T H E C H A N GIN G A R CTI C OC EAN | SPE C I A L ISS U E ON T H E INT ERN AT I ONA L POLAR YEAR 20072009 A Synthesis of Exchanges $rough the Main Oceanic Gateways to the Arctic Ocean

BY AGNIESZK A BESZCZYNSK A M Ö L L E R , R E B ECC A A . W O O D G AT E , C R A I G L E E , H U M FREY MELLING, A N D M I C HAE L KARCH ER

Deployment of a deep mooring with an Aanderaa current meter in Fram Strait during R/V Polarstern cruise ARKXIII in 2008. Photo courtesy of A. Beszczynska-Möller

82 Oceanography | Vol.24, No.3 ABSTRACT. In recent decades, the Arctic Ocean has changed dramatically. mass, liquid freshwater, and heat through Exchanges through the main oceanic gateways indicate two main processes of global the main Arctic gateways and the esti- A Synthesis of Exchanges climatic importance—poleward oceanic heat "ux into the Arctic Ocean and export mates derived from these observations. of freshwater toward the North Atlantic. Since the 1990s, in particular during the Year-round moorings, almost uninter- $rough the Main Oceanic Gateways International Polar Year (2007–2009), extensive observational e%orts were undertaken rupted since 1990, measure Paci!c to monitor volume, heat, and freshwater "uxes between the Arctic Ocean and the in"ow through Bering Strait. Since subpolar seas on scales from daily to multiyear. &is paper reviews present-day 2004, the program has been part of the to the Arctic Ocean estimates of oceanic "uxes and reports on technological advances and existing US National Oceanic and Atmospheric challenges in measuring exchanges through the main oceanic gateways to the Arctic. Administration (NOAA)-led Russian- American Long-Term Census of the INTRODUCTION by Paci!c in"ow through Bering Strait Arctic (RUSALCA) program. Since 2007, &e Arctic Ocean is tightly connected appears to play an important, triggering a high-resolution mooring array has to the global ocean system via water role in recent sea ice decline in the been deployed in the strait as part of a mass exchanges with the Paci!c and western Arctic (Woodgate et al., 2010). US National Science Foundation (NSF) Atlantic Oceans through several main Climate simulations indicate that International Polar Year (IPY) e%ort, oceanic gateways: Bering Strait, the increased freshwater out"ow from the and since 2010, as part of the NSF Arctic Canadian Arctic Archipelago, Davis Arctic may reduce convection in the Observing Network program. Strait, Fram Strait, and the entrance deepwater formation regions of the In the Atlantic in"ow sector, the to the Barents Sea (Figure 1). Changes North Atlantic, potentially in"uencing !rst concerted but regional e%ort to to subarctic seas thus can a%ect the the Atlantic meridional overturning monitor oceanic "uxes through the Arctic Ocean. For example, changing circulation (MOC; Wadley and Bigg, Nordic Seas started in the VEINS project poleward oceanic heat "ux can 2002; Vellinga et al., 2008). Arctic fresh- (Variability of Exchanges in Northern potentially impact perennial sea ice. water, stored in the upper few hundred Seas, 1997–2000) and was followed by In recent decades, a decline in sea meters, is delivered to the North Atlantic the international pan-Arctic ASOF study ice coverage, thickness, and volume in the form of liquid freshwater or sea (Arctic-Subarctic Ocean Fluxes, 2003– has been observed (e.g., Kwok et al., ice through Fram Strait or the Canadian 2006). &e subsequent DAMOCLES 2009), with the largest sea ice retreat in Arctic Archipelago. Details of how this project (Developing Arctic Modelling summer 2007 (e.g., Stroeve et al., 2008). "ux in"uences MOC are still poorly and Observing Capabilities for Long- A large-scale shi$ to warmer condi- known, although it is believed the term Environment Studies, 2006–2010) tions in the Arctic has been observed impacts on convection and Atlantic aimed to understand the e%ects of since the early 1990s, including warmer deepwater formation depend on the climate change in the Arctic (in partic- Atlantic in"ow (Holliday et al., 2008; exit route of the water from the Arctic ular, its potential to reduce perennial Schauer et al., 2008); sporadic warming (Koenigk et al., 2007). In recent years, sea ice). DAMOCLES was the European of the Paci!c Water in"ow, especially substantial changes have been observed contribution to IPY and provided the in 2007 (Woodgate et al., 2010); intru- in the storage and distribution of fresh- largest ever e%ort to assemble simul- sions of anomalously warm water over water in the Arctic Ocean, such as a shi$ taneous observations of the full Arctic Arctic shelves (Dmitrenko et al., 2010), in the position of the transpolar dri$ atmosphere-ice-ocean system, including along the Arctic continental margin (Rigor et al., 2002), changes in Arctic- oceanic exchanges through the Arctic- (Dmitrenko et al., 2008), into the central wide freshwater content (e.g., Polyakov Atlantic gateways. Arctic (Polyakov et al., 2005), and in the et al., 2008; Rabe et al., 2011), and trajec- To the west of Greenland, measure- Canada Basin (McLaughlin et al., 2009); tory shi$s in the Arctic-subarctic fresh- ments in Davis Strait carried out as and a large increase of ocean-atmosphere water cycle (Peterson et al., 2006). part of the NSF Freshwater Initiative heat "uxes due to sea ice losses (Kurtz In this paper, we review present-day (2004–2007) were followed by inten- et al., 2011). In addition, heat carried observational e%orts to quantify "uxes of sive observational e%orts undertaken

Oceanography | September 2011 83 1 50 E ° N N W

° ° 180° 150° 66°N 70 74 78°N BS 11 0°W Bering t Strait trai USA M’ClureS Victoria 0 Island Chukchi E Sea 120° -10 100° 60 W ° Beaufort N -25

Barro Sea Canada -50

w Laptev Russia St Canadian Sea -75 ra Basin LS it Canadian Arctic -100 90°W Lancaster

Devo Archipelago -250

n ian

Is s Ellesmere Island n 90°E . d Ell si -500 Sound es Sound m Jone ereI s. Eura Ba Kara Smith Sound Islan Nares Sea s -1000 Strait 80°W y Baffin NS Ba -1500 Greenland Island ffin Baffin d FS -2000 Ba Sval- an Barents DS bard Fram Strait Sea it -3000 ra Davis St Greenl BSO 60°N 91°W n Strait Greenland CS iga Sea -4000 Island 60 Card n °E -5000 90°W evo ate D llG He Norwegian Sea Norway -6000 89°W HG

Iceland Ellesmere Island

20 40 77 20 60 30 ’ ’ ° ’ ° N N W 60°N

0° ° E 30° Figure 1. Locations of moored arrays in the main Arctic oceanic gateways, overlaid on a bathymetry map (in color; bathymetry data from the International Bathymetric Chart of the Arctic Ocean [http://www.ngdc.noaa.gov/mgg/bathymetry/arctic; Jakobsson et al., 2003]). A schematic of the main pathways of the Atlantic (red) and Pacific (yellow) derived waters is shown (after Jones et al., 2001; Rudels et al., 2004; Aksenov et al., 2010). $e solid red line depicts the Atlantic Water (AW) inflow, and the dashed red line represents circulation of the modified AW in the Arctic Ocean. $e solid yellow line shows the Pacific inflow, and the dashed yellow line indicates the freshwater outflow (a mixture of Pacific-, river-, and ice-melt-origin waters) from the Arctic Ocean. Black bars represent the moored arrays: FS = Fram Strait. DS = Davis Strait. BSO = Barents Sea Opening. BS = Bering Strait. NS =Nares Strait. LS = Lancaster Sound. CS = Cardigan Strait. HG = Hell Gate. A magnified map of the Canadian Arctic Archipelago is shown on the upper left panel, and details of Cardigan Strait and Hell Gate are zoomed on the lower left panel.

from 2007 to present as part of the NSF continue northward as two separate BSO) as the North Cape Current. Only IPY and Arctic Observing Network branches of the Norwegian Atlantic the upper part of the AW layer can travel programs. Monitoring of oceanic "uxes Current (NwAC). &e western baroclinic through this 450 m deep passage into through three main gateways of the (i.e., varying with depth) "ow follows the Barents Sea. &e remaining AW Canadian Archipelago from 1998–2006 the Arctic Front through the Norwegian "ow continues along the shelf slope took place as part of the international and Greenland Seas and ultimately into Fram Strait. ASOF program and the NSF Freshwater converges with the West Spitsbergen From long-term measurements at Initiative, and was continued during Current (WSC) in Fram Strait (Orvik the BSO moored array, the North Cape IPY under the Canadian Archipelago and Niiler, 2002; Figure 1). &e eastern Current was found to depend on local &rough"ow Study (CATS). barotropic (i.e., similar at all depths) wind forcing (Ingvaldsen et al., 2004a), branch, the Norwegian Atlantic Slope and the AW in"ow exhibited strong BARENTS SEA OPENING Current (NwASC), is topographically seasonal variability with a maximum in AND BARENTS SEA trapped over the Norwegian continental winter (Ingvaldsen et al., 2004b). &e STRONGEST MODIFICATIONS slope and provides the main source of mean in"ow of warm, salty AW through OF THE ATLANTIC INFLOW Atlantic Water (AW) and heat to the BSO was estimated by Skagseth et al. &e Atlantic in"ow across the Barents Sea (Orvik and Skagseth, 2005). (2008) as 1.8 Sv, with large interannual Greenland-Scotland Ridge brings 8.5 Sv On the western slope of the Barents Sea, variations (between 0.8 and 2.9 Sv) (1 Sv = 106 m3 s–1) of warm, saline waters part of AW enters a 400 km wide passage and updated to 2 Sv for 1997–2007 by into the Nordic Seas (Østerhus et al., between Bear Island and the North Cape Smedsrud et al. (2010). However, a frac- 2005). More than half of those waters (known as the Barents Sea Opening, tion of the Atlantic in"ow recirculates

84 Oceanography | Vol.24, No.3 along a short pathway a$er passing Barents Sea and ~ 4 TW in its northern Sea exit took place under the Norwegian through BSO and returns as a 2–2.5°C part. Most of the AW leaving the Barents IPY Project BIAC (Bipolar Atlantic colder out"ow (leaving at 2°C) through Sea into the Arctic Ocean via St. Anna &ermohaline Circulation), when !ve the northern part of the opening Trough is already cooled to temperatures moorings were deployed for one year (Rudels, 1987; Skagseth et al., 2008). below 0°C (Schauer et al., 2002). (2007–2008) between Novaya Zemlya An additional in"ow enters the Barents In contrast to extensive observations and Franz Jozef Land. Sea along the Norwegian coast as the in the western Barents Sea, only a few warm, fresh Norwegian Coastal Current long-term measurements and sporadi- FRAM STRAIT (NCC), carrying a mixture of Atlantic cally repeated hydrographic sections are A DEEP AND WIDE PASSAGE and coastal waters. In a recent study, available in the eastern and northern FOR EXCHANGE OF ATLANTIC Skagseth et al. (2011) estimated the NCC Barents Sea. Using the only data avail- AND ARCTIC ORIGIN WATERS volume transport as 2.6 Sv (half of it able (i.e., from a one-year moored array Fram Strait provides a wide, deep (sill con!ned to the slope) with 86 coastal- between Novaya Zemlaya and Franz depth of 2600 m) conduit both for water components being 1.8 Sv of that Josef Land in 1991–1992), Schauer et al. Atlantic Water in"ow into the Arctic (including 0.6 Sv in the slope branch). (2002) estimated a net volume "ux into Ocean and export of polar waters and While no signi!cant trend was found the Arctic Ocean of 1.5 Sv, half of it sea ice out of the Arctic and into the in NwASC volume "ux, Mork and consisting of dense, cold (temperature of Nordic Seas (Figure 1). While AW in"ow Skagseth (2010) reported a positive trend about –0.5°C) bottom waters. Recently, is concentrated in the West Spitsbergen in the temperature, corresponding to a Gammelsrød et al. (2009) combined the Current (WSC) above the eastern shelf linear increase of 0.5°C in 1992–2009. In same observations with two numerical slope, the out"ow (both liquid and sea BSO, a temperature increase of 1°C over models and updated the earlier estimate ice) from the Arctic Ocean is carried in the period 1997–2006 (to values above to 2 ± 0.6 Sv. A modeling data assimila- the western part of the strait, both by 6°C), reported by Skagseth et al. (2008), tion study (using a variational inverse) of the East Greenland Current (EGC) and was accompanied by a positive trend in monthly hydrographic and atmospheric by waters above the Greenland shelf. In volume "ux of 0.1 Sv yr–1. climatologies (Panteleev et al., 2006) comparison to the strong atmospheric AW circulation in the Barents Sea provided estimates of BSO in"ow of modi!cation of the AW branch crossing consists of multiple narrow currents 3.2 Sv, recirculation in the northern the shallow Barents Sea, AW in Fram following bottom topography. During its BSO of 1.5 Sv, and the out"ows between Strait is spread over deeper layers, and transit through the shallow (mean depth Novaya Zemlya and Franz Jozef Land thus preserves its warm core, losing less of 230 m) sea, AW density increases and through Kara Gate (the passage heat to the atmosphere. signi!cantly due to atmospheric cooling between Novaya Zemla and the Russian A$er the detour of part of the and ice formation, despite dilution coast) of 1.1 and 0.7 Sv, respectively, Atlantic in"ow into BSO, AW continues by precipitation, sea ice melting, and similar to available observations. Recent northward along the continental shelf an admixture of fresh surface and observational activities in the Barents slope as the West Spitsbergen Current coastal waters. Warm AW in"ow keeps large areas of the Barents Sea ice-free Agnieszka Beszczynska-Möller ([email protected]) is Physical year-round, and, as a result, nearly all Oceanographer, Alfred Wegener Institute for Polar and Marine Research, Bremerhaven, oceanic heat entering through BSO Germany. Rebecca A. Woodgate is Associate Professor, Applied Physics Laboratory, is lost to the atmosphere. Smedsrud University of Washington, Seattle, WA, USA. Craig Lee is Associate Professor, Applied et al. (2010) showed that as much as Physics Laboratory, University of Washington, Seattle, WA, USA. Humfrey Melling is 67 TW (92% of the heat input of 73 TW Senior Researcher, Fisheries and Oceans Canada, Institute of Ocean Sciences, Sidney, BC, through BSO in the total in"ow of Canada. Michael Karcher is Physical Oceanographer, Alfred Wegener Institute for Polar 3.2 Sv; 1 TW = 1012 W) is lost to the and Marine Research, Bremerhaven, Germany, and also a researcher at Ocean Atmosphere atmosphere already in the southern Systems, Hamburg, Germany.

Oceanography | September 2011 85 (WSC). Farther to the north, WSC (CTD) sensors distributed at !ve levels: Fram Strait, Schauer and Beszczynska- splits into three separate branches in the subsurface layer (~ 50 m), in Möller (2009) obtained a mean value of (see Schauer et al., 2004, for discus- the AW layer (~ 250 m), at the AW 36 ± 6 TW for the 1997–2009 period. In sion). &e easternmost branch follows lower boundary (~ 750 m, since 2002), recent years, substantial warming was the upper shelf slope north of Svalbard in the deep layer (~ 1500 m), and also observed in the AW recirculating (Svalbard branch) and constitutes the above the bottom. directly in Fram Strait. &is warm water, main source of AW to the Arctic Ocean &e 13 years of measurements provide joining the southward "ow on the Boundary Current, the topographically estimates of the magnitude and vari- western side of the strait, may potentially steered pan-Arctic circulation of AW ability of oceanic "uxes through Fram modify water masses leaving the Arctic (Woodgate et al., 2001). &e second Strait over a wide range of time scales Ocean toward the North Atlantic. branch circumvents the northwestern (from hourly to multiyear), and suggest &e East Greenland Current in rim of the Yermak Plateau (Yermak a mean net volume "ux to the south of western Fram Strait carries fresh, cold Plateau branch), continues eastward, 2 ± 2.7 Sv (Schauer et al., 2008). &ere Arctic waters and sea ice into the Nordic and ultimately rejoins the Svalbard are large uncertainties in estimates of seas and toward the North Atlantic. In Branch, thus also entering the Arctic. volume "ux through the entire strait the Arctic freshwater budget studied &e westernmost branch of WSC mostly because the spatial variability of the "ow by Serreze et al. (2006), roughly equal carries waters from the western NwAC with a Rossby radius of ~ 10 km is only contributions from Fram Strait sea ice extension and recirculates directly in poorly resolved by the mooring spacing and liquid freshwater "uxes deliver about the northern Fram Strait, and exits in the deep part of the strait. (A Rossby half of the freshwater exported from the again to the south. radius is a theoretically derived scale Arctic Ocean. While the contributions Transport and properties of AW typical of oceanic horizontal variability; of Paci!c, meteoric (precipitation and carried by WSC west and north it is determined from the ocean depth riverine), and ice meltwaters to the EGC of Svalbard have been extensively studied and the density structure.) AW in"ow in freshwater out"ow vary signi!cantly since the beginning of the previous WSC, where mooring coverage is denser, between years (e.g., Dodd et al., 2009; century. Before the mid 1990s, most of is better constrained. On interannual Rabe et al., 2009) with an alternative the data was obtained by summer hydro- time scales between 1997 and 2010, exit route for these waters being via the graphic surveys or sparse moorings that a nearly constant volume "ux in the Canadian Arctic Archipelago, nearly the did not extend across the entire strait WSC Svalbard branch (long-term mean entire sea ice "ux from the Arctic passes (Aagaard et al., 1973; Foldvik et al., 1988). of 1.8 ± 0.1 Sv northward, including through Fram Strait (Kwok, 2009). Since 1997, long-term year-round obser- 1.3 ± 0.1 Sv of AW warmer than 2°C Long-term, year-round measurements vations have been carried out through and showing no seasonal variability) of EGC freshwater "ux (freshwater an extensive moored array in northern was accompanied by a highly variable being a relative measure, referenced to Fram Strait (Schauer et al., 2004, 2008). volume "ux of between 2 and 6 Sv in the an arbitrary chosen salinity) have been &e joint German-Norwegian array Yermak Plateau branch (long-term mean carried out since 1997 at the Fram Strait of 16 deep moorings extends from the of 5 ± 0.4 Sv, strong seasonal variability, moored array. &e preliminary estimates shelf edge west of Svalbard to the east 1–2 Sv of warm AW). of Holfort et al. (2008) of 32 mSv of Greenland shelf slope with a few shallow Signi!cant AW warming has been freshwater transport in EGC and an moorings on the western shelf dedicated observed at the moored array in Fram equal amount over the East Greenland to monitoring freshwater "uxes in EGC Strait since year-round measurements shelf for the period 1997–2005 were (Figure 2). &e moored array covers the began. Schauer et al. (2008) report updated by de Steur et al. (2009), using 300 km wide section with a spatial reso- distinctive temperature increases in two data from 1997–2008. Both studies yield lution from ~ 10 km at the upper shelf periods (1998–2000 and 2003–2006), similar liquid freshwater "ux through slope to ~ 30 km in the deep area. &e with a net increase over the last decade Fram Strait (64 and 66 mSv), but the array is instrumented with current meters of 1°C. Using a stream-tube approach to de Steur et al. (2009) study yields a and conductivity-temperature-depth estimate the oceanic heat "ux through higher estimate of freshwater transport

86 Oceanography | Vol.24, No.3 Steel top floatation 79° EGC

Benthos transponder d ARGOS transmitter and flashing light 78°

T

A

AC CTP sensor Greenlan S SE (MicroCat or Seacat) valbar

77 ° d

Float package

ADCP 76° Fram Strait WSC upward-looking RAFOS oceanographic tomographic STAR Seaglider sound mooring transceiver receiver with RAFOS source

° 0° 5° 10° 15

-5 ° Float package -20° -15° -10° Current meter Aanderaa RCM8 EGC WSC 0 200 400

Float package Anderaa current meter 600 TS sensor (SBEMicrocat/Seacat) 800 Current meter Pressure Inverted Echo Sounder 1000 Aanderaa RCM11 ) profiling top 1200 (m withCTD package 1400 long range acousticmodem 1600 depth Float package long range acoustic data transfer 1800 Current meter underwater profiling winch Aanderaa RCM8 upward-looking ADCP 2000 Double releases passive acoustic recorder 2200

Seaglider 2400 Anchor weight 2600 050 100 150 200 250 300 350 400 450 distance (km) Figure 2. Schematic of the Fram Strait Observatory. An example of a deep mooring is shown in the left panel: CTP = conductivity, temperature, pressure sensor. ADCP = acoustic Doppler current profiler. RCM = recording current meter. $e upper right panel shows positions of the oceano- graphic moorings, tomography moorings (STAR = Simple Tomographic Acoustic Receiver), and RAFOS (SOund Fixing And Ranging, spelled backward) sound sources in Fram Strait. $e lower right panel presents a distribution of instruments in the array (the view looking north). EGC = East Greenland Current. WSC = West Spitsbergen Current.

in EGC (with a long-term annual mean the freshwater "ux remained nearly total "ux, while an increase of salinity in of 40.4 mSv in 1997–2008) and a lower constant. Rabe et al. (2009) estimated the upper layer due to sea ice formation contribution from the shelf of 25.6 mSv. the average liquid freshwater transport was equivalent to a decrease of the total &e same study also shows that the total in summer (in the EGC and over the freshwater "ux by 60–80 mSv. EGC volume "ux more than doubled shelf) to be a net 80 mSv, based on since 2001 (from a minimum of 4 Sv in hydrographic data combined with BERING STRAIT 2002 to a maximum of nearly 10 Sv in current meter measurements. &ey also THE ONLY ENTRANCE 2007), mostly due to stronger transport derived the composition of freshwater FOR PACIFIC WATERS in deep layers (below 700 m). In the "ux from salinity and δ18O (oxygen In contrast to the broad entrance points upper layer above 700 m, the volume isotope) measurements. Meteoric fresh- for Atlantic Water into the Arctic, Paci!c transport increased only slightly, and water contributed 130–160 mSv to the Water has only one oceanic entrance, via

Oceanography | September 2011 87 the shallow (~ 50 m), narrow (~ 85 km waters, and provide part of the strong coherent across the section, con!rming wide) Bering Strait, which is divided into upper-ocean strati!cation that typi!es older ideas of a dominantly barotropic two channels by the Diomede Islands the Arctic and (in winter) isolates Arctic and homogeneous "ow throughout the and the US-Russian border (Figure 1). sea ice from subsurface oceanic heat. In strait except in the boundary currents, Sporadic observations in the strait have summer, however, Paci!c waters carry a hypothesis developed from analysis of been made for centuries (see Coachman signi!cant amounts of heat into the prior mooring and hydrographic data and Aagaard, 1966, for a review). Year- Arctic—although highly variable from and assumptions about the large-scale round moorings have been deployed in year to year, the heat "ux (relative to forcings of the "ow (see Woodgate et al., Bering Strait almost continuously since the freezing temperature of seawater) 2005b, for discussion). &is coherence 1990 (Roach et al., 1995; Woodgate is enough to melt 1–2 million km2 of in velocity is in stark contrast to the lack et al., 2006), with a one-year gap 1 m thick ice (Woodgate et al., 2010). of coherence across the much wider (summer 1996–1997). Certainly, the pathways of Paci!c Water Fram Strait, re"ecting the compara- Typically, moorings have been into the Arctic are clearly re"ected in the tive smallness of the strait (~ 85 km deployed in the center of each of the structure of the sea ice edge, implying versus ~ 350 km for Fram Strait) and channels (other than in several years in that Paci!c Water heat acts as a trigger the fact that the main Bering Strait the 1990s when access was limited to for the onset of western Arctic sea ice forcings have length scales larger than US waters) and at a site ~ 55 km north melt, especially in 2007 when the Bering the strait. Speci!cally, the "ow through of the strait proper, hypothesized to Strait oceanic heat "ux was over twice Bering Strait is generally attributed to give a useful average of the mean "ow that in 2001 (Woodgate et al., 2010). a northward pressure head between properties (Woodgate et al., 2007). &e above estimates are based on the Paci!c and the Arctic, modulated Due to the dangers of trawling by ice near-bottom data, using a simple by local winds (see Woodgate et al., keels, these moorings typically only approach (based on satellite data and 2005b, for discussion). While it has measured up to ~ 10 m above bottom summer hydrographic stations) to been noted that the estimated size of the (i.e., ~ 40 m below the surface). Since include the substantial contributions pressure-head forcing is comparable to 2001, an extra mooring has been added from the Alaskan Coastal Current and the di%erence in steric heights between to measure a seasonal, warm, fresh strati!cation in the upper water column. the basins (i.e., di%erences arising due to coastal current on the US side of the While only ~ 0.1 Sv in volume, the di%erent densities of water in the basins; strait, the Alaskan Coastal Current. Alaskan Coastal Current is believed to e.g., Stigebrandt, 1984), other causes of Since 2007, an array consisting of eight contribute one-third of the heat and one- the pressure head are also being consid- moorings, including upper-layer sensors quarter of the freshwater "uxes through ered. Recent results indicate signi!cant (Iscats, also called ICECATs, described the strait (Woodgate and Aagaard, 2005; temporal variability in this term, below in the new technology section), Woodgate et al., 2006). &e ~ 0.1 Sv suggesting that both far-!eld and local has been deployed in the strait to quan- contribution of another coastal current, forcings determine the net "ow through tify cross-strait structure and elucidate the Siberian Coastal Current (present the strait (Woodgate et al., 2005b, 2010; the physical driving mechanisms of the only in some summers on the Russian Aagaard et al., 2006). "ow (Woodgate, 2008). side of the strait and entering Bering Although only ~ 0.8 Sv in the long- Strait from the north) is still mostly CANADIAN POLAR SHELF term annual mean (Roach et al., 1995), unquanti!ed (Woodgate et al., 2005b). A NETWORK OF CHANNELS the Bering Strait through"ow is a large &e denser array of moorings (seven &e other half of the freshwater out"ow oceanic source of nutrients for the across the strait proper, with velocity from the Arctic Ocean takes routes Arctic (Walsh et al., 1989), bringing in pro!les and upper layer sensors at !ve west of Greenland across the Canadian one-third of the freshwater entering of the moorings) also allows the !rst polar shelf, a patchwork of islands and the Arctic Ocean (Serreze et al., 2006). year-round quantitative description of straits, basins, and sills (Figure 1). All Being generally fresher, Paci!c waters lie the "ow structure in the strait. &ese the Canadian Arctic &rough"ow (CAT) higher in the water column than Atlantic new data show the "ow to be strongly passes through four gateways, with Nares

88 Oceanography | Vol.24, No.3 Strait, Lancaster Sound, and Cardigan near the surface, and freshwater "ux is Ice thickness is measured by sonar. &e Strait (plus Hell Gate) being most impor- even more strongly surface intensi!ed spatial structure of the "ow was revealed tant (Melling et al., 2008). (Melling, 2000). Small-scale structure at high resolution by Doppler sonar from Pathways across the Canadian shelf necessitates high resolution in measure- the icebreaker USCGC Healy in 2003 are deeper than those for Paci!c Arctic ment (5 km), whereas surface intensi!ca- (Münchow et al., 2006). in"ow across the ~ 50 m deep Bering tion requires ingenuity in measuring Cardigan Strait and Hell Gate are and Chukchi shelves; the depths of above 30 m where ice risk is high. 8 and 5 km wide and 200 and 120 m obstructing sills vary, with the shallowest Year-round capability to measure CAT deep, respectively. Only Cardigan Strait being 80 m (Penny Strait) and the deepest has emerged only recently through tech- is instrumented, with three Doppler 220 m (Nares Strait). &e gateways are nological innovation (i.e., I-CYCLERS, sonars providing ice dri$ and current hydrodynamically wide (two to 10 times greater than the Rossby radius), despite being geometrically narrow (15–60 km). &is characteristic permits currents in ALTHOUGH MUCH OF THE RECENT IPY DATA both directions simultaneously, with IS STILL BEING ANALYZED, RESULTS ALREADY Arctic out"ow favoring the surface on the SHOW THE BENEFITS OF NEW TECHNOLOGIES, right (looking southeast) and the coun- ESPECIALLY WITH RESPECT TO SAMPLING ter"ow at greater depth on the le$. Each “ one of these gateways resembles Fram NEAR THE ICEOCEAN INTERFACE. Strait in miniature, with similarly compli- cated recirculation and mixing. &e Canadian shelf is the place where a tidal signal arriving from the discussed in the technology section below 30 m. Salinity is measured at west, weakened during travel around below; e.g., Fowler et al., 2004). the seabed, but 2 m s–1 tides preclude” the Arctic, meets a stronger tidal wave Synchronous observations in all gateways shallower measurements. &e section arriving directly from the Atlantic via were !rst acquired in 2003, measuring in Lancaster Sound is over 60 km Ba(n Bay (Kowalik and Proshutinsky, current, temperature, salinity, ice dri$ wide and 200 m deep. Four sites were 1994). Disparity in the amplitudes and and thickness, sea level, and wind (with instrumented during 2004–2005 phases of the waves creates strong tidal di%erent spatial and temporal resolutions (13 km spacing), but otherwise only one currents within the gateways, strong in the di%erent channels). or two sites are maintained. Here again, turbulence, and enhanced frictional &e section with a moored array in Doppler sonar cannot measure above damping of CAT. Also, the high terrain Nares Strait is 38 km wide and 300 m 30 m, but the moored ICYCLER pro!ler of the Arctic islands creates preferred deep (Figure 1b). Mooring observations (discussed below) deploys a salinity pathways for air"ow, which acceler- (deployed since 2003) at 5–7 km spacing sensor toward the ice daily at one loca- ates within the gateways, resulting in resolve the Rossby radius, except where tion (Fowler et al., 2004). &e detailed very strong along-strait winds at times moorings have been lost to icebergs. structure of "ow through Lancaster (Samelson et al., 2006). &e moored current measuring device Sound is not known. &e measurement of CAT faces chal- (a Doppler sonar) provides data above Volume and freshwater "uxes through lenges: remoteness, heavy sea ice that 40 m depth only if ice is present and then Nares Strait (and through other straits becomes land-fast in winter, strong only at the surface—fortunately, ice is with several moorings, for example, Fram currents, and an unreliable geomagnetic prevalent here. Salinity is measured at Strait, BSO, Davis Strait) are routinely reference direction because it occurs !ve levels on moorings that are drawn calculated from interpolated current and close to the north magnetic pole. Much down greatly by the tide. Salinity above salinity !elds. In Lancaster Sound, where of the volume "ux is carried by baroclinic 35 m is not known, except for the near- the cross sections of current and salinity "ow in narrow currents concentrated zero salinity layer (viz. ice) at the surface. are not routinely measured, "uxes are

Oceanography | September 2011 89 estimated from a single mooring using Volume "uxes through Cardigan export from the Arctic, with over half empirical scaling factors (Prinsenberg Strait and Hell Gate are –0.2 and –0.1 Sv, the freshwater export, and slightly more and Hamilton, 2005). Because there including surface current via ice tracking of the liquid freshwater export, exiting are few data available to constrain "ux (2000–2002; Melling et al., 2008). Given through Davis Strait. Arctic waters from estimates for the upper 30–40 m of the estimates are simplistic, derived from a the CAA "ow southward along Ba(n ocean, values are subject to unknown single sonar per strait without knowledge Island as the broad, surface-intensi!ed biases. For this reason, published "ux of spatial structure of the "ow. Moreover, Ba(n Island Current (Tang et al., 2004; values (Table 1) for Nares and Cardigan because of very strong tidal currents, it Cuny et al., 2005). On the eastern side of Straits do not include transport within is not technologically feasible to measure Davis Strait, the northward "ow consists the topmost 40 m. salinity over depth. &erefore, freshwater of the fresh West Greenland Current &ere are three data sources for "ux via these passages is not known. of Arctic origin above the shelf and Nares Strait "uxes. &e !rst is ship- &e program in Lancaster Sound the warm, salty West Greenland Slope based acoustic Doppler current pro!ler dates from 1998. Data from four sites Current of North Atlantic origin occu- (ADCP)/CTD snapshots from 2003, during 2004–2005 have been used to pying the slope. &ese in"owing waters yielding –0.9 Sv for volume and –31 mSv calibrate "uxes derived from fewer sites circulate cyclonically (anticlockwise) (relative to 34.8) for freshwater "ux, in other years. Surface-layer salinity in Ba(n Bay and join the southward with lower-bound estimates for surface- has been measured at one site in some out"ow in the western Davis Strait. layer transports included (Münchow years, but the salinity and current in Arctic waters undergo transformations et al., 2007). Two other studies based on the top 30 m have generally been esti- before passing through Davis Strait as long-term moored observations do not mated. &e average volume "ux during a result of terrestrial inputs from Ba(n include "uxes above 35 m depth. In the 1998–2006, including the surface, was Island and Greenland, an annual average !rst of these studies, current measure- –0.7 Sv (± 0.3 Sv interannual variation) heat loss of roughly 70 W m–2, brine and ments by Doppler sonar during 2003– and the freshwater "ux –48 ± 15 mSv freshwater inputs from sea ice formation 2006 yield a three-year average volume (Prinsenberg and Hamilton, 2005; and melting, and mixing with underlying "ux of –0.57 ± 0.1 Sv for "uxes below Melling et al., 2008). &e seasonal cycle warmer, more saline Atlantic waters 35 m depth (Münchow and Melling, is strong and variations appear to be (Rudels, 1986). Davis Strait thus provides 2008); seasonal variation is ± 50%. A driven by conditions in the Beaufort a strategic location for characterizing second estimate for the same period Gyre (Prinsenberg et al., 2009). the integrated CAA out"ow a$er trans- is based upon measured salinity and At the time of writing, June 2011, formation during its transit through temperature pro!les via the geostrophic instruments continue to record data in all Ba(n Bay and just prior to its entry into method (which assumes a balance three gateways, but these data will only the Labrador Sea. between the Coriolis force and oceanic become available for scienti!c interpreta- Although scienti!cally convenient, pressure gradients; Rabe et al., 2010). tion when the instruments are retrieved. Davis Strait presents severe observa- &e calculated "uxes, –0.47 ± 0.05 Sv tional challenges. As in Fram Strait and for volume and –20 ± 3 mSv for liquid DAVIS STRAIT in CAA, Davis Strait is dynamically freshwater, represent the instrumented A WIDE GATEWAY TO THE wide (i.e., much larger than the Rossby part of the strait only (61% of the NORTH ATLANTIC radius), and thus the "ow through the width and depth below 35 m). An &e narrow channels of the Canadian strait admits small-scale eddies and !la- unpublished study (Humfrey Melling, Arctic Archipelago (CAA) empty ments that must be resolved to produce Department of Fisheries and Oceans, into Ba(n Bay, ultimately funneling accurate "ux estimates. Sea ice cover Canada, pers. comm., 2011) of ocean- the Arctic out"ow from the western and icebergs with deep keels make it surface velocity via ice tracking suggests side of Greenland through the wide risky to measure in the upper 100 m that an additional 0.25 Sv of seawater (360 km), deep (650 m) Davis Strait (where the majority of the freshwater "ux may be transported within the (Figure 1). Fram and Davis Straits transport occurs) as well as over the neglected upper layer. account for 84% of the total freshwater broad (100 km), shallow (< 100 m)

90 Oceanography | Vol.24, No.3 Table 1. Summary of up-to-date estimates of fluxes through the critical oceanic gateways and their variability on different time scales, compiled from available sources. Values are positive for Arctic inflow.

Width Net volume flux (Sv) Variability Liquid freshwater Heat flux Gateway (max. depth) (mean for years) (Sv) flux (mSv) (TW=1012 W) Barents Sea ~ 350 km 2.01,2 55 to 601 501,2,* to 701,** interannual 0.8 to 2.92 Opening (500 m) (1997–2007) (1997 –2007) (1997–2007) NZ-FJL# ~ 400 km Barents Sea (400 m) 1.53 to 2.04 seasonal 0.6 to 2.63 43,* Exit SAT# (1991–1992) seasonal ± 0.64 ~ 200 km (600 m) 36 ± 69,** ~ 300 km, –2.0 ± 2.75 –80 to –667,8 monthly –8.4 to –0.26 (1997–2009) with shelf (1997 –2007) (1997–2008) Fram Strait interannual –4.7 to –0.36 2917 ~ 500 km –1.717 –6517 (2002–2008) (snapshots (2,600 m) (snapshots 1980–2005) (snapshots 1980–2005) 1980–2005) daily –2 to 312 ~ 85 km, 0.8 ± 0.2 monthly 0.4 to 1.312 80 ± 2014 10 to 2011,*** Bering Strait (50 m) (1990–2007)10,11,14 interannual 0.6 to 112 (1999–2005) (1998–2007) (uncertainties of 25%) –0.915 –31 ± 1015 (2003 snapshot, Not applicable (2003 snapshot, Not calculated 0–30 m estimated) 0–30 m estimated) –0.57 ± 0.116 ± 50% (2003–2006, neglects ~ 40 km (monthly to Not measured Not calculated Nares Strait 0–35 m, which prob- (220 m) interannual periods) ably adds ~ 0.25 Sv) –0.47 ± 0.0521 –20 ± 321 ± 50% (2003–2006, neglects (2003–2006, neglects (monthly to Not calculated 0–35 m which probably 0–35 m and ice flux, interannual periods) adds ~0.25 Sv) only 61% of width) Cardigan –0.313 ± 50% 8 km (180 m)/ Strait/Hell (2000–2002, (monthly to Not measured Not measured 4 km (125 m) Gate includes 0–30 m) interannual periods) Lancaster ~ 40 km –0.718 seasonal –1.3 to 018 –48 ± 1518 4 × 10–7 18,* Sound (125 m) (1998–2006) interannual –1.0 to –0.418 –2.6 ± 1.019 –92 ± 3419 18 ± 1719,* ~ 320 km (1987–1990) (1987–1990) (1987–1990) Davis Strait Not available (1,040 m) –2.3 ± 0.720 –116 ± 4120 20 ± 920,* (2004–2005) (2004–2005) (2004–2005) 1Smedsrud et al. (2010); 2Skagseth et al. (2008); 3Schauer et al. (2002); 4Gammelsrød et al. (2009); 5Schauer et al. (2008); 6Unpublished data of author Beszczynska-Möller; 7de Steur et al. (2009); 8Rabe et al. (2009); 9Schauer and Beszczynska-Möller (2009); 10Roach et al. (1995); 11Woodgate et al. (2010); 12Woodgate et al. (2005a) 13Melling et al. (2008); 14Woodgate et al. (2006); 15Münchow et al. (2007); 16Münchow and Melling (2008); 17Rudels et al. (2008); 18Prinsenberg et al. (2009); 19Cuny et al. (2005); 20Curry et al. (2011); 21Rabe et al. (2010) *heat flux calculated with reference temperature –0.1°C **heat flux for closed volume budget (reference temperature arbitrary) ***heat flux referenced to freezing temperature #NZ-FJL = a passage between Novaya Zemlya and Franz Jozef Land; SAT = Saint Anna Trough

Oceanography | September 2011 91 West Greenland shelf. &e Davis Strait 2004; Cuny et al., 2005), yielding and investigating processes to explain observing system addresses these chal- annual mean volume "ux estimates of the observed variability. lenges by combining a 14-element –3.3 to –2.6 (± 1.2) Sv and liquid fresh- Recent studies (e.g., Holland et al., moored array, which provides spatially water "uxes of –120 to –92 (± 34) mSv. 2008) show that northward penetra- coarse measurements of velocity, In contrast, Curry et al. (2011) use tion of warm Atlantic waters along the temperature, salinity, and ice dra$, the improved coverage and resolu- Greenland coast may accelerate the with an extensive marine chemistry tion provided by the current Davis motion of outlet glaciers by melting program, which provides measurements Strait observing system to estimate the and breakup of the "oating tongues that integrate broadly over time and annual mean volume (–2.3 ± 0.7 Sv) that resist seaward motion. Increased space, and new technologies developed and freshwater (–116 ± 41 mSv) "uxes northward heat transport through as part of the recent Davis Strait e%orts from 2004–2005. A$er normalizing for Davis Strait could thus accelerate (described in the technology section di%erences in coverage, the 1987–1990 Greenland ice cap melting, driving below; Figure 3). and 2004–2005 "ux estimates show sea level rise and injecting additional Davis Strait has been the site of the no changes that exceed the uncertain- freshwater into sensitive regions of the least scienti!c activity of all the major ties. Curry et al. (2011) also present subpolar North Atlantic. Arctic gateways. &e current e%ort, volume and freshwater budgets for started in 2004, builds on measurements Ba(n Bay, !nding that they close to HIGHLIGHTS OF NEW taken during 1987–1990 by a six-element within the somewhat large uncertain- TECHNOLOGIES IN THE moored array (Ross, 1992) that neglected ties in the constituent estimates, with MAIN GATEWAYS the upper 150 m and the shallow shelves. imbalances of 26% (volume) and 4% &e primary technological innovations &ree di%erent studies have employed (freshwater) of the net 2004–2005 Davis for measuring the "uxes of volume di%erent extrapolation schemes to Strait "uxes. Ongoing e%orts focus and freshwater through Arctic oceanic estimate "uxes from the 1987–1990 on calculating robust "ux estimates gateways have been (a) establishing data (Loder et al., 1998; Tang et al., for subsequent years (2005–present) ways of measuring salinity near the

Ba!n Island Greenland Length ~ 320 km 0 Iceberg ULS expendable CT ADCP & CT 200 RAFOS receiver current & CT 400 RAFOS source

600

Davis Strait 800 Depth (m)

1000

Depth ~ 1,040 m 1200 (Not to scale)

0 50 100 150 200 250 300 350 Distance (km) Figure 3. Schematic of the moored array in Davis Strait, looking north. ULS = upward-looking sonar. CT = conductivity and temperature sensor. ADCP = acoustic Doppler current profiler.

92 Oceanography | Vol.24, No.3 ocean surface in the ice-hazard zone commanded. Gliders and (b) deducing how to practically provide the ability to measure current direction so near to the resolve hydrographic north magnetic pole. structures (such as fronts, Several methods have been developed eddies, and !laments) at to acquire oceanographic data near the the roughly 10 km scale, ocean surface, close to the sea ice. In and to measure salinity Nares Strait, temperature and salinity near the ice-ocean inter- sensors have been placed on highly face. Sustained support compliant submerged moorings that are from the NSF O(ce passively dragged deeper, away from the of Polar Programs and ice, when the current is strong and the the US O(ce of Naval risk of ice impact is high. Loss to moving Research led to the ice is thereby minimized. In Lancaster successful development Sound, a novel ICYCLER has been used. and implementation of &is instrument incorporates a small Seagliders capable of "oat that carries a CTD and hazard- extended operation in avoidance sonar close to the surface ice-covered waters in while a larger mid-water "oat (containing Davis Strait since 2006. a controller, a power-minimizing winch Seagliders navigate on a system, and data storage) moves away regional scale (hundreds from the surface (Fowler et al., 2004). A of kilometers), deter- third under-ice measurement package, mining their location by the ICECAT, uses a (possibly) sacri!cial measuring sound travel temperature-salinity sensor package that time (viz. distance) from Figure 4. Applied Physics Laboratory-University of Washington transmits its data via inductive coupling an array of underwater team members launch a Seaglider from R/V Knorr into Davis to a recording module at a deeper (safer) sound sources. Because Strait in 2008. Photo courtesy of Craig Lee depth. &e connecting line is su(ciently under-ice gliders operate weak that it breaks if the near-surface without human interven- "oat is snagged and dragged by ice. &is tion for extended periods (weeks to in the upper 1,000 m layer. &e acoustic system has been used successfully in both months), they incorporate enhanced system for under-ice navigation, success- Bering Strait and Davis Strait, with some autonomy, including algorithms for navi- fully implemented in Davis Strait, is (anticipated) losses. gation, emergency response (if lost or currently being tested in Fram Strait An additional new technology— experiencing system failure), and identi- to permit gliders’ access to the ice- autonomous Seagliders—has been !cation and exploitation of ice leads for covered western part. used to gather hydrographic cross surfacing to telemeter data back to shore. &e challenge of reference current sections repeatedly over many months During a 2009 mission that included direction arises from the weakness and in Davis Strait (Figure 4). Seagliders are 51 days under the ice, a glider made variability of the horizontal geomag- underwater vehicles that glide without repeated sections across ice-covered netic !eld component at high latitudes. propellers from the surface to as deep as Davis Strait while !nding and exploiting &is problem was addressed by two 1,000 m while collecting information on eight leads (in 33 attempts) for commu- approaches: (1) the addition of a sensitive temperature, salinity, dissolved oxygen, nication through the ice. compass to the Doppler sonar package and chlorophyll. When they return to Seagliders are also used in Fram Strait and continuous monitoring of local the ocean surface, they transmit their to complement mooring observations by magnetic declination or (2) with installa- data via satellite and can be remotely providing high-resolution measurements tion of the Doppler sonar on a torsionally

Oceanography | September 2011 93 rigid mooring at the seabed (data are Atmospheric "uxes must also be strait, from freezing in winter to 10°C recorded in the instrument’s coordinate included in order to understand the heat at the surface in summer (Woodgate reference, and the true heading is estab- and freshwater budgets for the Arctic et al., 2005a), and the high day-to-day lished a posteriori from analysis of the Ocean. Serreze et al. (2006) achieved variability in "ow direction and water tidal oscillations of current). a balanced freshwater budget for the properties (Woodgate et al., 2005b) In recent years, acoustic tomog- Arctic Ocean by combining avail- necessitate year-round measurements. raphy has been used for instantaneous able observations with model results, In all critical gateways, with the measurements over long distances in the but they recognized that this balance possible exception of Bering Strait wide and deep Fram Strait. A$er a single was likely fortuitous, that uncertain- where there is also a strong ~ 40 m deep track acoustic thermometry experiment ties were still large, and that the vari- coastal current, the bulk of freshwater with acoustic transmissions between ability of freshwater "uxes remains is carried in a thin (order 10 m) surface a single source and a receiver array, highly undetermined. layer. Measurements in this layer (made carried out in 2008–2009, an acoustic With awareness of these challenges, possible, in part, by the new technologies tomography array consisting of three it is likely that better insight may be described above) are critical—including tomographic sound sources, one receiver gained from examining changes in increases in freshwater estimates in array, and passive listening systems oceanic "uxes through each gateway and this layer in some straits by 20–60% has operated in Fram Strait under the assessing their impacts on the Arctic (respectively, Lancaster Sound [Melling ACOBAR (Acoustic Technology for Ocean than from attempting to establish et al., 2008] and Davis Strait [Curry Observing the Interior of the Arctic a property "ux balance. Because instru- et al., 2011]). Near-surface access is thus Ocean) project since 2010. Acoustic menting all straits with moorings spaced essential for meaningful assessment of travel times along measured tracks at the Rossby radius is impractical, freshwater "uxes, and earlier estimates provide integrated measures of ocean e%orts must focus on (1) understanding from the literature need to be revised heat content and currents that will be the physics of the system, (2) estab- in this context. used for assimilation into numerical lishing the uncertainties and biases Another di(culty in estimating heat models, in combination with data from linked to under-resolving measurement "uxes into the Arctic Ocean arises from oceanographic moorings, gliders, and arrays, and (3) developing alternative the physical concept of oceanic heat satellites (Sagen et al., 2010). approaches for characterizing the "uxes. transport (see Schauer and Beszczynska- Hydrographic sections can resolve the Möller, 2009, for discussion). In general, UPTODATE ESTIMATES OF spatial scales of motion and provide esti- each gateway carries a nonzero volume OCEANIC FLUXES TO AND mates of geostrophic "uxes, but they are "ux. &us, to determine how much FROM THE ARCTIC OCEAN scarce and available mainly in summer. heat is delivered to the Arctic through a Despite extensive e%orts to estimate In an attempt to balance geostrophic gateway, it is necessary to ascertain the oceanic "uxes through the critical gate- volume "uxes to and from the Arctic temperatures of all those waters exiting ways, the goal of balanced seawater and Ocean, Rudels et al. (2008) used summer the Arctic (possibly through di%erent freshwater budgets is still far from being hydrography and estimated net volume gateways). Useful progress can be made achieved. Using available estimates of transports in Fram and Davis Straits to in speci!c cases. For example, Paci!c long-term means of the "uxes (Table 1), be smaller than mooring observations, waters are believed to exit the Arctic the budget for seawater volume can although agreeing within observational with temperatures close to freezing, so be balanced to within uncertainties; uncertainties (Rudels et al.: Fram 1.7 Sv, this value can be used as a reference for however, the uncertainties in the Fram Davis 1.4 Sv; mooring-based transports: the heat "ux of Paci!c in"ow (Woodgate, and Davis Strait "uxes are larger than the Fram 2 Sv and Davis 2.3 Sv). However, et al., 2010). However, the general total "uxes through other gateways. &is summer sections are likely not repre- conclusion is that oceanic heat transport problem also contaminates attempts to sentative of the whole year. A good to the Arctic Ocean can be estimated quantify freshwater "uxes because they example is Bering Strait, where the large only when all of the in"ows and out"ows are strongly dependent on volume "uxes. seasonal cycle in water properties in the are taken into account.

94 Oceanography | Vol.24, No.3 NUMERICAL SIMULATIONS properties of in"ows into the Arctic between CAA and Fram Strait still exists COMPLEMENTING Ocean via the Barents Sea Opening are (Gerdes et al., 2008). Jahn et al. (2010) OBSERVATIONAL EFFORTS greatly modi!ed on their roughly two- suggest that the dominant freshwater Given that many of the observational year-long transit across the Barents Sea. source for the western CAA is the Paci!c e%orts are regional in nature, basin-scale Model simulations may assist in under- and runo% from North America, while models are needed to provide connec- standing the details of these conversion in Fram Strait the freshwater is mainly tivity to the rest of the Arctic/subarctic processes, which are a consequence of composed of Eurasian runo% and Paci!c- system. Ideally, the combination of variable atmospheric forcing and strong derived water. A recent modeling study modeling and observations allows us surface heat and salt "uxes (e.g., Harms by Houssain and Herbaut (2011) explains to validate model performance in some et al., 2005; Aksenov et al., 2010). the links between freshwater export regions while interpolating between data On the Paci!c side, there are few through the Canadian Archipelago points in other (data-sparse) regions, recent dedicated Bering Strait through- and surface atmospheric forcing in the while dedicated sensitivity experiments "ow modeling e%orts, perhaps because western Arctic (wind stress curl) and in allow us to test hypotheses and suggest of the complexity of open boundary the Labrador Sea (air-sea "ux anomalies). causal relationships (Proshutinsky et al., conditions required for a regional model, Improved knowledge of the vari- 2011, in this issue). or because the strait’s small size means it ability in the export of freshwater and One major challenge for modeling is poorly represented in global models. the properties of the over"ows are e%orts is to simulate and understand A notable exception is a data assimila- important to better understand past and interannual variability in the Arctic and tion study of the Chukchi Sea (Panteleev future changes in the MOC (Holland how it relates to in"ows and out"ows et al., 2010), which combines a year of et al., 2007). Model studies indicate, for through the gateways. In the early mooring data into a dynamically consis- example, that the transport of water to 1990s, intense warming was observed tent model solution, and considers the the North Atlantic may substantially in the Eurasian and Makarov basins force balances within the strait. in"uence the MOC (e.g., Mauritzen and (Quadfasel et al., 1991). Model studies Model experiments also have contrib- Hakkinen, 1997; Holland et al., 2001), (e.g., Häkkinen and Geiger, 2000; uted to a better insight into the dynamics as do changes in the routes of freshwater Karcher et al., 2003) attributed the of freshwater export through the gate- out of the Arctic (Koenigk et al., 2007; observed warming to an increased AW ways of the Arctic, linking large-scale Rennermalm et al., 2007). volume in"ow and reduced heat loss Arctic Ocean changes with the freshwater to the warmer atmosphere. Since the "ux to subpolar waters. Haak et al. (2003) SUMMARY AND OUTLOOK second half of the 1990s, more recent found that large freshwater and sea ice &is paper reviews our current state of in"ows of anomalously warm water export events through the Canadian knowledge of the "ow through gate- into the Arctic Ocean have been docu- Arctic Archipelago and Fram Strait, ways into and out of the Arctic Ocean. mented (Schauer et al., 2008; Skagseth which led to the Great Salinity Anomalies Although much of the recent IPY data is et al., 2008). Karcher et al. (2008, 2011) (GSA) of the 1970s, 1980s, and 1990s still being analyzed, results already show point out that, both in observations and (when large freshenings were observed the bene!ts of new technologies, espe- model results, the anomalously warm in the subarctic seas; Belkin et al., 1998), cially with respect to sampling near the in"ows were also characterized by nega- had their source in the Arctic. Karcher ice-ocean interface. One next challenge is tive density anomalies. Karcher et al. et al. (2005) found that due to the change to design a sustainable Arctic Observing (2011) suggest that these anomalies, of the Arctic Ocean’s thermohaline Network for the Arctic gateways. What though reduced in magnitude, survive structure in the early 1990s, a large are the most e(cient approaches for the long (decadal) passage through volume of freshwater exited through quantifying, with known uncertainties, the Arctic Ocean, exit via Fram Strait, CAA and Fram Strait to enter the North oceanic "uxes of volume, heat, and fresh- and go on to alter the properties of the Atlantic deepwater production regions. water? A decadal-scale observing system Denmark Strait over"ow. Uncertainty with respect to what deter- will likely operate under severe resource In contrast to Fram Strait, the water mines the partition of freshwater export constraints, forcing di(cult choices.

Oceanography | September 2011 95 What approach should be used to opti- maneuvering, a variety of new technolo- from the European Union under the mize such a system? gies are gradually providing moored 5th FP project ASOF-N (contract number In many of the straits, "ow and water access to the upper water column and EVK2-CT-200200139), the 6th FP properties vary on the Rossby radius the ice-ocean interface. project DAMOCLES (contract number (usually ~ 10 km or less). For a sustained, Wider use could also be made of 018509GOCE) and the 7th FP project decadal-scale e%ort, it is impractical observational data by incorporating in ACOBAR. &e US National Science and too costly to instrument the straits situ and satellite data in high-quality Foundation supported CL and the Davis with moorings at this spacing. Statistical computer models of the regions Strait project under grants ARC0632231 relationships have been exploited to infer (i.e., data assimilation e%orts such as and ARC0633885. HM acknowledges "uxes from measurements made at a that of Panteleev et al., 2010). &ese !nancial and in-kind support from small number of critical sites. However, techniques are still being developed. the Canadian Federal Programme such empirical models rely on stationary Such a system, employing mooring for the International Polar Year (IPY statistics. Will such relationships remain observations, acoustic tomography, 2006-SR1-CC-135), from Fisheries and static despite the dramatic changes gliders, and "oats (see Figure 2) together Oceans Canada and the US National observed in the Arctic and subpolar seas? with an eddy-resolving numerical Science Foundation (grants 1022843 and Because the use of such empirical models model, was proposed for Fram Strait 0230254). &e success of !eld work is a may mask signals of Arctic change, the during the European DAMOCLES study. tribute to the diligence of captains, crews, straits would have to be instrumented Implementation of this system started and technicians during numerous cruises regularly at su(cient resolution to last year under the subsequent European devoted to mooring operations in the re-assess their validity. Only by under- ACOBAR program (Sagen et al., 2010). Arctic oceanic gateways. standing the physics controlling the Fundamentally, however, optimization through"ows and by using this under- of a decadal-scale, sustained observing REFERENCES standing to develop physical (rather than system for characterizing the Arctic Aagaard, K., C. Darnall, and P. Greisman. 1973. Year-long measurements in the statistical) models of the through"ow can gateways will !rst require a careful, Greenland-Spitsbergen passage. Deep-Sea useful and a%ordable long-term moni- quantitative assessment of goals. What Research Part I 20:743–746, http://dx.doi. org/10.1016/0011-7471(73)90090-9. toring systems be developed. Physical improvements in estimates of "uxes Aagaard, K., T.J. Weingartner, S.L. Danielson, understanding must therefore be a are required and can be made? To what R.A. Woodgate, G.C. Johnson, and T.E. Whitledge. 2006. Some controls on "ow primary goal of current research. accuracy must they be quanti!ed in and salinity in Bering Strait. Geophysical New technologies and approaches order to resolve long-term changes in the Research Letters 33, L19602, http://dx.doi. could be exploited to improve existing ocean and climate system of the Arctic? org/10.1029/2006GL026612. Aksenov, Y., S. Bacon, A. Coward, A. Nurser. arrays (Lee et al., 2010). For example, in What biological and biogeochemical 2010. &e North Atlantic in"ow to the Arctic the deep straits (Fram and Davis Straits), parameters should also be studied, and Ocean: High-resolution model study. Journal of Marine Systems 79(1–2):1–22, http://dx.doi. autonomous gliders can provide access at what spatial and temporal scales? org/10.1016/ j.jmarsys.2009.05.003. to the region near the ice-ocean interface Careful consideration of our goals will Belkin, I.M, S. Levitus, J. Antonov, S.-A. Malmberg. 1998. “Great Salinity Anomalies” in the and resolve the horizontal Rossby radius, provide guidance for the di(cult deci- North Atlantic. Progress In Oceanography although at coarse temporal resolution sions that lie ahead. 41(1):1–68, http://dx.doi.org/10.1016/ (7–10 days). Combining glider data S0079-6611(98)00015-9. Coachman, L.K., and K. Aagaard. 1966. On with moored measurements at coarse ACKNOWLEDGEMENTS the water exchange through Bering Strait. spatial resolution allows us to re!ne Comments from Laura de Steuer and one Limnology and Oceanography 11:44–59, http:// dx.doi.org/10.4319/lo.1966.11.1.0044. "ux estimates and better de!ne "ow anonymous reviewer helped to improve Cuny, J., P. Rhines, and R. Kwok. 2005. Davis Strait composition in water mass classes. For the manuscript. RW and the Bering volume, freshwater and heat "uxes. Deep-Sea Research Part I 52(3):519–542, http://dx.doi. broad, ice-covered shelves and shallow Sea project were supported by grants org/10.1016/ j.dsr.2004.10.006. straits (such as Bering Strait and CAA), NSF-ARC-0632154, NSF-ARC-0855748, where gliders cannot safely operate due and NOAA-RUSALCA. &e Fram Strait to high currents or lack of depth for observations and AB-M received support

96 Oceanography | Vol.24, No.3 Curry, B., C.M. Lee, and B. Petrie. 2011. Harms, I.H., C. Schrum, and K. Hatten. 2005. Jones, E.P. 2001. Circulation in the Arctic Ocean. Volume, freshwater, and heat "uxes through Numerical sensitivity studies on the variability Polar Research 20:139–146, http://dx.doi. Davis Strait, 2004–05. Journal of Physical of climate-relevant processes in the Barents Sea. org/10.1111/j.1751-8369.2001.tb00049.x. Oceanography 41:429–436, http://dx.doi. Journal of Geophysical Research 110, C06002, Karcher, M., A. Beszczynska-Möller, F. Kauker, org/10.1175/2010JPO4536.1. http://dx.doi.org/10.1029/2004JC002559. R. Gerdes, S. Heyen, B. Rudels, and U. Schauer. de Steur, L., E. Hansen, R. Gerdes, M. Karcher, Holfort, J., E. Hansen, S. Østerhus, S. Dye, 2011. Arctic Ocean warming and its conse- E. Fahrbach, and J. Holfort. 2009. Freshwater S. Jonsson, J. Meincke, J. Mortensen, and quences for the Denmark Strait over"ow. "uxes in the East Greenland Current: M. Meredith. 2008. Freshwater "uxes east of Journal of Geophysical Research 116, C02037, A decade of observations. Geophysical Greenland. Pp. 263–288 in Arctic-Subarctic http://dx.doi.org/10.1029/2010JC006265. Research Letters 36, L23611, http://dx.doi. Ocean Fluxes. R.R. Dickson, J. Meincke, and Karcher, M., R. Gerdes, and F. Kauker. 2008. Long- org/10.1029/2009GL041278. P. Rhines, eds, Springer, Dordrecht. term variability of Atlantic water in"ow to the Dmitrenko, I.A., S.A. Kirillov, L.B. Tremblay, Holland, M.M., C.M. Bitz, M. Eby, and A.J. Weaver. Northern Seas: Insights from model experi- D. Bauch, J.A. Hölemann, T. Krumpen, 2001. &e role of ice-ocean interactions in the ments. Pp. 111–130 in Arctic-Subarctic Ocean H. Kassens, C. Wegner, G. Heinemann, variability of the North Atlantic thermohaline Fluxes. R.R. Dickson, J. Meincke, and P. Rhines, and D. Schröder. 2010. Impact of the Arctic circulation. Journal of Climate 14:656–675, eds, Springer, Dordrecht. Ocean Atlantic water layer on Siberian http://dx.doi.org/10.1175/1520-0442(2001) Karcher, M., R. Gerdes, F. Kauker, and C. Köberle. shelf hydrography. Journal of Geophysical 014<0656: TROIOI>2.0.CO;2. 2003. Arctic warming: Evolution and spreading Research 115, C08010, http://dx.doi. Holland, M.M., J. Finnis, A.P. Barrett, and of the 1990s warm event in the Nordic seas org/10.1029/2009JC006020. M.C. Serreze. 2007. Projected changes in and the Arctic Ocean. Journal of Geophysical Dmitrenko, I.A., I.V. Polyakov, S.A. Kirillov, Arctic Ocean freshwater budgets. Journal of Research 108(C2), 3034, http://dx.doi. L.A. Timokhov, I.E. Frolov, V.T. Sokolov, Geophysical Research 112, G04S55, http:// org/10.1029/2001JC001265. H.L. Simmons, V.V. Ivanov, and D. Walsh. 2008. dx.doi.org/10.1029/2006JG000354. Karcher M., R. Gerdes, F. Kauker, C. Köberle, Toward a warmer Arctic Ocean: Spreading Holland, D.M., R.H. &omas, B. DeYoung, and I. Yashayev. 2005. Arctic Ocean change of the early 21st century Atlantic Water warm M.H. Ribergaard, and B. Nyberth. 2008. heralds North Atlantic freshening. Geophysical anomaly along the Eurasian Basin margins. Acceleration of Jakobshavn Isbrae trig- Research Letters 32, L21606, http://dx.doi. Journal of Geophysical Research 113, C05023, gered by warm subsurface ocean waters. org/10.1029/2005GL023861. http://dx.doi.org/10.1029/2007JC004158. Nature Geosciences 1:659–664, http://dx.doi. Koenigk, T., U. Mikolajewicz, H. Haak, and Dodd, P.A., K.J. Heywood, M.P. Meredith, org/10.1038/ngeo316. J. Jungclaus. 2007. Arctic freshwater export A.C. Naveira-Garabato, A.D. Marca, and Holliday, N.P., S.L. Hughes, S. Bacon, in the 20th and 21st centuries. Journal of K.K. Falkner. 2009. Sources and fate of fresh- A. Beszczynksa-Möller, B. Hansen, A. Lavin, Geophysical Research 112, G04S41, http:// water exported in the East Greenland Current. H. Loeng, K.A. Mork, S. Østerhus, T. Sherwin, dx.doi.org/10.1029/2006JG000274. Geophysical Research Letters 36, L19608, http:// and W. Walczowski. 2008. Reversal of the 1960s Kowalik, Z., and A.Y. Proshutinsky. 1994. &e dx.doi.org/10.1029/2009GL039663. to 1990s freshening trend in the Northeast Arctic Ocean tides. Pp. 137–158 in "e Polar Foldvik, A., K. Aagaard, and T. Törresen. 1988. On North Atlantic and Nordic Seas. Geophysical Oceans and "eir Role in Shaping the Global the velocity !eld of the East Greenland Current. Research Letters 35, L03614, http://dx.doi. Environment. Geophysical Monograph Deep-Sea Research Part I 35:1,335–1,354, http:// org/10.1029/2007GL032675. 85, O.M. Johannessen, R.D. Muench, and dx.doi.org/10.1016/0198-0149(88)90086-6. Houssais, M.-N., and C. Herbaut. 2011. J.E. Overland, eds, American Geophysical Fowler, G.A., G.R. Siddall, and S.J. Prinsenberg. Atmospheric forcing on the Canadian Arctic Union, Washington, DC. 2004. An energy-conserving oceanographic Archipelago freshwater out"ow and implica- Kurtz, N.T., T. Markus, S.L. Farrell, D.L. Worthen, pro!ler for use under mobile ice cover: tions for the Labrador Sea variability. Journal and L.N. Boisvert. 2011. Observations of ICYCLER. International Journal of O!shore and of Geophysical Research 116, C00D02, http:// recent Arctic sea ice volume loss and its Polar Engineering 14(3):176–181. dx.doi.org/10.1029/2010JC006323. impact on ocean-atmosphere energy exchange Gammelsrød, T., Ø. Leikvin, V. Lien, W.P. Budgell, Ingvaldsen, R.B., L. Asplin, and H. Loeng. and ice production. Journal of Geophysical H. Loeng, and W. Maslowski. 2009. Mass 2004a. Velocity !eld of the western entrance Research 116, C04015, http://dx.doi. and heat transports in the NE Barents Sea: to the Barents Sea. Journal of Geophysical org/10.1029/2010JC006235. Observations and models. Journal of Marine Research 109, C03021, http://dx.doi. Kwok, R. 2009. Out"ow of Arctic Ocean sea ice Systems 75:56–69. org/10.1029/2003JC001811. into the Greenland and Barents Seas: 1979– Gerdes, R., M. Karcher, C. Köberle. and K. Fieg. Ingvaldsen, R.B., L. Asplin, and H. Loeng. 2004b. 2007. Journal of Climate 22:2,438–2,457, http:// 2008. Simulating the long-term variability of &e seasonal cycle in the Atlantic transport to dx.doi.org/10.1175/2008JCLI2819.1. liquid freshwater export from the Arctic Ocean. the Barents Sea during the years 1997–2001. Kwok, R., G.F. Cunningham, M. Wensnahan, Pp. 405–425 in Arctic-Subarctic Ocean Fluxes. Continental Shelf Research 24(9):1,015–1,032, I. Rigor, H.J. Zwally, and D. Yi. 2009. &inning R.R. Dickson, J. Meincke, and P. Rhines, eds, http://dx.doi.org/10.1016/j.csr.2004.02.011. and volume loss of the Arctic Ocean sea ice Springer, Dordrecht. Jahn, A., L.B. Tremblay, R. Newton, M.M. Holland, cover: 2003–2008. Journal of Geophysical Haak, H., J. Jungclaus, U. Mikolajewicz, and L.A. Mysak, and I.A. Dmitrenko. 2010. Research 114, C07005, http://dx.doi. M. Latif. 2003. Formation and propaga- A tracer study of the Arctic Ocean’s liquid org/10.1029/2009JC005312. tion of great salinity anomalies. Geophysical freshwater export variability. Journal of Latif, M., C. Böning, J. Willebrand, A. Biastoch, Research Letters 30(9), 1473, http://dx.doi. Geophysical Research 115, C07015, http://dx.doi. J. Dengg, N. Keenlyside, G. Madec, and org/10.1029/2003GL017065. org/10.1029/2009JC005873. U. Schweckendiek. 2006. Is the thermo- Häkkinen, S., and C.A. Geiger. 2000. Simulated Jakobsson, M., A. Grantz, Y. Kristo%ersen, haline circulation changing? Journal of low-frequency modes of circulation in and R. Macnab. 2003. Physiographic Climate 19:4,631–4,637, http://dx.doi. the Arctic Ocean. Journal of Geophysical provinces of the Arctic Ocean. GSA org/10.1175/JCLI3876.1. Research 105(C3):6,549–6,564, http://dx.doi. Bulletin 115(12):1,443–1,455, http://dx.doi. Lee, C.M., H. Melling, H. Eicken, P. Schlosser, org/10.1029/2000JC900003. org/10.1130/B25216.1. J.-C. Gascard, A. Proshutinsky, E. Fahrbach, C. Maurtizen, J. Morison, and I. Polyakov.

Oceanography | September 2011 97 2010. Autonomous platforms in the Arctic Orvik, K.A., and P. Niiler. 2002. Major pathways of Proshutinsky, A., Y. Aksenov, J. Clement Observing Network. In Proceedings of Atlantic water in the northern North Atlantic Kinney, R. Gerdes, E. Golubeva, D. Holland, OceanObs ’09: Sustained Ocean Observations and Nordic Seas toward Arctic. Geophysical G. Holloway, A. Jahn, M. Johnson, and Information for Society, vol. 2. Venice, Research Letters 29, 1896, http://dx.doi. E. Popova, M. Steele, and E. Watanabe. Italy, September 21–25, 2009, J. Hall, org/10.1029/2002GL015002. 2011. Recent advances in Arctic Ocean D.E. Harrison, and D. Stammer, eds, ESA Orvik, K.A., and Ø. Skagseth. 2005. Heat "ux studies employing models from the Publication WPP-306. variations in the eastern Norwegian Atlantic Arctic Ocean Model Intercomparison Project. Loder, J., B. Petrie, and G. Gawarkiewicz. 1998. &e Current toward the Arctic from moored Oceanography 24(3):102–113, http://dx.doi. coastal ocean o% northeastern North America: instruments, 1995–2005. Geophysical org/10.5670/oceanog.2011.61. A large-scale view. Pp. 105–133 in "e Global Research Letters 32, L14610, http://dx.doi. Quadfasel, Q., A. Sy, D. Wells, and A. Tunik. 1991. Coastal Ocean: Regional Studies and Syntheses. org/10.1029/2005GL023487. Warming in the Arctic. Nature 350:385, http:// A. Robinson and K. Brink, eds, &e Sea, vol. 11, Østerhus, S., W.R. Turrell, S. Jónsson, and dx.doi.org/10.1038/350385a0. Harvard University Press. B. Hansen. 2005. Measured volume, Rabe, B., M. Karcher, U. Schauer, J.M. Toole, Mauritzen, C., and S. Hakkinen. 1997. In"uence heat, and salt "uxes from the Atlantic to R.A. Krish!eld, S. Pisarev, F. Kauker, R. Gerdes, of sea ice on the thermohaline circulation in the Arctic Mediterranean. Geophysical and T. Kikuchi. 2011. An assessment of Arctic the Arctic-North Atlantic Ocean. Geophysical Research Letters 32, L07603, http://dx.doi. Ocean freshwater content changes from the Research Letters 24:3,257–3,260, http://dx.doi. org/10.1029/2004GL022188. 1990s to the 2006–2008 period. Deep-Sea org/10.1029/97GL03192. Panteleev, G.G., D.A. Nechaev, and M. Ikeda. Research Part I 58:173–185, http://dx.doi. McLaughlin, F.A., E.C. Carmack, W.J. Williams, 2006. Reconstruction of summer Barents org/10.1016/ j.dsr.2010.12.002. S. Zimmermann, K. Shimada, and M. Itoh. Sea circulation from climatological data. Rabe, B., A. Münchow, H.L. Johnson, and 2009. Joint e%ects of boundary currents and Atmosphere-Ocean 44(2):111–132, http://dx.doi. H. Melling. 2010. Nares Strait hydrography thermohaline intrusions on the warming of org/10.3137/ao.440201. and salinity !eld from a 3-year moored array. Atlantic water in the Canada Basin, 1993–2007. Panteleev, G.G., D.A. Nechaev, A. Proshutinsky, Journal of Geophysical Research 115, C07010, Journal of Geophysical Research 114, C00A12, R. Woodgate, and J. Zhang. 2010. http://dx.doi.org/10.1029/2009JC005966. http://dx.doi.org/10.1029/2008JC005001. Reconstruction and analysis of the Chukchi Rabe, B., U. Schauer, A. Mackensen, M. Karcher, Melling, H. 2000. Exchanges of freshwater through Sea circulation in 1990–1991. Journal of E. Hansen, and A. Beszczynska-Möller. 2009. the shallow straits of the North American Geophysical Research 115, C08023, http://dx.doi. Freshwater components and transports in the Arctic. Pp. 479–502 in Freshwater Budget of the org/10.1029/2009JC005453. Fram Strait: Recent observations and changes Arctic Ocean. E.L. Lewis, ed., NATO/WCRP/ Peterson, B., J. McClelland, R. Curry, R.M. Holmes, since the late 1990s. Ocean Science 5:219–233, AOSB, Kluwer Academic Publishers, Boston. J.E. Walsh, and K. Aagaard. 2006. Trajectory http://dx.doi.org/10.5194/os-5-219-2009. Melling, H., T.A. Agnew, K.K. Falkner, shi$s in the Arctic and subarctic freshwater Rennermalm, A.K., E.F. Wood, A.J. Weaver, D.A. Greenberg, C.M. Lee, A. Münchow, cycle. Science 313:1,061, http://dx.doi. M. Eby, and S.J. Dery. 2007. Relative sensi- B. Petrie, S.J. Prinsenberg, R.M. Samelson, and org/10.1126/science.1122593. tivity of the Atlantic meridional overturning R.A. Woodgate. 2008. Fresh-water "uxes via Polyakov, I.V., V.A. Alexeev, G.I. Belchansky, circulation to river discharge into Hudson Bay Paci!c and Arctic out"ows across the Canadian I.A. Dmitrenko, V.V. Ivanov, S.A. Kirillov, and the Arctic Ocean. Journal of Geophysical polar shelf. Pp. 193–247 in Arctic-Subarctic A.A. Korablev, M. Steele, L.A. Timokhov, and Research 112, G04S48, http://dx.doi. Ocean Fluxes. R.R. Dickson, J. Meincke, and I. Yashayaev. 2008. Arctic Ocean freshwater org/10.1029/2006JG000330. P. Rhines, eds, Springer, Dordrecht. changes over the past 100 years and their Rigor, I.G., J.M. Wallace, and R.L. Colony. 2002. On Mork, K.A., and Ø. Skagseth. 2010. A quantita- causes. Journal of Climate 21:364–384, http:// the response of sea ice to the Arctic Oscillation. tive description of the Norwegian Atlantic dx.doi.org/10.1175/2007JCLI1748.1. Journal of Climate 15(18):2,546–2,663, http:// Current by combining altimetry and hydrog- Polyakov, I.V., A. Beszczynska, E.C. Carmack, dx.doi.org/10.1175/1520-0442(2002)015<2648: raphy. Ocean Science 6:901–911, http://dx.doi. I.A. Dmitrenko, E. Fahrbach, I.E. Frolov, ROSITT>2.0.CO;2. org/10.5194/os-6-901-2010. R. Gerdes, E. Hansen, J. Holfort, Roach, A., K. Aagaard, C. Pease, S. Salo, Münchow, A., and H. Melling. 2008. Ocean V.V. Ivanov, and others. 2005. One more T. Weingartner, V. Pavlov, and M. Kulakov. current observations from Nares Strait to step toward a warmer Arctic. Geophysical 1995. Direct measurements of trans- the west of Greenland: Interannual to tidal Research Letters 32, L17605, http://dx.doi. port and water properties through the variability and forcing. Journal of Marine org/10.1029/2005GL023740. Bering Strait. Journal of Geophysical Research 66(6):801–833. Prinsenberg, S.J., and J. Hamilton. 2005. Research 100(C9):18,443–18,457, http://dx.doi. Münchow, A., K.K. Falkner, and H. Melling. Monitoring the volume, freshwater and heat org/10.1029/95JC01673. 2007. Spatial continuity of measured seawater "uxes passing through Lancaster Sound in the Ross, C. 1992. Moored current meter measure- and tracer "uxes through Nares Strait, a Canadian Arctic Archipelago. Atmosphere- ments across the Davis Strait. Technical dynamically wide channel bordering the Ocean 43(1), http://dx.doi.org/10.3137/ Report 92/70. Northwest Atlantic Fisheries Canadian Archipelago. Journal of Marine ao.430101. Organization (NAFO) Research Document. Research 65(6):759–788, http://dx.doi.org/ Prinsenberg, S., J. Hamilton, I. Peterson, and Rudels, B. 1986. &e out"ow of polar water 10.1357/002224007784219048. R. Pettipas. 2009. Observing and interpreting through the Arctic Archipelago and the Münchow, A., H. Melling, and K.K. Falkner. the seasonal variability of the oceanographic oceanographic conditions in Ba(n Bay. 2006. An observational estimate of volume "uxes passing through Lancaster Sound of the Polar Research 4:161–180, http://dx.doi. and freshwater "ux leaving the Arctic Ocean Canadian Arctic Archipelago. Pp. 125–143 org/10.1111/j.1751-8369.1986.tb00528.x. through Nares Strait. Journal of Physical in In$uence of Climate Change on the Rudels, B. 1987. On the mass balance of the Polar Oceanography 36:2,025–2,041, http://dx.doi. Changing Arctic and Sub-Arctic Conditions. Ocean, with special emphasis on the Fram org/10.1175/JPO2962.1. NATO Science for Peace and Security Series Strait. Norsk Polarinstitutt Skri%er 188, 53 pp. C: Environmental Security, http://dx.doi. org/10.1007/978-1-4020-9460-6_10.

98 Oceanography | Vol.24, No.3 Rudels, B., E. Jones, U. Schauer, and P. Eriksson. Skagseth, Ø., T. Furevik, R. Ingvaldsen, H. Loeng, boundary current along the Eurasian slope 2004. Atlantic sources of the Arctic K.A. Mork, K.A. Orvik, and V. Ozhigin. 2008. and the adjacent Lomonosov Ridge: Water Ocean surface and halocline waters. Polar Volume and heat transports to the Arctic Ocean mass properties, transports and transforma- Research 23(2):181–208, http://dx.doi. via the Norwegian and Barents seas. Pp. 45–64 tions from moored instruments. Deep-Sea org/10.1111/j.1751-8369.2004.tb00007.x. in Arctic-Subarctic Ocean Fluxes. R.R. Dickson, Research Part I 48:1,757–1,792, http://dx.doi. Rudels, B., M. Marnela, and P. Eriksson. 2008. J. Meincke, and P. Rhines, eds, Springer, org/10.1016/S0967-0637(00)00091-1. Constraints on estimating mass, heat and Dordrecht. Woodgate, R.A., K. Aagaard, J.H. Swi$, freshwater transports in the Arctic Ocean: Smedsrud, L.H., R. Ingvaldsen, J.E.Ø. Nilsen, and W.M. Smethie, and K.K. Falkner. 2007. Atlantic An exercise. Pp. 315–341 in Arctic-Subarctic Ø. Skagseth. 2010. Heat in the Barents Sea: water circulation over the Mendeleev Ridge Ocean Fluxes. R.R. Dickson, J. Meincke, and Transport, storage, and surface "uxes. Ocean and Chukchi Borderland from thermohaline P. Rhines, eds, Springer, Dordrecht, http:// Science 6:219–234, http://dx.doi.org/10.5194/ intrusions and water mass properties. Journal of dx.doi.org/10.1007/978-1-4020-6774-7_14. os-6-219-2010. Geophysical Research 112, C02005, http://dx.doi. Sagen, H., S. Sandven, A. Beszczynska-Möller, Stigebrandt, A. 1984. &e North Paci!c: org/10.1029/2005JC003416. O. Boebel, T.F. Duda, L. Freitag, J.C. Gascard, A global estuary. Journal of Physical Woodgate R.A., K. Aagaard, and T.J. Weingartner. A. Gavrilov, C.M. Lee, D.K. Mellinger, and Oceanography 14:464–470, http://dx.doi. 2005a. Monthly temperature, salinity, and trans- others. 2010. Acoustic technologies for org/10.1175/1520-0485(1984)014<0464: port variability of the Bering Strait through"ow. observing the interior of the Arctic Ocean. TNPAGS>2.0.CO;2. Geophysical Research Letters 32, L04601, http:// Pp. 146–147 in Proceedings of the OceanObs´09: Stroeve, J., M. Serreze, S. Drobot, S. Gearheard, dx.doi.org/10.1029/2004GL021880. Sustained Ocean Observations and Information M. Holland, J. Maslanik, W. Meier, and Woodgate, R.A., K. Aagaard, and T.J. Weingartner. for Society Conference (Annex). Venice, Italy, T. Scambos. 2008. Arctic sea ice extent plum- 2005b. A year in the physical oceanography September 21–25, 2009, J. Hall, D.E. Harrison, mets in 2007. Eos, Transactions, American of the Chukchi Sea: Moored measurements and D. Stammer, eds, ESA Publication Geophysical Union 89(2), http://dx.doi. from autumn 1990–1991. Deep-Sea Research WPP-306, Abstract AC-5B-19. org/10.1029/2008EO020001. Part II 52(24–26), http://dx.doi.org/10.1016/ Samelson, R.M., T. Agnew, H. Melling, and Tang, C., C. Ross, T. Yao, B. Petrie, B. DeTracey, j.dsr2.2005.10.016. A. Münchow. 2006. Evidence for atmospheric and E. Dunlap. 2004. &e circulation, water Woodgate, R.A., K. Aagaard, and T.J. Weingartner. control of sea-ice motion through Nares Strait. masses and sea-ice of Ba(n Bay. Progress in 2006. Interannual changes in the Bering Geophysical Research Letters 33, L02506, http:// Oceanography 63(4):183–228, http://dx.doi. Strait "uxes of volume, heat and fresh- dx.doi.org/10.1029/2005GL025016. org/10.1016/ j.pocean.2004.09.005. water between 1991 and 2004. Geophysical Schauer, U., and A. Beszczynska-Möller. 2009. Vellinga, M., B. Dickson, and R. Curry. 2008. &e Research Letters 33, L15609, http://dx.doi. Problems with estimation and interpreta- changing view on how freshwater impacts the org/10.1029/2006GL026931. tion of oceanic heat transport: Conceptual Atlantic meridional overturning circulation. Woodgate, R.A., T. Weingartner, and R. Lindsay. remarks for the case of Fram Strait in the Arctic Pp. 289–314 in Arctic-Subarctic Ocean Fluxes. 2010. &e 2007 Bering Strait oceanic heat Ocean. Ocean Science 5:487–494, http://dx.doi. R.R. Dickson, J. Meincke, and P. Rhines, "ux and anomalous Arctic sea‐ice retreat. org/10.5194/os-5-487-2009. eds, Springer, Dordrecht, http://dx.doi. Geophysical Research Letters 37, L01602, http:// Schauer, U., A. Beszczynska-Möller, org/10.1007/978-1-4020-6774-7_13. dx.doi.org/10.1029/2009GL041621. W. Walczowski, E. Fahrbach, J. Piechura, and Wadley, M.R., and G.R. Bigg. 2002. Impact of E. Hansen. 2008. Variation of measured heat "ow through the Canadian Archipelago on "ow through the Fram Strait between 1997 the North Atlantic and Arctic thermohaline and 2006. Pp. 65–85 in Arctic-Subarctic Ocean circulation: An ocean modelling study. Fluxes. R.R. Dickson, J. Meincke, and P. Rhines, Quarterly Journal of the Royal Meteorological eds, Springer, Dordrecht. Society 128(585):2,187–2,203, http://dx.doi. Schauer, U., E. Fahrbach, S. Osterhus, and org/10.1256/qj.00.35. G. Rohardt. 2004. Arctic warming through Walsh, J.J., C.P. McRoy, T.H. Blackburn, the Fram Strait: Oceanic heat transport L.K. Coachman, J.J. Goering, K. Henriksen, from 3 years of measurements. Journal of P. Andersen, J.J. Nihoul, P.L. Parker, Geophysical Research 109, C06026, http://dx.doi. A.M. Springer, and others. 1989. &e role of org/10.1029/2003JC001823. Bering Strait in the carbon/nitrogen "uxes Schauer, U., H. Loeng, B. Rudels, V.K. Ozhigin, and of polar marine ecosystems. Pp. 90–120 in W. Dieck. 2002. Atlantic water in"ow through Proceedings of the Sixth Conference of the Comité the Barents and Kara Seas. Deep-Sea Research Arctique International. I. Ecosystems of Northern Part I 49(12):2,281–2,298, http://dx.doi. Seas. L. Rey, V. Alexander, and E.J. Brill, eds, org/10.1016/S0967-0637(02)00125-5. Leiden, Netherlands. Serreze, M.C., A.P. Barrett, A.G. Slater, Woodgate, R.A. 2008. Mooring Cruise Report for R.A. Woodgate, K. Aagaard, R.B. Lammers, RUSALCA Sever Cruise to the Bering Strait, Aug/ M. Steele, R. Moritz, M. Meredith, and Sept 2007. University of Washington, 17 pp. C.M. Lee. 2006. &e large‐scale freshwater Available online at: http://psc.apl.washington. cycle of the Arctic. Journal of Geophysical edu/BeringStrait.html. Research 111, C11010, http://dx.doi. Woodgate, R.A., and K. Aagaard. 2005. Revising org/10.1029/2005JC003424. the Bering Strait freshwater "ux into the Arctic Skagseth, Ø., K. Drinkwater, and E. Terrile. In Ocean. Geophysical Research Letters 32, L02602, press. Wind and buoyancy induced transport of http://dx.doi.org/10.1029/2004GL021747. the Norwegian Coastal Current in the Barents Woodgate, R.A., K. Aagaard, R.D. Muench, Sea. Journal of Geophysical Research, http:// J. Gunn, G. Bjork, B. Rudels, A.T. Roach, dx.doi.org/10.1029/2011JC006996. and U. Schauer. 2001. &e Arctic Ocean

Oceanography | September 2011 99