiss and Görgen (Reference Bareiss and Görgen2005) showed that in November–June of 1979–1980 to 2001–2002, the mean area of the West New Siberian polynya in the southeast Laptev Sea averaged 4,000 km2 and had a mean duration of 14 days, while the Annabar–Lena polynya averaged 3,000 km2 and had a mean duration of 22 days. The mean cumulative areas of the two were 1,713 × 103 km3 and 1,152 × 103 km2, respectively, associated with a mean frequency of 12.4 polynya events during November–June in all investigated regions of the Laptev Sea.
Winter polynya areas to the northeast of Svalbard have long been known as Whaler’s Bay. Their persistence and re-emergence between 2011 and 2014 was investigated by Ivanov et al. (Reference Ivanov2016). Recent increased seasonality of Arctic sea ice cover enables an enhanced influence of oceanic heat on sea ice and on the heat transported by Atlantic Water. The “memory” of ice-depleted conditions in summer is transferred to the autumn through excess heat content in the upper mixed layer, which in turn transfers this “memory” via thinner and younger ice to midwinter conditions. This thinner ice facilitates the formation of polynyas and leads. Thermohaline convection-induced upward heat flux from the Atlantic layer retards ice formation, either keeping ice thin or blocking ice formation entirely.
An open ocean polynya was observed in the Weddell Sea near Maud Rise during three austral winters, 1974–1976. It was identified in Electrically Scanning Microwave Radiometer (ESMR) imagery, but has not recurred since that time. Holland (Reference 332Holland2001) explains this feature through a mechanism by which modest variations in the large-scale oceanic flow past the Maud Rise seamount caused a horizontal cyclonic eddy to be shed from its northeast flank. The shed eddy transmitted a divergent Ekman stress into the sea ice, leading to a crescent-shaped opening in the pack. Thermodynamic interaction with the atmosphere further enhanced the opening by inducing oceanic convection. The Maud Rise polynya has not re-formed, probably as a result of enhanced ocean stratification due to freshening but had not recurred until September 2017 when an area of 80,000 km2 unexpectedly opened up. The causes are uncertain.
Almost all of the area between latitudes 60° and 65° S is ocean. Except near Antarctica, the waters of the Southern Ocean move eastward at 40–60 cm s−1, forced by the westerlies. The Antarctic Circumpolar Current (ACC) has a volume transport of approximately 173 Sv, three fourths of which is baroclinic. The Antarctic Polar Front (APF) is located south of the ACC axis and the maximum westerlies. Antarctic Bottom Water forms in autumn and winter when shelf waters sink down the continental slope at a rate of 8–10 Sv, mainly in the southwest Weddell and Ross seas. Rather than three fronts (as at Drake Passage), the circumpolar ACC comprises multiple frontal filaments. The Subantarctic Surface Water (SASW) and Antarctic Surface Water (AASW) water masses are found in the upper 500 m around Antarctica. Below 1,500 m there are three water masses: Circumpolar Deep Water (CDW), North Atlantic Deep Water (NADW), and Antarctic Bottom Water (AABW). The number and intensity of fronts are determined largely by the bathymetry. The number is reduced where the ACC’s path is constricted. Two sub-Antarctic fronts and a Polar Front have strong temperature and salinity gradients.
The Ross Sea embayment covers approximately 960,000 km2. The southern part is the Ross Ice Shelf; the remainder has sea ice for much of the year. The surface ocean circulation has a wind-driven cyclonic (clockwise) gyre, accompanied by upwelling deep water. The Weddell Sea, east of the Antarctic Peninsula, covers approximately 2.8 million km2. In the south is the Ronne–Filchner Ice Shelf. The wind-driven cyclonic gyre transports sea ice northward in the western part. Offshore winds lead to the formation of coastal polynyas. Surface energy budget analysis has shown that sensible heat fluxes from the 5–7 percent of the area that comprises leads or coastal polynyas largely balance downward fluxes over the sea ice.
The Arctic Ocean is considered to be a Mediterranean sea due to its relative isolation. Continental shelves cover 2.5 million km2 and receive massive amounts of river runoff. The Arctic Ocean is divided by the Lomonosov Ridge into the Canadian and Eurasian basins. Water masses in the upper 500 m are the Atlantic Subarctic Upper Water (ASUW) and Pacific Subarctic Upper Water (PSUW). Below 1,500 m are deep and abyssal waters and Arctic Bottom Water (ABW).
Surface circulation and ice drift in the Arctic Ocean reflect a clockwise gyre driven by the Beaufort Sea high pressure. On the Eurasian side, the Transpolar Drift Stream takes water and sea ice from the Eurasian coast to Fram Strait, where it forms the East Greenland Current. Atlantic water enters the Norwegian and Barents seas, sinks, and flows counterclockwise around the Arctic Basin. The Arctic has low surface salinity owing to the immense river runoff and fresh water entering this ocean via Bering Strait. Water temperatures have risen since the 1950s, especially in the entryways. The upper Barents Sea has warmed 4 °C since the 1970s.
Many diverse seas surround the Arctic Ocean. The rather shallow Barents Sea (1.4 million km2) has a main branch of the warm, salty North Atlantic Current that forms a Polar Front with cold water from the north. The Barents Sea is ice free year-round to about 75° N. The Kara Sea (888,000 km2) is part of the Siberian shelf, with nearly half of it having a depth less than 50 m. The sea is very cold and frozen for eight months of the year. Runoff from the Ob and Yenisei rivers causes very low salinity.
The Laptev Sea (660,000 km2) has a mean depth of approximately 50 m. It is a major region of winter sea ice growth and export. Runoff from the Lena and four other major rivers causes low salinity. Three channels connect the Laptev Sea to the East Siberian Sea (910,000 km2). The East Siberian Sea is nearly all continental shelf, with depths less than 50 m, and has sea ice cover for most of the year. The Chukchi Sea (approximately 600,000 km2) has a mean depth of 70 m. Pacific water enters via Bering Strait, forced by a slope in sea level. Northward flow is stronger in summer. Upwelling events of Atlantic Water occur in Barrow Canyon. A large winter polynya forms along the Alaskan coast. The Beaufort Sea (475,000 km2) has a mean depth of approximately 1,000 m. The continental shelf is narrow in the west but wide off the Mackenzie delta. The surface circulation is clockwise around the Beaufort gyre, but an eastward subsurface jet occurs along the shelf break. The upper water temperature is approximately −1.4 °C in summer and −1.8 °C in winter. The Lincoln Sea (64,000 km2) is covered all year with the thickest ice in the Arctic. There is a narrow, eastward undercurrent along the continental slope, suggesting a link to the undercurrent off Alaska.
The Greenland Sea (1.2 million km2) is linked to the Arctic Ocean via Fram Strait. It has an average depth of 1,450 m and features deep troughs. North-flowing North Atlantic Current (NAC) water sinks in the Arctic, and part of it returns south as the East Greenland Current (EGC), which transports Arctic pack ice south from October to August. The eastern part of the sea is home to the warm West Spitsbergen Current, which, with the NAC, forms a counterclockwise gyre. The EGC has a southward transport of 8.6 Sv with a net outflow of 2 Sv. At 75° N, a mooring showed a mean transport of 11 Sv in summer and 37 Sv in winter, with two thirds of this transport being wind driven and one third attributable to thermohaline circulation. Deep convection events were reported in the central Greenland Sea in the 1990s.
The Canadian Arctic archipelago (CAA) has three main straits with shallow sills that open to Baffin Bay. Total outflow of Arctic water through the CAA is approximately 1.5 Sv. Sea ice persists for much of the year in the north and west, whereas channels in the east and south usually clear by late summer.
Adjacent seas of the North Atlantic are Baffin Bay, Davis Strait, Labrador Sea, and Hudson Bay. Baffin Bay (approximately 890,000 km2) has an average depth of 760 m. The West Greenland Current flows northward before recirculating off northwest Greenland and joining Arctic water from Nares Strait to form the Baffin Current, which transports sea ice and Greenland icebergs southward to Newfoundland. Davis Strait links Baffin Bay to the Labrador Sea (840,000 km2). Two thirds of the sea is ice covered in winter. North Atlantic Deep Water forms by deep convection in about six out of ten winters and flows at depth to Antarctica. NADW is overlain by lighter Labrador Sea Water. Hudson Bay (approximately 1.2 million km2) is linked to the North Atlantic by Hudson Strait. It has an average depth of approximately 100 m and is ice covered from mid-December to early June. There is a slow cyclonic gyre in this area, and the large runoff makes it function as an estuary.
Marginal seas of the North Pacific are the Bering Sea and the Sea of Okhotsk. The former has an area of 2.3 million km2, with a shallow shelf in the east and north and a deep basin in the south and west. There is a cyclonic gyre with the south-flowing Kamchatka Current in the west. The Alaska stream of Pacific water enters through gaps in the Aleutian Island chain. Sea ice in November is advected south from polynyas and by mid-March covers one third to one half of the sea. The Sea of Okhotsk (approximately 1.6 million km2) has an average depth of 860 m, with a broad shelf in the north. There is a cyclonic gyre and southward current along Sakhalin Island. Runoff from the Amur River gives this sea low salinity, facilitating freezing from October–November to June.
Ocean warming and expansion caused 1.1 mm of annual sea level rise from 1992 to 2010. Over this period, Greenland produced 0.33 mm of sea level rise, Antarctica 0.27 mm, and terrestrial storage 0.38 mm. Glacier melt accounted for 0.86 mm of this amount for 1993–2008.
There are three categories of sea ice: first-year ice (FYI), multiyear ice (MYI), and (land)fast ice. The sea begins to freeze at around −1.8 °C with average ocean salinity. Only the layer above the pycnocline has to cool to enable freezing. Growing sea ice rejects brine, which drains out. In calm conditions, frazil crystals form on the surface and freeze into a sheet of nilas. Congelation ice grows on the base and the ice thickens, turning first gray and then white. Waves can lead to the formation of cakes of slush and pancakes that can be several meters in diameter in the interior of the pack. FYI can reach 1.5–2 m in thickness. The sea ice edge is defined by the 15 percent concentration line. The 100- to 200-km-wide marginal ice zone (MIZ) extends to concentrations of 80 percent.
Arctic ice is currently about 60 percent FYI and 40 percent MYI, a reversal from the pattern in the 1980s. In winter, ice divergence accounts for half as much volume change as ice growth. In summer, basal and lateral melt exceed that at the top surface.
Arctic ice circulates around the Beaufort gyre and moves from the Asian coast to Fram Strait, where it exits the basin. For 1978–2002, the annual ice area export was 866,000 km2 and the annual ice volume flux was 2,218 km3 (0.07 Sv). There has been a 0.45 m decrease in mean ice thickness in Fram Strait.
The Arctic Ocean has two oceanic circulation patterns that alternate every five to seven years. One is anticyclonic, where the Beaufort gyre expands and that in the Laptev Sea shrinks. The other is cyclonic, with the reverse tendency.
Cloud cover does not respond to sea ice loss in summer, but low clouds form over open water in autumn. The Laptev, Chukchi, and Beaufort seas have the largest cloud–sea ice covariance in summer–autumn; for the Barents–Kara Sea, the maximum occurs in winter.
Persistence of sea ice characteristics affects predictability. July–August sea ice area is correlated with that in September. Re-emergence of sea ice anomalies also occurs from the melt to the growth season, and vice versa; these phenomena are related to SST anomalies and sea ice thickness anomalies, respectively. Analysis of sea ice predictions has highlighted the role of “easy” and “difficult” years as reflecting climate versus weather. Predictions based on statistics and coupled models have shown the best performance in predicting sea ice extent.
Landfast ice comprises bottom fast ice in 2-m water depths as well as attached floating ice. Off northern Alaska, the latter extends to the 18-m isobaths. In parts of the Kara, Laptev, and East Siberian seas, fast ice extends hundreds of kilometers offshore. It provides a basis for ice massifs, each of which is associated with a flaw polynya.
Sea ice leads form in only approximately 1 percent of the ice area, but account for about half of the turbulent heat transfer to the atmosphere in winter. Lead densities are highest in early winter, decreasing 20 percent by April in the western Arctic. Preferred orientations are north–south in the Beaufort Sea and east–west in the East Siberian Sea. There is a strong correlation between divergence and lead density, and between shear and lead orientation. Leads show a rectilinear pattern, with an intersection angle of approximately 30°. MODIS IR data for January to April 2003–2015 over the entire Arctic have shown that the main lead activity occurs in the MIZ of Fram Strait and the Barents Sea.
Major fracture zones are in the Beaufort Sea and along flaw polynyas off Siberia.
Energy balance data for Arctic pack ice show small turbulent fluxes and about 60 W m−2 available for melting in June–July.
Arctic sea ice trends from a data compilation from 1850 to the present show that a decrease in sea ice extent from the 1920s to the 1940s was apparent in summer, whereas the recent decline has affected sea ice extent in all seasons. This decline appears unprecedented for the past 1,450 years. The East Greenland Sea had more sea ice during the 1960s and 1970s, coincident with the Great Salinity Anomaly. The CAA and the Chukchi Sea have seen major recent declines in summer and autumn ice cover.
The passive microwave record since 1978 is the most consistent source of information on sea ice extent in the Arctic. In September, the average decrease in ice extent was 13 percent decade−1, compared with less than 3 percent decade−1 in March. September 2012 saw a record minimum of 4.3 million km2 of Arctic sea ice.
Internal variability seems to account for approximately 40 percent of the September sea ice decline since 1979. It involves Pacific SSTs and atmospheric circulation over Greenland. September open water fractions in the Pacific and Atlantic sectors over 1979–2014 show breakpoints around 1988 and 2007. Increases in open water have been much greater in the Pacific sector than in the Atlantic, with the largest increases occurring in the Chukchi and East Siberian seas. Since 2003, there has been increased Atlantic water inflow into the eastern Eurasian Basin, with oceanic heat reducing winter sea ice growth.
Links exist between the tropical Madden Julian Oscillation (MJO) and mid-tropospheric AO circulation and both winter and summer sea ice concentrations.
Ice in the Eurasian Arctic has generally decreased since 1933, but experienced a partial recovery from the mid-1950s to the mid-1980s.
Arctic ice thickness for the 1990s shows a 40 percent thinning compared with 1958–1976 data. Winter ice thinned by 1.75 m from 1980 to 2008–2009. MYI coverage decreased 42 percent from 2005 to 2008, and thinned by 0.6 m during this span. Landfast ice measurements in the CAA from 1957 to 2014 have shown thinning of 0.25 m (except at Resolute) that reflects changes in snow cover.
A record of Fram Strait ice export for 1935–2014 shows no long-term trend in ice export, except for an increase in the 2000s due to stronger northerly winds.
Antarctic ice is nearly all seasonal. Its mean extent occurs between February and September and ranges from 3 to 18.5 million km2. Two decades of observations show a mean thickness value of 0.87 m. Wind and wave interaction lead to rafting, while the snow load depresses the ice, flooding the surface, which freezes on top. Ice drifts northward and forms a circumpolar frontal ice zone, up to 1,000 km wide, that is older, thicker, and rougher. It protects the younger ice in the interior of the pack.
Ship’s logs from 1897 to 1917 suggest ice extent comparable to now, except in the Weddell Sea where ice coverage reached 1.0–1.7° farther north. Whaling records indicate the summer ice edge in the 1970s and 1980s was 2.4° south of that in the 1930s. Satellite data for 1979–2014 show increasing ice cover in the summer. Increases in the western Ross Sea have been offset by decreases in the Bellingshausen–Amundsen seas. Changes in the tropical Pacific since 2000 and deepening of the Amundsen Sea low have affected the extent of ice in the Ross Sea. An improved passive microwave record (November 1978–December 2015) shows strong negative correlations between sea ice extent and one-month lagged surface temperature in the growth season, and similar positive correlations in the melt season.
Polynyas are areas of largely ice-free water surrounded by sea ice or by ice and a coastline. They may form by upwelling warm water (sensible-heat polynya) or by winds driving ice away from a coast or other barrier. New ice forms in the open water, releasing latent heat, and is pushed downwind until it is consolidated onto the pack ice. Sixteen Arctic polynyas were shown to generate about 144 km3 of sea ice. Thirteen Antarctic polynyas produced 1,410 km3 of sea ice and accounted for approximately 10 percent of southern hemisphere sea ice production. The winter coastal polynyas of Antarctica cover approximately 245,000 km2. The Ross Sea polynya is approximately 20,000 km2 and is forced by strong southerly winds. The largest polynyas (Ross Sea and Cape Darnley) also form dense Antarctic Bottom Water.
The northwestern Okhotsk polynya has the highest ice production in the northern hemisphere. The North Water (NOW) polynya in northern Baffin Bay (85,000 km2 in spring) is maintained by both latent heat and sensible heat from upwelling Atlantic water off West Greenland. Polynyas in the Laptev Sea have been shown to have a mean duration of 14 days, with twelve events noted between November and June.
In 1974–1976, an open ocean polynya was observed by ESMR close to Maud Rise. It has not re-formed since then.
1. Describe the main features of the Antarctic Circumpolar Current.
2. Compare the Ross and Weddell seas.
3. What are the major differences between the Arctic Ocean and the Southern Ocean?
4. How does river runoff affect the Arctic Ocean?
5. Compare the Arctic seas north of Eurasia with those north of North America and Greenland.
6. Compare the East Greenland Sea and Baffin Bay–Davis Strait.
7. What are the components of global sea level rise?
8. Compare sea ice growth processes in the Arctic and the Antarctic.
9. How has Arctic sea ice changed since the 1980s?
10. How do landfast ice and pack ice differ?
11. What roles do sea ice leads and polynyas play in the ice balance and energy balance?
12. Describe the characteristics of polynyas and what maintains them.