Generated using V3.2 of the official AMS LATEX template–journal page layout FOR AUTHOR USE ONLY, NOT FOR SUBMISSION!

Volume and Freshwater Flux Observations from Nares Strait to the West of Greenland

at Daily Time Scales from 2003 to 2009

Andreas Munchow ∗

University of Delaware, Newark, Delaware

ABSTRACT

Time series observation of velocity, salinity, pressure, and ice draft provide estimates of advectivefluxes in Nares Strait from 2003 to 2009 at daily to interannual time scales. Velocity and salinityare integrated across the 36 km wide and 350 m deep channel for two distinct multi-year periodsof sea ice cover. These observations indicate multi-year mean fluxes that range from 0.71±0.09 to1.03±0.11 Sv (Sv=10−6m3s−1=31,536 km3yr−1) for volume and from 32±5.7 to 54±9.3 mSv foroceanic freshwater for the first 2003-06 and second 2007-09 periods, respectively. Advection of iceadds another 8±2 mSv or 260±70 km3 per year to the freshwater export. Flux values are largerwhen the sea ice is mobile all year. About 75% of the oceanic volume and freshwater flux variabilityis correlated at daily to interannual time scales. Flux variability peaks at a 20-day time scale andcorrelates strongly with along-channel pressure gradients which explain 68% of the volume flux.The along-channel pressure gradient peaks in early spring when the sea ice is often motionless withhigher sea level in the Arctic that drives the generally southward ocean circulation. Local windscontribute only when the sea ice is mobile and explain 60% its variance. Observed changes inthe duration of motionless sea ice conditions impact ocean stratification and freshwater flux whileseasonal variations are small. [Submitted to J. Phys. Oceanogr., May-19, 2015]

1. Introduction

The Canadian Archipelago and Fram Strait constitutethe two pathways for the exchange of water and ice betweenthe Arctic and Atlantic Oceans. Arctic outflows via thesepathways return freshwater to the North Atlantic that wasevaporated from tropical oceans, transported by the at-mosphere, and delivered to the Arctic Ocean via precipi-tation, terrestrial run-off, and inflow from the North Pa-cific (Emile-Geay et al. 2003). This study focuses on thethrough-flows west of Greenland where Nares Strait is amajor conduit of southward flux into Baffin Bay (Munchowet al. 2015) and from there via Davis Strait (Curry et al.2014) onto the Labrador shelf and slope of the AtlanticOcean (Khatiwala et al. 1999). Fluxes through Nares Straitreflect the impacts of freshwater storage and release asso-ciated with the Beaufort gyre (Proshutinsky et al. 2009;Timmermans et al. 2011), diminishing sea ice in the Arctic(Parkinson and Cavalieri 2008; Stroeve et al. 2012), disin-tegrating ice shelves of northern Canada (Copland et al.2007), and potentially surging glaciers of northern Green-land (Rignot and Steffen 2008; Johnson et al. 2011). Mod-els of global climate projections are sensitive to the param-eterizations of both processes and pathways of freshwaterflux from the Arctic to the North Atlantic (Holland et al.2007; Curry and Mauritzen 2005).

The pack ice and upper Arctic Ocean are a reservoir ofnearly 100,000 km3 of freshwater (Aagaard and Carmack1989). This volume represents about 15 years of local inputthat buffers imbalances in the annual freshwater budget.Conversely, the reservoir can deliver large pulses of fresh-water to the Atlantic. The factors influencing storage andrelease relate to the wind stress curl over the central Arctic(Proshutinsky et al. 2009). Changes in this atmosphericproperty also contribute to a myriad of observed Arcticchange such as the shift in the boundary between Atlantic-and Pacific-derived waters (Morison et al. 1998); a shiftin the position of the trans-polar drift (Rigor et al. 2002);retreat and return of the cold halocline in the EurasianBasin (Alkire et al. 2007); reduced pack ice (Lindsay et al.2009); changes in freshwater content (Polyakov et al. 2008);increased run-off into the Eurasian sector (Peterson et al.2002); and many more (White et al. 2007).

Meanwhile, Arctic freshwater export impacts downstreamdeep water formation zones in the Greenland and LabradorSeas such as the Great Salinity Anomaly spreading over thesurface of the North Atlantic (Belkin 2004), freshening ofdeep waters both in the North-Atlantic (Curry and Mau-ritzen 2005) and adjacent basins such as Baffin Bay (Zwengand Munchow 2006). The propagation of these freshwatersignals is linked to atmospheric variations by both modelsand observations (Sundby and Drinkwater 2007; Hurrell

1

and Deser 2009). Furthermore, models predict that in-creased surface freshening in the Greenland and LabradorSeas will slow down down the global meridional overturn-ing circulation (Koenigk et al. 2007). Possible early evi-dence for such slowing has been reported by Bryden et al.(2005), although Wunsch and Heimbach (2006) suggeststhat aliasing may explain some of the observed trends inhydrographic observations.

The potential climate impact of Arctic freshwater ex-port motivates observational studies in the central Arctic(Proshutinsky et al. 2009) as well as its gates to the Pacific(Woodgate et al. 2012) and the Atlantic Oceans via Fram(Rabe et al. 2013; de Steur et al. 2009) and Davis Strait(Curry et al. 2014). Numerical models often compare theirpredictions against this metric (Jahn et al. 2012) as it con-nects high latitude physical forcing to mid-latitude envi-ronmental responses (Greene et al. 2008). The respectiveroles of freshwater flux to the west and east of Greenlandfeatures in most modern large scale coupled ice, ocean, at-mospheric circulation studies (Holland et al. 2007). Forsimilar reasons McGeehan and Maslowski (2012), Wekerleet al. (2013), and Lu et al. (2014) all simulate decadalvariability of freshwater flux through the Canadian ArcticArchipelago. The models generally agree with observationsfor integrated linear properties such as freshwater content,volume flux, and sea level, however, model and data divergeon nonlinear products of properties such as freshwater fluxwhich requires both velocity and salinity estimates.

Efforts to compile estimates of various component fluxesthrough the Arctic gateways have generally asserted thatthe freshwater fluxes east and west of Greenland are simi-lar in size (Beszczynska-Moller et al. 2011). Dickson et al.(2007) cautioned, however, that the fluxes to the east andwest of Greenland are potentially unsteady and may impactthe overturning circulation differently. More specifically,the impact of the Arctic freshwater outflow on the globaloverturning circulation depends on whether (a) it is spreadto depth in the Atlantic Ocean via the overflow system ofNordic Seas east of Greenland (Dickson et al. 2002), (b) itis inserted at the surface of the North-Atlantic via eddies,or (c) it is trapped on continental shelves (Khatiwala et al.1999; Myers 2005). Koenigk et al. (2007) predict a likely48% increase in the freshwater flux through the CanadianArchipelago, compared to only a 3% increase through FramStrait, as a result of a much diminished Arctic ice cover.

Direct current measurements are sparse within the Cana-dian Archipelago (Melling et al. 2008). Strong and persis-tent currents have been measured in Robeson (Sadler 1976)and Kennedy Channel (Munchow et al. 2007) of NaresStrait as well as in Smith Sound (Melling et al. 2001) and inBarrow Strait (Peterson et al. 2012). Mean currents in thedirection of Kelvin wave propagation are often intensifiednear the surface adjacent to channel boundaries (LeBlond1980; Munchow et al. 2007).

Munchow and Melling (2008) provided initial descrip-tions and analyses of ocean currents in Nares Strait from2003 through 2006. Focusing on depth-averaged propertiesonly, they found a steady sectionally- averaged mean flow ofabout 0.06 m/s to the south that corresponds to a volumeflux of 0.72 ± 0.11 Sv. This study extend this work by (a)introducing a more complete 2003-09 record, (b) extendingvelocity observations to the surface using new estimates ofice velocity, (c) presenting full section estimates of volumeand freshwater flux, and, finally, (d) demonstrating thatboth volume and freshwater flux correlate strongly witheach other and the along-channel pressure gradients.

2. Study Area, Data, and Methods

Nares Strait connects the Lincoln Sea of the Arctic toBaffin Bay of the Atlantic Ocean (Figure 1). Across a36 km wide section an array of 75 kHz acoustic Dopplercurrent profilers (ADCP) of Teledyne RD Instruments wasdeployed in 2003 at the bottom at depths ranging from170-m off Greenland to 370-m in the center of KennedyChannel near 80.5 N latitude (Figure 2). Munchow andMelling (2008) describe mooring details and discuss verti-cally averaged properties at tidal to interannual time scalesfor five instruments recovered in 2006. A second deploy-ment of these five ADCPs in 2007 was recovered withoutlosses in 2009. Ocean currents were sampled in bursts for5 minutes every half hour in 8-m vertical bins from about15-m above the seafloor to about 30-m below the surface(Munchow et al. 2006). Tidal oscillations are removed fromthe record via harmonic analysis prior to application of aLanczos raised cosine filter with a half-power point near34 hours (Munchow and Melling 2008). Table 1 providessalient mooring details.

Data from bottom-mounted ADCP moorings are alsoused to estimate ice velocity from water-tracking pings.Following methodology developed independently by Mellinget al. (1995) and Visbeck and Fischer (1995), I use theDoppler shift from the vertical bin that contains a near-surface maximum of acoustic backscatter. This maximumis distinct across all four beams in the presence of level icewhen all four acoustic beams track the same piece of ice.The resulting estimates were filtered and subsampled at 3-hour intervals to retain the tides while reducing noise. Icedraft measurements originate from an ASL Environmen-tal Sciences ice profiling sonar to distinguish an ice surfacefrom open water (Melling 1998). Two sonar were mooredadjacent to ADCPs at 100 m depth and measured ice draftto within 0.1 m (Ryan, pers. comm., 2015).

Moored strings held four SeaBird SBE37SM Microcatseach that measured conductivity and temperature (C/T)to estimate salinity above 200-m as well as pressure attwo vertical locations from which to infer mooring motions.Rabe et al. (2010, 2012) describes details of the mooring

2

design, location, and data processing for the 2003-06 pe-riod that we here apply to the 2007-09 period. AdditionalC/T sensors were placed 3-m above the bottom at ADCPmooring locations (Munchow et al. 2011).

Bottom pressure observations originate from a Digi-quartz pressure sensor of ParaScientific Inc. that is accu-rate to within 0.2 mbar (Munchow and Melling 2008) andwas deployed in a shallow bay of Foulke Fjord, Greenland(78.30 N, 72.57 W) at the southern end of Nares Strait.Data from a surface tide gauge at Alert, Canada (82.49 N,75.80 W) was used and corrected for the inverted barom-eter effect in order to estimate the bottom pressure gradi-ent along Nares Strait by subtracting corrected Alert fromFoulke Fjord bottom pressure estimates.

Wind data originate from numerical simulations de-scribed by Samelson and Babour (2008). The Nares Straitmodel uses boundary conditions from operational globalmeteorological forecast models via 3-step nesting proce-dures at 54, 18, and finally 6-km resolutions. The 6 kmmodel provides realistic 36 hour predictions of wind fieldsfor Nares Strait at hourly intervals. I here use model out-put from a single grid point near the center of the mooringssection near 80.5 N latitude and 68 W longitude.

I use remotely sensed surface conditions from daily ob-servations of microwave SSM/I and optical MODIS satel-lites, respectively. The SSM/I data used are sea ice concen-trations obtained from NSIDC archives at 25 km resolution(Steffen and Schweiger 1991) while MODIS data are sur-face reflectance at 865 nm obtained from NASA’s GoddardSpace Flight Center at 250 m resolution.

3. Sea Ice Conditions

Nares Strait is covered by sea ice for most of the year(Figure 1). From July through November the ice is mobilewhile from December through June it is generally landfastas a result of ice arches that form at the northern entranceat 83◦ N and southern exit at 78◦ N latitudes (Kwok et al.2010). Figure 2 shows an optical MODIS-Terra satelliteimage from April-21, 2008 at 250-m spatial resolution. Thelow sun angle naturally illuminates the high mountainousterrain of Ellesmere Island, Canada in the west and Green-land in the east. Black ocean areas to the south of 80◦ Nindicate open water or very thin ice as an ice bridge blocksall ice motion to the north. The ocean under the ice moves,however, as indicated by low-pass filtered current vectorsobserved at the mooring locations on April-21, 2008. Theocean under the ice moves southward at speeds reaching0.2 m s−1 while the ice above it is stationary.

Almost exactly a year later on April-22, 2009 Figure 3shows a different winter distribution of the sea ice in NaresStrait. No southern ice bridge formed this year, however, anorthern ice arch prevents advection of ice into the channelwhich is covered by thin mobile ice or open water. Note

that surface currents are again southward. Largest down-stream velocities at this time occur in the center of thechannel reaching 0.3 m s−1 while the year prior largestflow was adjacent to the western coast. The presence ofmobile versus immobile and landfast ice will impact dy-namics, volume, and freshwater flux.

4. Circulation

Measuring ocean currents within 30 m of an ice-oceaninterface challenges available technologies. For example,the bottom-mounted ADCPs in Nares Strait measure ve-locity reliably only to within 30 or 40-m of the surface, be-cause acoustic sidelobe interference from a strong scattersuch as the surface mask the water column signal. Whilethis interferences prevents retrieval of water velocity mea-surements, the scatter from the surface allows estimationof an ice surface velocity (Melling et al. 1995) than is thenused to with subsurface water velocities to interpolate ve-locity data to a dynamically motivated vertical shear model(Munchow et al. 2007). These data and methods will pro-vide for accurate volume and freshwater flux estimates atdaily to interannual time scales.

a. Ocean Velocity

Figure 4 shows vertically averaged ocean currents as therecord mean (arrow) and subtidal variability (filled ellipse)for the 2003-06 and 2007-09 observational periods that arealso summarized in Table 2. Record mean flows are alwaysto the south. Strongest mean currents occur within 10-kmoff the Ellesmere Island coast in the west. In contrast, wefind largest variability off Greenland where the ellipse islong and slender (Munchow and Melling 2008).

The depth-dependent flow is always to the south alsoand statistically significantly different from zero at 95%confidence (Figure 5). Veering occurs below 200 m depthtowards the bottom and above 30 m towards the surface.Record mean currents are identical for the two periodswithin their confidence interval except for the near sur-face flow that is significantly stronger during the 2007-09relative to the 2003-06 period. Figure 6 shows time seriesof wind and ocean current vectors within 115 m of the sur-face. Winds are persistent from the north and exceed 10m s−1 frequently. Ocean currents are detided and low-passfiltered. Wind and 35 m currents have strong seasonal-ity with strongest southward wind and currents occurringin January and February with weaker southward flow andnorthward winds in June and July. Strong winds and oceancurrents along with mobile sea ice will result in large valuesfor both volume and freshwater flux during the 2007-09 pe-riod. In contrast, during the 2003-06 period persistent icearches form to the south of our study area and shut downall ice motion and atmospheric momentum transfer to theocean for 5 to 8 months each year (Kwok et al. 2010).

3

b. Ice Velocity

Figure 7 shows the scatter of ice velocity vectors forthe 2003-06 period at three locations across Nares Straitwhile Figure 8 depicts the corresponding time series of thedominantly along-channel flow along with the ice draft. Incombination these data document the passage of rafted iceduring the mobile season as well as slowly growing level iceduring the winter periods when ice motion is small. Duringthe summer ice motions frequently exceed 1 m s−1 and aregenerally stronger over the center and off Greenland ascompared to the western channel off Canada’s EllesmereIsland.

I find the same lateral ice velocity distribution for thesecond 2007-09 period of observation (Figure 9). Note theabrupt change in ice draft in February 2009 after a northernice arch forms (Figure 3). A southern ice arch never formsin 2009 and all thick ice is flushed out of Nares Strait andreplaced by thin, mobile first year ice until the failure of thenorthern ice arch in July 2009 marks the arrival of thickerice from the Arctic Ocean to the north.

Almost 60% of the variance of along-channel ice motioncorrelates with the variance of the local winds. Figure 10shows the complex (vector) correlation Cxy of wind with icemotions for the entire 2007-09 period as a function of lag.Largest correlation occurs at zero lag at an angle Dir(Cxy)of -10◦ which represents a clock-wise rotation of the windinto the sea ice vector. The lagged auto-correlation of theice velocity indicates ∼1 and ∼10 day decorrelation timescales for the across- and along-channel velocity compo-nents, respectively. Tidal oscillations are visible in thealong- but not across-channel auto-correlations. Harmonicanalyses reveal coherent tidal oscillations of the ice motionthat are about 60% of the tidal currents 40-m below thesurface (not shown).

c. Surface Velocity

Sub-surface ocean and ice velocity vectors are combinedto estimate surface layer velocities via interpolation. Morespecifically, I interpolate velocity observations above 100m depth towards the surface by fitting a linear shear anda vertical Ekman layer profile. Both along- and across-channel velocity components (u,v) are fitted concurrently,that is,

u(z) = u0 + u1 ∗ z + vE ∗ s(z) + uE ∗ c(z) (1)

v(z) = v0 + v1 ∗ z + vE ∗ c(z) − uE ∗ s(z) (2)

where z is the vertical distance from the surface scaledby the vertical Ekman layer depth D=(2*Av/f)1/2=10 m,f is the Coriolis parameter, and Av is the vertical eddyviscosity, while s(z)=sin(z)*ez and c(z)=cos(z)*ez are thefunctions to describe the Ekman layer profile. The 6 freeparameters (u0,v0), (u1,v1), and (uE,vE) are determined

by minimizing the least square errors for data from the top100 m of the water column at each time step for each pro-file. The root mean square error varies between 3.3 cm/sand 4.6 cm/s. The largest errors always occur within 5km of the Ellesmere Island coast (not shown). Munchowet al. (2007) applied a similar procedure to vessel-mountedADCP observations and discuss both sensitivity and un-certainty of the method.

5. Salinity

Salinity observation from moored sensors reveal dra-matically different structures in the 2003-06 observationalperiod when it is compared to 2007-09 observations. Fig-ures 11 and 12 show both the time-averaged and instanta-neous salinity distributions across Nares Strait from moor-ings and ship-based summer profiling. Rabe et al. (2010)and Rabe et al. (2012) discuss the 2003-06 data in detailand estimate geostrophic flux below 30 m depth, respec-tively. Extending this methodology to the entire section, Ihere construct daily salinity distribution by assuming (a)a vertically uniform surface mixed layer with the time-varying salinity observed near 30 m depth, (b) a time-varying observed vertical profile between 30 and 200 mdepth, and (c) a steady salinity distribution below 200 mwhere we use observed data from the summer surveys in2003 and 2007 for all days.

Both the summer snapshots and interannual mean sec-tions indicate isohalines that slope upward from EllesmereIsland in the west to Greenland in the east. This reflectsthe baroclinic contribution to the geostrophic circulation(Rabe et al. 2012). Above 100 m depth minimal salini-ties of about 32 psu occur near the center of the channelin 2003-06. In contrast, the 2007-09 salinity minima oc-curs closer the Ellesmere Island coast where record-meansurface salinities reach as low as 31 psu decreasing contin-uously from east to west. This response is consistent witha wind-forced surface circulation shown in Figure 6 dueto mean winds to the south (Samelson and Babour 2008).Recall that the second deployment period coincided withlargely mobile ice conditions year-round.

6. Volume and Freshwater Flux

a. Flux Definitions

Velocity and salinity fields u(y,z,t) and s(y,z,t) are de-composed into two components , e.g., u = U + u′ ands = S+s′ where U(y,z) and S(y,z) are time-mean fields withtemporal fluctuations u’(y,z,t) and s’(y,z,t). The time aver-age < • > of the fluctuations is zero by construction, thatis, < u′ >=< s′ >= 0 while < u >= U , and < s >= S.The mean volume flux < Q(t) > is then defined as

< Q(t) >=

∫∫

A

U dy dz (3)

4

where the integral is over the wet area A=10.59 km2

of the section (y, z). The measurements are interpolatedonto a regular grid with dz = 4 m and dy = 6 km spac-ing over a 400 m vertical and 36 km lateral domain usinga bi-harmonic spline interpolation that is equivalent to aminimum curvature criteria as implemented in the GeneralMapping Tools routine ”surface” (Wessel 2009).

The flux of ocean freshwater Qf (t) combines along-channel velocity u with salinity anomaly (s-S0)/S0 whichfor the time mean Qf0 becomes

Qf0 =< Qf (t) >=<

∫∫

A

u × (s/S0 − 1) dy dz >

=

∫∫

A

U × (S/S0 − 1) dy dz

+ <

∫∫

A

u′× s′/S0 dy dz > (4)

for a reference salinity S0 that we take as 34.8. Thegridding details are identical to those of volume flux, butuncertainty estimates now consist of both velocity and salin-ity contributions. The time average of the second term isnegligible only if velocity and salinity fluctuations are or-thogonal. Many observation-based Arctic freshwater fluxnumbers assume this contribution to be negligible, becausedata to evaluate them are rarely available. In contrast, ourdaily velocity and salinity estimates from moored observa-tions reveal that the second term never exceeds 0.8 mSvand thus falls within our flux uncertainty estimates.

b. Daily Flux Time Series

Using daily salinity and velocity data from moorings,I apply the same objective gridding procedures to evalu-ate Q(t) and Qf (t). Figure 13 shows the resulting timeseries as well as its low-passed filtered, roughly monthlycomponent thereof. The temporal mean flux increases by68% from 32 mSv (1053 km3 y−1) in 2003-06 to 54 mSv(1780 km3 y−1) in 2007-09 (Table 3). Only the first valuecompares favorably to the geostrophic estimate of 28 mSv(921 km3 y−1) for the same section and period (Rabe et al.2012). The likely cause for this dramatic increase relatesto the almost year-round mobile ice seasons in 2007-09.In most years a persistent ice arch across southern SmithSound blocks ice motion throughout Nares Strait for 4-6months each winter (Kwok et al. 2010). In contrast, icewas mobile for the 2007-09 period except for 6 weeks in2008 as indicated in Figures 2 and 9.

While daily flux estimates range from a southward -2.5to a northward +1.2 Sv for volume and -130 to 45 mSvfor freshwater, monthly flux values are always to the south(Figure 13). Furthermore, flux exhibits little seasonalityeven though largest southward freshwater flux occurs dur-ing December and January of each year except in 2004/05.

The standard deviation about the respective mean flux issimilar for the two observational periods, about 21 versus26 mSv for freshwater and 0.44 Sv for volume (Figure 13).

Volume and freshwater flux correlate strongly as 76%and 74% of the variances relate to each other for the 2003-06 and 2007-09 periods, respectively. Figure 14 shows thiscorrelation as well as the corresponding linear regression,that is, each Sv of volume flux accounts for 40±1.4 and51±2.1 mSv of freshwater flux for the two periods. Theuncertainties are 95% confidence limits in multiple regres-sion (Fofonoff and Bryden 1975) that indicate significantlydifferent gains from the first to the second period. Volumeflux estimates for Nares Strait can be used to estimatefreshwater flux to within 25% at daily time scales.

c. Mean Flux and Random Uncertainty

Figures 15 and 16 show the distribution of the record-mean along-channel currents U for the 2003-06 and 2007-09 periods, respectively, as well as U × (S/S0 − 1). Asurface intensified high-velocity core occupies the center ofthe channel where southward speeds reach 0.4 m s-1 about15 km from the coast off Ellesmere Island. The mean flowreduces to 0.1 m s-1 off Ellesmere Island while within about10 km off Greenland the mean surface flow is close to zero.Northward flows only occur within 3 km off Greenlandwhere they reach 0.05 m s-1 in water 100 m deep. Theflow within 200 m of the bottom elsewhere in the sectionis to the south and varies between 0.05 and 0.1 m s-1 .

The mean volume flux < Q > ±δ is 0.71±0.06 Sv and1.03±0.07 Sv for the 2003-06 and 2007-09 periods, respec-tively. The uncertainty δ is a 95% confidence level due toa random velocity uncertainty of δu=0.01 m s-1 (Table 2),that is,

δ = 1.96 ×

√

( ∫∫

Aδu dy dz

)2

2N − 1(5)

where the degrees of freedom are 2N-1 for N mooringlocations. I here assume that each mooring location rep-resent 2 independent velocity estimates, because the firsttwo empirical orthogonal functions (not shown) generallyexplain 85% of the variance at each mooring location with15 to 39 velocity timeseries separated by 8 m in the ver-tical. Mean across-channel velocities (not shown) are lessthan 10−3 m s-1 when averaged over the section.

The record-mean freshwater flux increases from 32±2.5mSv for the 2003-06 to 54±3.9 mSv for the 2007-09 periods.Uncertainties here are estimated from error propagationwhich for a linear product such as u × s is in the sum ofthe errors due to velocity and salinity. Each contributionis estimated by an integral similar to Equation-5. Randomsalinity and velocity uncertainty of δs=0.1 and δu=0.01 ms-1 result in freshwater flux uncertainties of 1.8 (2.0) mSvfrom velocity and 0.8 (1.9) mSv for salinity for the 2003-06

5

(2007-09) period.Freshwater flux estimates are lower bounds, because

salinities at 30 m depth are assumed to characterize a uni-form surface layer, Actual surface salinities decrease to-wards the surface at times when vertical salinity stratifi-cation exists in this upper layer. Data from CTD surveysdemonstrate that such stratification exists during the shortsummer when sea ice melt contributes to vertical stratifi-cation (Figures 11 and 12).

d. Surface Layer Flux and Systematic Errors

Table 3 summarizes record-mean fluxes and their ran-dom errors, however, flux estimations contain systematicbesides random errors. For example, flux values include in-terpolated surface velocity and extrapolated surface salin-ity estimates. Evaluating the impact of the surface layer,I construct flux estimates that exclude an increasing frac-tion of the surface layer from integrals such as Eqs. 3 and4. Figure 17 then shows how much of the flux is containedwithin the excluded surface layer that is varied from zeroto 30 m. Volume and freshwater flux are normalized by itsthe full-section value given in Table 3.

Observations from Nares Strait suggest that more than50% of the freshwater and more than 20% of the volumeflux are contained within 30 m of the surface. This esti-mate reflects data from multiple seasons and multiple yearswhen the ice is mobile, landfast, or not present. These largeflux contributions from an unmeasured surface layer implypotentially large systematic error that depend on how ob-servations are extended into this surface layer. For exam-ple, if ice velocity estimates are consistently high or lowby 20%, say, then this introduces a small error for volumeflux (4% or ∼0.04 Sv), but it will introduce a large errorfor freshwater flux (10% or ∼ 5 mSv), because the freshwa-ter flux is more sensitive to surface velocity than is volumeflux. Table 3 presents both the random and the randomplus systematic error for all record-mean flux estimates.

7. Wind and Pressure Gradient Forcing

Winds drive ocean circulation both locally via a surfacewind stress and remotely via horizontal pressure gradients(Garvine 1985). For local winds to drive ocean currentsin ice-covered seas, however, the ice must be sufficientlymobile to allow the transfer of momentum. Sizeable icevelocities appear during the 2007-09 period at all timesexcept for April and May of 2008 (Figure 9) that drive iceand oceans below. Figure 10 shows correlations of windwith ice velocity vectors and demonstrates the dominanceof local winds in Nares Strait sea ice motions at daily timescales when ice is mobile.

In contrast, along-channel pressure gradient forcing ofthe ocean does not depend on ice cover or motion. Derivingan estimate of this pressure gradient, I use the difference

of bottom pressure recordings from a northern and south-ern station and remove the record-mean pressure.Munchowand Melling (2008) use identical data for a frequency-domaincorrelation analysis that reveals maximal correlation r2

∼

0.8 near a monthly time scale at zero lag. Figure 18 showstime series of pressure gradient and volume flux that arelow-passed filtered with a Lanczos raised cosine filter thathas a half-power point near 2 weeks. The difference in sealevel dP along the 450 km long Nares Strait varies by about±0.1 m while volume flux Q varies in the range from zeroto 1.8 Sv to the south. A fraction of these data are used inthe regression

Q = a + b ∗ dP (6)

where volume flux Q and pressure difference dP fromthe first year are used to find coefficients a=-0.92±0.023Sv and b=9.0±0.66 Sv/m for the remainder of the timeseries. The uncertainties are 95% confidence limits in mul-tiple linear regressions (Fofonoff and Bryden 1975). I thencompare volume flux predictions Qp=a+b*dP with volumeflux observations Q and show the results in Figure 18. Theagreement of this simple forecasting model is excellent asthe residuals are always less than 0.5 Sv with a root meansquare error of 0.2 Sv. Figure 19 shows both the regres-sion and the scatter of the data. The correlation betweenthe two time series is 0.82, that is, 68% of the volume fluxvariance relates to the variance in the pressure differencealong Nares Strait.

8. Discussion and Conclusions

Ice and ocean sensors deployed in Nares Strait between2003 and 2009 quantified the export of volume and freshwa-ter from the Arctic to the North Atlantic Oceans betweennorthern Canada and Greenland. These data demonstrateddramatic change in all observed and estimated parametersfrom a first 2003-06 to a second 2007-09 study period. Bothice and ocean velocities increased in southward magnitudewhile surface waters also became fresher. More specifi-cally, volume flux increased by 45%, ocean freshwater fluxincreased by 69%, and ice freshwater flux increased by 46%from the first to the second period (Table 3).

These observed changes relate to different winter sea iceregimes during these two periods. Sea ice during the first2003-06 period was motionless for 5-8 months of the yearas the result of a southern ice arch that forms in late winterand collapses in early summer (Kwok et al. 2010). This icearch failed to form in the winters of 2006/07, 2007/08, and2009/10 while it existed for less than 2 months in 2008/09.

While Kwok et al. (2010) documents these first recordedfailures using remotely sensed ice properties, this studyquantifies how changes in winter ice regime impact oceanvolume and liquid freshwater flux. More specifically, thenew regime without a southern ice arch facilitates larger

6

flux in winter, enhances momentum transfer via coupledice-ocean motions, results in an ”earlier” open water sea-son, and allows enhanced solar heating. These changes allincrease the time window that thick Arctic ice can streamsouthward through Nares Strait in summer, partially meltin transit in Nares Strait and thus freshening ocean surfacewaters. Furthermore, when the sea ice cover is mobile, localwind forcing provides a more efficient momentum transferfrom the atmosphere to the ocean.

All prior flux studies of Nares Strait were limited to ob-servations below 30 m depth (Munchow and Melling 2008;Rabe et al. 2012) or were based on daily snapshots of sur-vey data (Munchow et al. 2007). In contrast, this studyprovided daily flux estimates for the complete section byinterpolating observations into the 30-m surface layer us-ing ice velocities that were derived from the same ADCPsthat measured water column velocities (Melling et al. 1995;Visbeck and Fischer 1995). Statistical analyses of these icevelocities reveal both tidal oscillations as well as strongcorrelations with the local winds at a 10 degree angle atzero lag. This gives confidence that both ice and surfacevelocity estimates are accurate and physically meaningful.

Furthermore, the complete 2003-09 data set includesa second period of observations that differs qualitativelyand quantitatively from the prior 2003-06 record. We thusconclude that present flux estimates are more robust, com-plete, and comprehensive. All freshwater flux estimatescontain uncertainty estimates which incorporate contribu-tions from random velocity and salinity errors as well as apotential systematic error from the surface layer interpola-tion.

Along-channel bottom pressure gradients explain 68%of the volume flux for the 2003-06 period that includes pe-riods with and without ice motions. Furthermore, about75% of the freshwater flux variance relates linearly to thevolume flux variance for both 2003-06 and 2007-09 peri-ods even though the gain of volume to freshwater flux dif-fers. These analyses suggest that two tide gauges or bottompressure sensors placed at the northern entrance and south-ern exit of Nares Straits may capture 60% of freshwater fluxvariability. The expensively acquired measurements of thisstudy thus provide the basis for a inexpensive monitoringarray.

The coupling of Arctic to North-Atlantic Oceans inNares Strait to the west of Greenland depends on the mo-bility of the sea ice cover. While along-channel pressuregradients dominate the ocean flow irrespectively of the icecover, the flux of freshwater is increased by almost 70%during years when the sea ice is mobile in winter. Inciden-tally, the large-scale general circulation climate model ofKoenigk et al. (2007) predicts an increase of liquid fresh-water flux through the Canadian Archipelago by 38% from80 to 110 mSv for the second half of the 21st century as aconsequence of a warming world with diminished and more

mobile sea ice cover (Stroeve et al. 2012).The precise dynamics is unclear from an observational

perspective, but I speculate that more efficient local mo-mentum transfer from the wind field to the ocean plays acentral role as does the baroclinic adjustment of a bound-ary current along the Canadian Ellesmere Island coast thatis caused by enhanced onshore surface Ekman flux. Onlydetailed process oriented numerical studies motivated byobservations and analyses will provide the dynamical un-derstanding needed to make meaningful predictions intothe future.

The southern ice arch formed again each winter 2010/11through 2014/15 at its normal location for several monthseach year. The 2007-09 period may thus be regarded as anextreme anomaly caused by an intermittent release of Arc-tic freshwater from the Beaufort Gyre (Curry et al. 2014;Timmermans et al. 2011).

Acknowledgments.

The National Science Foundation supported the ini-tial fieldwork with OPP-0230236 and analyses with OPP-1022843. The University of Delaware provided substantialadditional support with capital equipment and salaries. In-strument preparation, deployment, and recoveries were ac-complished through dedicated efforts by a large group of in-ternational scientists, engineers, technicians, students, andstaff in addition to Captains and crews of multiple shipson multiple expeditions. Humfrey Melling of the CanadianDepartment of Fisheries and Oceans provided critical lead-ership and design ideas on all moored ocean sensor systemsthat were implemented creatively by Peter Gamble, DavidHuntley, Ron Lindsey, and Jonathan Poole. Pat Ryan pro-cessed ice draft data shown in Figures 8 and 9.

REFERENCES

Aagaard, K. and E. C. Carmack, 1989: The role of seaice and other fresh-water in the Arctic circulation. J.

Geophys. Res., 94 (C10), 14 485–14498.

Alkire, M. B., K. K. Falkner, I. Rigor, M. Steele, andJ. Morison, 2007: The return of Pacific waters to theupper layers of the central Arctic Ocean. Deep-Sea Res.,54 (9), 1509–1529.

Belkin, I. M., 2004: Propagation of the ”Great SalinityAnomaly” of the 1990s around the northern North At-lantic. Geophys. Res. Let., 31 (8), L08 306.

Beszczynska-Moller, A., R. A. Woodgate, C. Lee,H. Melling, and M. Karcher, 2011: A synthesis of ex-changes through the main oceanic gateways to the ArcticOcean. Oceanography, 24 (3), 82–99.

7

Bryden, H. L., H. R. Longworth, and S. A. Cunningham,2005: Slowing of the Atlantic meridional overturning cir-culation at 25 degrees N. Nature, 438 (7068), 655–657.

Copland, L., D. R. Mueller, and L. Weir, 2007: Rapidloss of the Ayles ice shelf, Ellesmere Island, Canada.Geophys. Res. Let., 34 (21), L21 501, doi:10.1029/2007GL031809.

Curry, B., C. M. Lee, B. Petrie, R. E. Moritz, and R. Kwok,2014: Multiyear volume, liquid freshwater, and seaice transports through Davis Strait, 2004-10. J. Phys.

Oceanogr., 44 (4), 1244–1266.

Curry, R. and C. Mauritzen, 2005: Dilution of the north-ern North Atlantic Ocean in recent decades. Science,308 (5729), 1772–1774.

de Steur, L., E. Hansen, R. Gerdes, M. Karcher,E. Fahrbach, and J. Holfort, 2009: Freshwater fluxes inthe East Greenland Current: A decade of observations.Geophys. Res. Let., 36, L23 611.

Dickson, B., I. Yashayaev, J. Meincke, B. Turrell, S. Dye,and J. Holfort, 2002: Rapid freshening of the deepNorth Atlantic Ocean over the past four decades. Na-

ture, 416 (6883), 832–837.

Dickson, R., B. Rudels, S. Dye, M. Karcher, J. Meincke,and I. Yashayaev, 2007: Current estimates of fresh-water flux through Arctiac and subarctic seas. Progr.

Oceanogr., 73 (3-4), 210–230.

Emile-Geay, J., M. A. Cane, N. Naik, R. Seager, A. C.Clement, and A. van Geen, 2003: Warren revisited: At-mospheric freshwater fluxes and ”why is no deep wa-ter formed in the North Pacific”. J. Geophys. Res.,108 (C6), 3178, doi:10.1029/2001JC001058.

Fofonoff, N. and H. Bryden, 1975: Density of sea waters.J. Mar. Res., 41, 69–82.

Garvine, R. W., 1985: A simple-model of estuarine subtidalfluctuations forced by local and remote wind stress. J.

Geophys. Res., 90 (NC6), 1945–1948.

Greene, C., J. Pershing, T. Cronin, and N. Ceci, 2008:Arctic climate change and its impacts on the ecology ofthe North Atlantic. Ecology, 89 Suppl., S24–S38.

Holland, M. M., J. Finnis, A. P. Barrett, and M. C. Ser-reze, 2007: Projected changes in Arctic Ocean freshwa-ter budgets. J. Geophys. Res., 112 (G4), G04S55, doi:10.1029/2006JG000354.

Hurrell, J. W. and C. Deser, 2009: North Atlantic climatevariability: The role of the North Atlantic Oscillation.J. Mar. Sys., 78 (1), 28 – 41.

Jahn, A., et al., 2012: Arctic Ocean freshwater: How ro-bust are model simulations? J. Geophys. Res., 117,C00D16.

Johnson, H., A. Munchow, K. Falkner, and H. Melling,2011: Ocean circulation and properties in PetermannFjord, Greenland. J. Geophys. Res., 116, C01 003, doi:10.1029/2010JC006519.

Khatiwala, S. P., R. G. Fairbanks, and R. W. Houghton,1999: Freshwater sources to the coastal ocean offnortheastern North America: Evidence from (h2o)-o-18/(h2o)-o-16. J. Geophys. Res., 104 (C8), 18 241–18 255.

Koenigk, T., U. Mikolajewicz, H. Haak, and J. Jungclaus,2007: Arctic freshwater export in the 20th and 21stcenturies. J. Geophys. Res., 112 (G4), G04S41, doi:10.1029/2006JG000274.

Kwok, R., L. T. Pedersen, P. Gudmandsen, and S. S. Pang,2010: Large sea ice outflow into the Nares Strait in2007. Geophys. Res. Let., 37 (3), L03 502, doi:10.1029/2009GL041872.

LeBlond, P. H., 1980: On the surface circulation in somechannels of the Canadian Arctic Archipelago. Arctic,33 (1), 189–197.

Lindsay, R. W., J. Zhang, A. Schweiger, M. Steele, andH. Stern, 2009: Arctic sea ice retreat in 2007 followsthinning trend. J. Clim., 22 (1), 165–176.

Lu, Y. Y., S. Higginson, S. Nudds, S. Prinsenberg, andG. Garric, 2014: Model simulated volume fluxes throughthe Canadian Arctic Archipelago and Davis Strait: Link-ing monthly variations to forcing in different seasons. J.

Geophys. Res., 119 (3), 1927–1942.

McGeehan, T. and W. Maslowski, 2012: Evaluation andcontrol mechanisms of volume and freshwater exportthrough the Canadian Arctic Archipelago in a high-resolution pan-Arctic ice-ocean model. J. Geophys. Res.,117, C00D14.

Melling, H., 1998: Sound scattering from sea ice: As-pects relevant to ice-draft profiling by sonar. J. Atmos.

Oceanic Technol., 15 (4), 1023–1034.

Melling, H., 2000: Exchange of freshwater transport

through the sahllow straits of the North American Arctic,chap. 20, 479–502. Kluwer Academic Publishers, Dor-drecht, Netherlands.

Melling, H., Y. Gratton, and G. Ingram, 2001: Ocean cir-culation within the North Water polynya of Baffin Bay.Atmos.-Oceans, 39 (3), 301–325.

8

Melling, H., P. Johnston, and D. Riedel, 1995: Measure-ments of the underside topography of sea-ice by mooredsubsea sonar. J. Atmos. Oceanic Technol., 12 (3), 589–602.

Melling, H., et al., 2008: Fresh-water fluxes via Pacific and

Arctic outflows across the Canadian polar shelf, chap. 9,193–261. Springer Verlag, Dordrecht, Netherlands.

Morison, J., M. Steele, and R. Andersen, 1998: Hydrog-raphy of the upper Arctic Ocean measured from thenuclear submarine USS Pargo. Deep-Sea Res., 45 (1),15–38.

Munchow, A., K. Falkner, and H. Melling, 2007: Spa-tial continuity of measured seawater and tracer fluxesthrough Nares Strait, a dynamically wide channel bor-dering the Canadian Archipelago. J. Mar. Res., 65 (6),759–788.

Munchow, A., K. Falkner, and H. Melling, 2015: BaffinIsland and West Greenland Current systems in northernBaffin Bay. Progress In Oceanography, 132, 305–317.

Munchow, A., K. K. Falkner, H. Melling, B. Rabe,and H. L. Johnson, 2011: Ocean warming of NaresStrait bottom waters off northwest Greenland, 2003-2009. Oceanography, 24 (3), 114–123.

Munchow, A. and H. Melling, 2008: Ocean current ob-servations from Nares Strait to the west of Greenland:Interannual to tidal variability and forcing. J. Mar. Res.,66 (6), 801–833.

Munchow, A., H. Melling, and K. K. Falkner, 2006: Anobservational estimate of volume and freshwater fluxleaving the Arctic Ocean through Nares Strait. J. Phys.

Oceanogr., 36 (11), 2025–2041.

Myers, P. G., 2005: Impact of freshwater from the Cana-dian Arctic Archipelago on Labrador Sea Water forma-tion. Geophys. Res. Let., 32 (6), L06 605, doi:10.1029/2004GL022082.

Parkinson, C. L. and D. J. Cavalieri, 2008: Arctic seaice variability and trends, 1979-2006. J. Geophys. Res.,113 (C7), C07 003, doi:10.1029/2007JC004558.

Peterson, B. J., R. M. Holmes, J. W. McClelland, C. J.Vorosmarty, R. B. Lammers, A. I. Shiklomanov, I. A.Shiklomanov, and S. Rahmstorf, 2002: Increasing riverdischarge to the Arctic Ocean. Science, 298 (5601),2171–2173.

Peterson, I., J. Hamilton, S. Prinsenberg, and R. Pettipas,2012: Wind-forcing of volume transport through Lan-caster Sound. J. Geophys. Res., 117, C11 018.

Polyakov, I. V., et al., 2008: Arctic ocean freshwaterchanges over the past 100 years and their causes. J.

Clim., 21 (2), 364–384.

Proshutinsky, A., et al., 2009: Beaufort Gyre freshwaterreservoir: State and variability from observations. J.

Geophys. Res., 114, C00A10.

Rabe, B., H. Johnson, A. Munchow, and H. Melling, 2012:Geostrophic currents and freshwater fluxes throughNares Strait to the west of northern Greenland. J. Mar.

Res., 70, 603–640.

Rabe, B., A. Munchow, H. Johnson, and H. Melling, 2010:Nares Strait hydrography and salinity field from a 3-year moored array. J. Geophys. Res., 115, C07 010, doi:10.1029/2009JC005966.

Rabe, B., et al., 2013: Liquid export of Arctic freshwatercomponents through the Fram Strait 1998-2011. Ocean

Science, 9 (1), 91–109.

Rignot, E. and K. Steffen, 2008: Channelized bottom melt-ing and stability of floating ice shelves. Geophys. Res.

Let., 35 (2), L02 503, doi:10.1029/2007GL031765.

Rigor, I. G., J. M. Wallace, and R. L. Colony, 2002: Re-sponse of sea ice to the Arctic oscillation. J. Clim.,15 (18), 2648–2663.

Sadler, H., 1976: Water, heat, and salt transport throughNares Strait, Ellesmere Island. J. Fish. Res. Board Can.,33, 2286–2295.

Samelson, R. and P. Babour, 2008: Low-level windsin Nares Strait: a model-based mesoscale climatology.Mon. Weather Rev., 136, 4746–4759.

Serreze, M. C. and J. A. Francis, 2006: The Arctic ampli-fication debate. Clim. Change, 76 (3-4), 241–264.

Steffen, K. and A. Schweiger, 1991: NASA team algorithmfor sea ice concentration retrieval from Defense Mete-orological Satellite Program Special Sensor MicrowaveImager - comparison with Landsat satellite imagery. J.

Geophys. Res., 96 (C12), 21 971–21987.

Stroeve, J. C., M. C. Serreze, M. M. Holland, J. E. Kay,J. Malanik, and A. P. Barrett, 2012: The Arctic’s rapidlyshrinking sea ice cover: a research synthesis. Climatic

Change, 110 (3-4), 1005–1027.

Sundby, S. and K. Drinkwater, 2007: On the mechanismsbehind salinity anomaly signals of the northern NorthAtlantic. Progr. Oceanogr., 73 (2), 190–202.

Timmermans, M. L., A. Proshutinsky, R. A. Krishfield,D. K. Perovich, J. A. Richter-Menge, T. P. Stanton,

9

and J. M. Toole, 2011: Surface freshening in the Arc-tic Ocean’s Eurasian Basin: An apparent consequenceof recent change in the wind-driven circulation. J. Geo-

phys. Res., 116, C00D03.

Visbeck, M. and J. Fischer, 1995: Sea-surface conditionsremotely-sensed by upward-looking ADCP. J. Atmos.

Oceanic Technol., 12 (1), 141–149.

Wekerle, C., Q. Wang, S. Danilov, T. Jung, and J. Schroter,2013: The Canadian Arctic Archipelago throughflow in amultiresolution global model: Model assessment and thedriving mechanism of interannual variability. J. Geophys.

Res., 118 (9), 4525–4541.

Wessel, P., 2009: A general-purpose Green’s function-based interpolator. Computers & Geosciences, 35 (6),1247–1254.

White, D., et al., 2007: The Arctic freshwater system:Changes and impacts. J. Geophys. Res., 112 (G4),G04S54, doi:10.1029/2006JG000353.

Woodgate, R. A., T. J. Weingartner, and R. Lindsay, 2012:Observed increases in Bering Strait oceanic fluxes fromthe Pacific to the Arctic from 2001 to 2011 and theirimpacts on the Arctic Ocean water column. Geophys.

Res. Let., 39, L24 603.

Wunsch, C. and P. Heimbach, 2006: Estimated decadalchanges in the North Atlantic meridional overturningcirculation and heat flux 1993-2004. J. Phys. Oceanogr.,36 (11), 2012–2024.

Yang, J. Y., 2005: The Arctic and subarctic ocean flux ofpotential vorticity and the Arctic Ocean circulation. J.

Phys. Oceanogr., 35 (12), 2387–2407.

Zweng, M. M. and A. Munchow, 2006: Warming andfreshening of Baffin Bay, 1916-2003. J. Geophys. Res.,111 (C7), C07 016, doi:10.1029/2005JC003093.

10

−100˚ −90˚−80˚

−70˚

−60˚

−50˚

−40˚

−30˚

−20˚

−10

˚

74˚

76˚

78˚

80˚

82˚

84˚

86˚

0 25 50 75 100Ice Concentration

%−100˚ −90˚

−80˚

−70˚

−60˚

−50˚

−40˚

−30˚

−20˚

−10

˚

74˚

76˚

78˚

80˚

82˚

84˚

86˚

2008

April−21

0˚

90˚

180˚

−90˚

0˚

90˚

180˚

−90˚

Fig. 1. Map of the study area off North Greenland withNares Strait in the center. The insert shows the the Arcticand adjacent North-Atlantic Oceans. Colors indicate seaice concentrations on Apr.-21, 2008 as seen from SSM/Isatellite while black contours are 500-m and 1000-m iso-baths.

11

−76˚ −74˚ −72˚ −70˚ −68˚ −66˚ −64˚ −62˚ −60˚ −58˚

78˚00'

78˚30'

79˚00'

79˚30'

80˚00'

80˚30'

81˚00'

81˚30'

82˚00'

82˚30'

0.2 m/s

Fig. 2. Nares Strait with 5-day filtered current vectorsat the mooring section in Kennedy Channel centered intime on April-21, 2008 overlaid on a MODIS/Terra imageof surface reflectance sensed that day. A low sun angleilluminates mountainous terrain of Ellesmere Island in thewest and Greenland in the east. Open water or thin ice isindicated by the black color near 74 W longitude

12

−76˚ −74˚ −72˚ −70˚ −68˚ −66˚ −64˚ −62˚ −60˚ −58˚

78˚00'

78˚30'

79˚00'

79˚30'

80˚00'

80˚30'

81˚00'

81˚30'

82˚00'

82˚30'

0.2 m/s

Fig. 3. As Figure 3, but for April-22, 2009. The blackcolor indicates Nares Strait to be covered by open water orthin ice due to a northern ice arch at 60 W longitude.

13

−70˚ −68˚ −66˚ −64˚

80˚00'

80˚15'

80˚30'

80˚45'

81˚00'

−300

−300

−300

−200

−200

−200

−70˚ −68˚ −66˚ −64˚

80˚00'

80˚15'

80˚30'

80˚45'

81˚00'

0 10 20

km

0.2 m/s

Ellesmere Island

Kane Basin

Greenland

−70˚ −68˚ −66˚ −64˚

80˚00'

80˚15'

80˚30'

80˚45'

81˚00'

−400

−300

−300

−300

−200

−200

−200

−70˚ −68˚ −66˚ −64˚

80˚00'

80˚15'

80˚30'

80˚45'

81˚00'

0 10 20

km

0.2 m/s

Ellesmere Island

Kane Basin

Greenland

Fig. 4. Record mean and principal components of verti-cally averaged ocean current vectors for the 2007-09 (top)and 2003-06 (bottom) periods. The co-ordinate system foralong- and across-channel directions is indicated with anorigin on the Ellesmere Island coastline. Contours are bot-tom depths between 100 m and 400 m. See Figure 2 forlocations.

14

Fig. 5. Mean current speed and direction near the cen-ter of Kennedy Channel (KS10) for the 2003-06 and 2007-09 periods at all available depths. Uncertainties represent95% confidence levels for speed; values at -16 m and -24 mare from interpolating measured velocities into the surfacelayer using ice velocity estimates.

15

Fig. 6. Low-pass filtered time series of wind 10 m aboveand ocean current vectors at 25, 75, and 115 m below thesurface as well as air temperature for the 2007-09 period.Wind and temperature are from model predictions (Samel-son and Babour 2008) while ocean currents are from stationKS10; see Figures 4 and 5 for locations.

16

Fig. 7. Ice velocity for the 2003-06 period off EllesmereIsland in the west (KS02), the center of Kennedy Channel(KS10), and off Greenland in the East (KS14). North isvertically up.

17

Fig. 8. Timeseries of ice velocity across Nares Strait fromwest (KS02) to east (KS14) with ice draft at KS20 for the2003 to 2006 deployment.

18

Fig. 9. Timeseries of ice velocity across Nares Strait fromwest (KS04) to east (KS10) with ice draft at KS30 for the2007 to 2009 deployment.

19

Fig. 10. Time-lagged correlation of wind- with ice velocityvector for the entire 2007 to 2009 deployment (top panel).Ice velocities are the across-channel average from 5 loca-tions. Negative direction Dir(Cxy) indicates that the icevector is rotated clock-wise relative to the wind vector. C2

xy

indicates the fraction of correlated variance. Bottom panelshows the time-lagged auto-correlation of along (black) andacross-channel (red) ice velocity.

20

−400

−300

−200

−100

0

Dep

th (

m)

30

31

32

33

34

35

33

34

34.4

34.6

34.835 2003−06

−400

−300

−200

−100

0

Dep

th (

m)

0 10 20 30

Distance (km)

30

31

32

33

34

3533

34

34.4

34.6

34.835 Aug.−2003

Fig. 11. Salinity across Kennedy Channel section of NaresStrait from moorings (top, record mean) and CTD profil-ing (bottom, snapshot) for 2003-06 deployment. Mooringlocations are indicated by black triangles while CTD profilelocations are by gray symbols. View is to the north withEllesmere Island on the left and Greenland on the right.

21

−400

−300

−200

−100

0

Dep

th (

m)

30

31

32

33

34

35

33

34

34.434.6

2007−09

−400

−300

−200

−100

0

Dep

th (

m)

0 10 20 30

Distance (km)

30

31

32

33

34

35

3334

34.4

34.6

Aug.−2007

Fig. 12. As Figure 11, but for the 2007-09 deployment.

22

Fig. 13. Freshwater (top) and volume (bottom) fluxthrough Nares Strait from Aug.-2003 through Jul.-2009 atdaily (gray) and monthly (red) time scales. The mean fluxfor 2003-06 and 2007-09 are indicated in red, no data werecollected from 2006-07.

23

Fig. 14. Correlations (black symbols) and regression (redline) of freshwater versus volume flux. Note the differentslope of 2003-06 (40±1.4 mSv/Sv) and 2007-09 (51±2.1mSv/Sv) regressions.

24

−400

−200

0

Dep

th (

m)

−100

1020304050

0

10

2003−06Volume: 0.71 Sv

−400

−200

0

Dep

th (

m)

0 5 10 15 20 25 30 35

Distance (km) from Canada

02003−06

Freshwater: 32 mSv −1

012345

Fig. 15. Record mean flow (top) and freshwater flux perunit area (bottom) in the along-channel direction definedas 120 degrees clockwise from true East. View is to thenorth with Greenland on the right and Canada on theleft. Black and grey symbols indicate mooring locationsfor 2007-09 and 2003-06 periods of observations, respec-tively. Contours above the thick black line at z=-30 metersseparate direct observations from estimated values usingEkman layer function fitting. Positive flow is towards thesouth-west.

25

−400

−200

0

Dep

th (

m)

−100

1020304050

−1000

10102030

40

2007−09Volume: 1.03 Sv

−400

−200

0

Dep

th (

m)

0 5 10 15 20 25 30 35

Distance (km) from Canada

0

000

2

2007−09Freshwater: 54 mSv

−1012345

Fig. 16. Record mean flow (top) and variability (bottom)in the along-channel direction defined as 120 degrees clock-wise from true East. View is to the north with Greenlandon the right and Canada on the left. Black and grey sym-bols indicate mooring locations for 2007-09 and 2003-06periods of observations, respectively. Contours above thethick black line at z=-30 meters separate direct observa-tions from estimated values using Ekman layer functionfitting. Positive flow is towards the south-west.

26

Fig. 17. Volume and freshwater (with symbols) flux esti-mates in the surface layer normalized by the surface value.The upper limit of the sectional integral is the depth forwhich the integrated flux estimate is shown.

27

Fig. 18. Regression of along-channel pressure differencewith volume flux. The regression is derived for the 2003/04data and applied to the 2005/06 data.

28

Fig. 19. Correlation (r2=0.68) and regression of along-channel pressure difference dP with volume flux Q.

29

Table 1. Mooring locations and records. Distance is mea-sured from the coast of Ellesmere Island.

Name Year Longitude, W Latitude, N Depth, m Distance, km Bins Record, days

KS02 2003-06 68.8744 80.5538 302 2.2 32 1103KS10 2003-06 67.9296 80.4388 299 23.9 32 1105KS12 2003-06 67.6709 80.4092 263 29.6 27 1104KS14 2003-06 67.4458 80.3884 157 34.4 15 1103KS04 2007-09 68.7393 80.5378 366 5.2 39 718KS06 2007-09 68.4555 80.5038 358 11.7 38 718KS08 2007-09 68.1850 80.4715 356 17.8 38 719KS10 2007-09 67.8930 80.4355 294 24.6 31 719KS12 2007-09 67.5927 80.3993 228 31.5 23 717

Table 2. Basic statistics of depth-averaged currents.Mean directions and ellipse orientations are in degrees pos-itive counter-clockwise from true east. The degrees of free-dom (dof) are determined as the ratio of the record lengthT and and integral decorrelation time scale Td.

Name km Year Mean Speed Direction Major/Minor Axis Orientation dof

KS02 2.2 2003-06 6.7±0.7 -126 6.6/1.0 -123 290KS04 5.4 2007-09 7.9±1.1 -122 6.8/2.0 -121 141KS06 11.7 2007-09 13.4±2.0 -117 7.9/3.6 -112 59KS08 17.8 2007-09 9.7±1.4 -120 6.8/3.9 -114 92KS10 23.9 2003-06 6.6±0.9 -123 7.3/3.1 -124 274KS10 24.6 2007-09 7.4±0.9 -124 6.9/4.2 -120 235KS12 29.6 2003-06 4.8±1.1 -108 10.0/3.2 -127 274KS12 31.5 2007-09 1.1±1.2 -91 13.4/3.1 -125 331KS14 34.4 2003-06 4.4±1.1 63 13.6/2.1 -119 548

30

Table 3. Record-mean flux estimates for volume Q andfreshwater Qf . Uncertainties are 95% confidence limits;uncertainties in brackets include a systematic error due toa 20% uncertainty in surface ice velocity estimation. Valuesfor Lancaster Sound and Davis Strait are from 1998 to 2006(Peterson et al. 2012) and from 2004 to 2010 (Curry et al.2014).

Year Q, Sv Qf Ocean, mSv Qf Ice, mSv

Nares Strait 2003-06 0.71±0.06 (±0.09) 32±2.5 (±5.7) 6.7±0.33 (±1.7)Nares Strait 2007-09 1.03±0.07 (±0.11) 54±3.9 (±9.3) 9.8±0.25 (±2.2)Lancaster Sound 0.53±0.13 32±6.0 2.1Davis Strait 2004-10 1.6±0.2 93±6 10±1

31