On the waters upstream of Nares Strait, Arctic Ocean, from 1991 to 2012

Continental Shelf Research 73 (2014) 83–96

Contents lists available at ScienceDirect

Continental Shelf Research journal homepage: www.elsevier.com/locate/csr

Research papers

On the waters upstream of Nares Strait, Arctic Ocean, from 1991 to 2012 Jennifer M. Jackson a,n, Camille Lique b,1, Matthew Alkire a, Michael Steele a, Craig M. Lee a, William M. Smethie c, Peter Schlosser c a

Applied Physics Laboratory, University of Washington, Seattle, WA, USA Joint Institute for the Study of the Atmosphere and Ocean, University of Washington, Seattle, WA, USA c Lamont-Doherty Earth Observatory, Earth Institute at Columbia University, Palisades, NY, USA b

art ic l e i nf o

a b s t r a c t

Article history: Received 6 June 2013 Received in revised form 29 October 2013 Accepted 28 November 2013 Available online 6 December 2013

The Lincoln Sea is a bifurcation point, where waters from the Canadian and Eurasian Basins flow to Nares or Fram Strait. Mechanisms that control which waters are found in the Lincoln Sea, and on its continental shelves, are unknown. Using conductivity–temperature–depth (CTD; from hydrographic and ice-tethered profiler surveys), nutrient, and mooring data with the DRAKKAR global 3-D coupled ocean/sea-ice model, the Lincoln Sea was examined from 1991 to 2012. Although both Pacific and Atlantic waters were observed on the North Ellesmere and North Greenland shelves, Atlantic water was shallower on the North Greenland shelf. Thus, deeper than 125 m, water was warmer and saltier on the North Greenland shelf than the North Ellesmere shelf. Three different water types were identified on the North Ellesmere shelf – waters from the Canadian Basin were observed 1992, 1993, 1996, 2005, and 2012, waters from both the Canadian and Eurasian Basins were observed in 2003, 2004, and 2008, and waters with no temperature minima or maxima below the surface mixed layer were observed in 1991, 2006, 2009, and 2010. Mixing with vertical advection speeds of 1
10  4 m s  1 were observed on the continental slope and this mixing could cause the disappearance of the temperature maxima. Model results suggest that currents on the North Ellesmere shelf were weak (less than 10 cm s  1), baroclinic, and directed away from Nares Strait while currents on the North Greenland shelf were stronger (less than 15 cm s  1), and primarily directed towards Nares Strait. CTD, mooring, and model results suggest that the water advected to Nares Strait is primarily from the North Greenland shelf while water on the North Ellesmere shelf is advected westward. & 2013 Elsevier Ltd. All rights reserved.

Keywords: Arctic Ocean Lincoln Sea Nares Strait Ocean circulation Water mass modification

1. Introduction Water from the Arctic Ocean flows to the North Atlantic Ocean through Fram Strait and the Canadian Arctic Archipelago (CAA) (Aagaard and Carmack, 1989). The Arctic Ocean is fresher than the North Atlantic, so water exported from the Arctic has the potential to influence the global freshwater cycle (Dickson et al., 1988) and meridional overturning circulation (Dickson et al., 1996; Koenigk et al., 2007; Rennermalm et al., 2007). Nares Strait is one of the two main passages through the CAA and year-round mooring data have estimated that the volume flux through Nares Strait ranges from 0.47 70.05 Sv (Rabe et al., in press) to 0.577 0.09 Sv (Münchow and Melling, 2008), or roughly half of the total CAA transport (McGeehan and Maslowski, 2012). Nares Strait separates northern Ellesmere Island from northern Greenland. At its northern end, Robeson Channel, Nares Strait is about n Corresponding author. at: ASL Environmental Sciences Inc., Victoria, British Columbia, Canada. E-mail address: [email protected] (J.M. Jackson). 1 Now at: Department of Earth Sciences, University of Oxford, Oxford, UK.

0278-4343/$ - see front matter & 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.csr.2013.11.025

35 km wide and about 500 m deep (Fig. 1). At the southern end of Nares Strait, about 530 km south of Robeson Channel, Smith Sound separates Nares Strait from Baffin Bay. Nares Strait is covered yearround by sea-ice. Despite its relatively narrow width, the circulation through Nares Strait is complex. Based on a 3-year mooring study across southern Robeson Channel, Rabe et al. (2010) found that flow was strongest on the Ellesmere Island side of Nares Strait, with average speeds of 0.2 m s  1 southward. Circulation varied with sea-ice conditions – under mobile ice, a second southward current was often observed in the middle of Robeson Channel while under land-fast ice a single southward current against Ellesmere Island was observed (Rabe et al., in press). It is thought that variability in the southward flow through Nares Strait can be explained by the sea surface height in northern Baffin Bay (Münchow et al., 2006; Houssais and Herbaut, 2011; McGeehan and Maslowski, 2012; Rudels, 2012), wind speed and direction under mobile ice conditions (Samelson et al., 2006; Münchow et al., 2006; Rabe et al., in press), and tides (Münchow and Melling, 2008). Hydrographic data collected in Nares Strait in the 1970s and 1980s showed that all of the water in Robeson Channel was from the Arctic Ocean (Sadler, 1976; Bourke et al., 1989). Both Pacific and

84

J.M. Jackson et al. / Continental Shelf Research 73 (2014) 83–96

Fig. 1. Bathymetric map of the Lincoln Sea. Regions discussed in the text are labeled in the figure. Bathymetric data are from version 2.23 of the International bathymetirc chart of the Arctic Ocean (IBCAO; http://www.ngdc.noaa.gov/mgg/bathymetry/arctic/arctic.html). Colored symbols show the locations of the stations that were sampled each year. The large black ex on the North Ellesmere shelf marks the location mooring SY01, which collected data from 2008 to 2009. The large gray ex on the North Ellesmere shelf shows the location of mooring SY04, which collected data from 2010 to 2011. The large black ex on the North Greenland shelf shows the location of the virtual mooring used to calculate current speed and direction with the 3-D DRAKKAR model in Section 6. The colored lines mark the different regions defined as follows – the green lines show the North Ellesmere shelf (the region south of 841N and to the west of Robeson Channel), the blue lines show the continental slope (the region between 84 and 851N from 42 to 721W), and the red lines denote the North Greenland shelf (the region north of Robeson Channel from 83 to 841N and from 42 to 541W).

Atlantic water masses have been identified in Nares Strait (Jones and Eert, 2006; Münchow et al., 2007; Alkire et al., 2010). Recent studies have suggested that there is variability in the water mass composition that is advected through Nares Strait. For example, mooring data collected near the bottom at southern Robeson Channel showed a freshening and warming trend from 2003 to 2006 followed by a salinification and warming trend from 2007 to 2009 (Münchow et al., 2011). Thus, it is likely that the source of water that is advected through Nares Strait is variable. Despite the global importance of Nares Strait, very little is known about its upstream region, the Lincoln Sea. The Lincoln Sea encompasses a continental shelf that extends north from Ellesmere Island and Greenland and a continental slope that leads to the Lomonosov Ridge. A deep (about 500 m) trough, the northern extension of Robeson Channel, extends northeast from Nares Strait, along the Greenland coast. For this paper, we define the North Ellesmere shelf as the continental shelf west of Robeson Channel that extends north from Ellesmere Island to the continental slope. We define the North Greenland shelf as the continental shelf that is north and east of Robeson Channel between 83–841N and 42–541W. The presence of the trough makes the shelf north of Greenland narrower than the shelf north of Ellesmere Island. The Lomonosov Ridge separates the Canadian Basin (Canada and Makarov Basins) from the Eurasian Basin (Amundsen and Nansen Basins). Hydrographic surveys of the Lincoln Sea have been sparse. A single temperature and salinity profile was collected on the North Ellesmere shelf in June 1967 (Seibert, 1968). As part of the ICESHELF project, an examination of springtime temperature and salinity profiles from 1989 to 1994 showed that waters on the North Ellesmere shelf and continental slope had similar characteristics to those in the Canada Basin (Newton and Sotirin, 1997). Based on the 1991–1996 ICESHELF data, Steele et al. (2004) suggest that

Canada Basin water was advected to the Lincoln Sea via the Transpolar Drift from the Chukchi Sea. There are no known observational studies on the pathway of water from the Lincoln Sea to Nares Strait. Using hydrographic, nutrient, and mooring data together with the DRAKKAR 3-D coupled ocean/sea-ice model and year-round conductivity–temperature–depth (CTD) data from an ice-tethered profiler, we examine the hydrography of waters on the continental shelf and slope of the Lincoln Sea to understand the source of waters that are transported to Nares Strait. In Section 2 we present a description of the data, the methods of nutrient analyses, and a description of the 3-D model used to estimate circulation. The spatial variability of the Lincoln Sea will be examined in Section 3. In Section 4, the interannual variability of the North Ellesmere shelf will be examined using data from 1991 to 2012. In Section 5 we discuss how mechanisms such as vertical diffusive heat flux and upwelling can modify water masses in the Lincoln Sea. In Section 6, water properties from Nares Strait will be compared with those from the North Greenland and North Ellesmere shelves and results from the 3-D model will be presented to examine the source of water to Nares Strait.

2. Data and methods 2.1. CTD data Temperature and salinity CTD data were collected as a part of the ICESHELF project between 1991 and 1996 and as a part of the Switchyard project from 2003 to 2012 using instruments reported in Table 1 (Fig. 1). All samples were collected in spring (late April to mid-May) when the study area was fully ice-covered and

J.M. Jackson et al. / Continental Shelf Research 73 (2014) 83–96

unreachable by icebreaker. Data were acquired by helicopter from 1991 to 2006 and by a ski-equipped DeHavilland DHC-6 aircraft, commonly referred to as a Twin Otter, from 2005 to 2012. 2.2. Mooring data Year-round temperature and salinity data were collected from moorings using instruments reported in Table 2. Two years of data

Table 1 Data collection dates and information on CTD and expendable CTD (AXCTD) data collected during each field campaign. The number of casts (No.casts) denotes the number within the study area for this paper (south of 851N from 42 to 721W). NS stands for the citation Newton and Sotirin (1997). Jea14 stands for this paper. The unit for temperature precision (Temp precision) is degrees Celsius. Year

Date range

No. casts

References CTD make

CTD model

Temp precision

OS100

0.02

OS100

0.02

OS100

0.02

OS100

0.02

OS100

0.02

SBE-19 Plus SBE-19 Plus SBE-19 Plus SBE-19 Plus SBE-19 Plus SBE-19 Plus SBE-19 Plus SBE-19 Plus AXCTD

0.005

1991

9

NS97

1992

11

NS97

1993

6

NS97

1994

2

NS97

1996

2

NS97

2003 6–8 May

10

Jea14

Ocean Sensors Ocean Sensors Ocean Sensors Ocean Sensors Ocean Sensors Seabird

2004 1–5 May

11

Jea14

Seabird

8

Jea14

Seabird

16

Jea14

Seabird

2

Jea14

Seabird

2008 27 April–7 May 2009 3–18 May

12

Jea14

Seabird

9

Jea14

Seabird

2010 5–17 May

12

Jea14

Seabird

2010 7 May

6

Jea14

2011 28 April– May 12 2011 7 May

6

Jea14

TsurumiSeiki Seabird

6

Jea14

2012 3–19 May

9

Jea14

2005 29 April–5 May 2006 10–15 May 2007 2–7 May

TsurumiSeiki Seabird

SBE-19 Plus AXCTD SBE-19 Plus

85

were collected from these moorings – the first from 27 April 2008 to 16 May 2009 and the second from 7 May 2010 to 9 May 2011. For 2008–2009, mooring SY01 was located near the edge of the North Ellesmere continental shelf at a depth of 274 m. For 2010– 2011, mooring SY04 was located about 17 km southwest of mooring SY01 in water that was 202 m deep. Each mooring was composed of six thermistors located 15–95 m apart and two salinity sensors located 55–60 m apart. A Twin Otter airplane was used to deploy the moorings in 2008 and 2010 and the recorded data were downloaded acoustically the following springs. 2.3. Ice-tethered profile data Data collected from Ice Tethered Profiler number 15 (ITP15) between 11 September 2007 and 9 October 2008 were downloaded from the project website at http://www.whoi.edu/itp (Krishfield et al., 2008; Toole et al., 2011). ITP15 was deployed in the Makarov Basin at 86.71N, 177.31E and drifted southeast to the Lincoln Sea. Although ITP15 eventually traveled through Nares Strait, it stopped collecting data when it reached the North Ellesmere continental slope at 84.21N, 61.41W. ITP15 was eventually advected through Nares Strait and was retrieved south of Nares Strait in August 2010. Uncertainties in the CTD data are estimated at 0.005 salinity units and 0.005 1C (Krishfield et al., 2008). Vertical profiles were collected about three times daily between about 7 and 800 m with a vertical resolution of 1 m. 2.4. Nutrient data and water-type analysis

0.005 0.005 0.005

0.005 0.005 0.005 0.1 0.005 0.1 0.005

From 2005 to 2011, discrete seawater samples were collected for chemical analysis during some CTD casts. The CTD/Rosette consisted of 4-bottle sample modules that were stacked over a CTD (Smethie et al., 2011). The volume of each of the water bottles was 4 l. At the end of each cast, the sample modules were placed in coolers and the end caps, which were connected by an epoxy coated stainless steel spring inside the bottle, were secured externally so they could not open. Bags of snow were placed into coolers to provide thermal stability close to 0 1C and the modules were returned to the base camp for sub sampling on the day of acquisition. Water samples for nutrients (NO3 , NO2 , NH4þ , PO34  , and Si (OH)4), salinity, stable oxygen isotopes (δ18O), and dissolved oxygen were drawn into appropriate containers. Oxygen and salinity samples were measured at the base camp with an automated Winkler titration system, and an Autosal salinometer standardized with IAPSO standard seawater, respectively. Oxygen isotopes were analyzed by the CO2 equilibration method on a Finnegan Mat 251 mass spectrometer at

Table 2 Information on the moorings used to collect year-round data from April 2008 to May 2009 and from May 2010 to May 2011. Temperature, salinity, and pressure data were collected every 30 min. The terms avg temp (std) and avg sal (std) stand for the average temperature (1C) and average practical salinity (with standard deviation), respectively. The temperature precision for the Seabird SBE37-IM and the Seabird SBC39-IM is 0.002 1C. The conductivity precision for the Seabird SBE37-IM is 0.0003 S m  1, or about 0.008 practical salinity units.The bottom depth was 274 m for the 2008–2009 mooring and 202 m for the 2010–2011 mooring. Date range

Location

Instrument

Depth

Avg temp (std)

Avg sal (std)

27 April 2008–16 May 2009

83.731N 65.181W

SBE37-IM SBE39-IM SBE39-IM SBE37-IM SBE39-IM SBE39-IM

55 m 70 m 85 m 110 m 160 m 255 m

 1.54 (0.1)  1.43 (0.1)  1.33 (0.1)  1.39 (0.1)  0.95 (0.1) 0.18 (0.1)

31.42 (0.4)

SBE37-IM SBE39-IM SBE39-IM SBE37-IM SBE39-IM SBE39-IM

34 m 54 m 74 m 94 m 144 m 194 m

 1.58 (0.1)  1.50 (0.1)  1.43 (0.0)  1.37 (0.0)  0.91 (0.1)  0.39 (0.1)

7 May 2010–9 May 2011

83.581N 64.951W

33.07 (0.5)

30.45 (0.4)

33.05 (0.2)

86

J.M. Jackson et al. / Continental Shelf Research 73 (2014) 83–96

Rutgers University by Rick Mortlock. Results were reported as δ18O relative to VSMOW. Nutrient samples were kept frozen until subsequent analysis at the Ocean Data Facility at Scripps Institute of Oceanography using an autoanalyzer and following standard JGOFS protocols. In this paper we use the chemical data to determine whether Pacific water was present when CTD data were collected. Fractional contributions of Pacific and Atlantic seawater, meteoric (precipitation, river runoff, and/or glacial melt), and net sea-ice melt were computed using two different approaches (POn4 parameter versus DIN:POn4 relationships) for distinguishing Pacific from Atlantic water and the results compared. In simplest terms, a sample of water collected anywhere from the Arctic Ocean water (Sobs) can be considered a mixture of one or more of four water types: Pacific water (PAC), Atlantic water (ATL), meteoric water (MW), and sea-ice melt (SIM). Assuming that the endmember values characterizing each of these water types are known (see Table 3), their fractional contributions to a given water sample can be computed from a set of coupled equations conserving mass (or volume), salinity, δ18O, and a third geochemical tracer (e.g., POn4 as discussed in Ekwurzel et al., 2001): SPAC
f PAC þSATL
f ATL þSMW
f MW þSSIM
f SIM ¼ Sobs

ð1Þ

δ18 OPAC
f PAC þ δ18 OATL
f ATL þ δ18 OMW
f MW þ δ18 OSIM
f SIM ¼ δ18 Oobs ð2Þ PO4 nPAC
f PAC þ PO4 nATL
f ATL þ PO4 nMW
f MW þ PO4 nSIM
f SIM ¼ PO4 nobs

ð3Þ

f PAC þ f ATL þ f MW þ f SIM ¼ 1

ð4Þ

where f is the fraction of mass (or volume) contributed by a given component. Note that net sea-ice formation (formation exceeding melting) will generate a negative SIM fraction (f SIM o 0) as this process removes water from the liquid ocean and stores it as solid ice (Ostlund and Hut, 1984). Various authors have used this approach to quantitatively track changes in the distribution of water types throughout the Arctic Ocean (e.g. Guay and Falkner, 1997; Macdonald et al., 1999; Taylor et al., 2003; Yamamoto-Kawai et al., 2005; Bauch et al., 2009; Alkire et al., 2010; Bauch et al., 2011). However, the combination of chemical tracers and their specified endmember values vary somewhat among the different studies. For example, Eqs. (1)–(4) exhibit the use of the POn4 parameter (POn4 ¼ PO4 þ ½O2 =1751:95), a quasi-conservative tracer that combines the biologically active (and therefore non-conservative) variables of phosphate and dissolved oxygen via the expected, stoichiometric relationship describing the production and aerobic respiration of organic matter (Broecker et al., 1985). Pacific waters entering the Arctic Ocean through the highly productive Bering and Chukchi Seas are marked by relatively high phosphate and silicic acid concentrations compared to Atlantic waters entering via Fram Stait and the Barents Sea (Jones and Anderson, 1986; Aagaard and Carmack, 1989). Thus, Pacific versus Table 3 Endmember assignments for water types considered in this study. Uncertainty in the assigned values were designated as 7 1 standard deviation from the associated mean or median with the exception of the meteoric water and sea ice melt δ18O values, which were adopted from Cooper et al. (2008) and Yamamoto-Kawai et al. (2005), respectively. POn4 endmember values were assigned based on values provided in Ekwurzel et al. (2001). Further details are provided in the text. Water type

Salinity

δ18O

POn4

Uncertainty

Pacific seawater Atlantic seawater Meteoric water Sea-ice melt

32.5 7 0.4 34.86 7 0.07 0 47 4

 0.8 7 0.3 0.247 0.09  18.8 7 3 0.3 7 1

2.4 7 0.3 0.707 0.05 0.17 0.1 0.4 7 0.2

0.19 0.19 0.03 0.03

Table 4 Locations for North Ellesmere shelf and slope stations where nutrient data were collected and analyzed from 2005 to 2009. These are different stations that are presented in Fig. 2. The length of the line lowered into the water in 2005–2009 was 600 m so the actual bottom depth was underestimated in 2005 and 2007. Year

Latitude (1N)

Longitude (1W)

Bottom depth (m)

2005 2006 2007 2008 2009

84.45 84.11 84.16 83.73 84.05

65.23 64.60 64.67 64.17 64.96

601 542 596 240 801

Atlantic contributions to the halocline can be distinguished using this method. However, due to the effect of air–sea exchange on the O2 concentration, the application of this tracer is restricted to deeper layers that have not been in recent contact with the atmosphere (i.e., below the mixed layer depth). A different group of geochemical tracers, based on the relationship of dissolved inorganic nitrogen to phosphate (DIN:P), has also been used to discern water types in the Arctic Ocean (e.g. Jones et al., 1998; Yamamoto-Kawai et al., 2008; Azetsu-Scott et al., 2013). This method exploits the deficit in DIN associated with Pacific water as a result of sedimentary denitrification on the Chukchi shelf, itself a consequence of the high rates of export production and the shallow depth of the shelves. The application of this method requires that relative Pacific and Atlantic water fractions be estimated from DIN:P relationships that characterize these waters (Yamamoto-Kawai et al., 2008): P PAC ¼ ðDIN obs þ3:072Þ=17:499

ð5Þ

P ATL ¼ ðDIN obs þ 11:306Þ=13:957

ð6Þ

f PAC ¼ ðP ATL  P obs Þ=ðP ATL  P PAC Þ

ð7Þ

f ATL ¼ 1  f PAC

ð8Þ

It is important to note that the DIN:P relationship associated with Atlantic waters are assumed to also be related to meteoric water and sea-ice melt, an assumption that is not necessarily correct (Jones et al., 1998; Taylor et al., 2003). Once these initial Pacific and Atlantic fractions are computed, they are used to define a weighted saline endmember (SE) that reflects the specific mixture of Pacific and Atlantic water in the sample. Then, a set of coupled equations similar to 1–4 are solved that incorporate this saline endmember: SSIM
f SIM þSMW
f MW þSSE
f SE ¼ Sobs

ð9Þ

δ18 OSIM
f SIM þ δ18 OMW
f MW þ δ18 OSE
f SE ¼ δ18 Oobs

ð10Þ

f SIM þ f MW þ f SE ¼ 1

ð11Þ

The final Pacific and Atlantic seawater fractions are computed by multiplying the initial fractions (f PAC and f ATL ) and the saline endmember fraction (fSE). While both the POn4 and the DIN:P method for defining end-members have been used in the literature, Bauch et al. (2011) have recently shown that denitrification on Siberian shelves is likely to complicate the estimation of Pacific fractions using the DIN:P method, resulting in the overestimation of Pacific water. In this study, Pacific fractions were computed using both endmember methods (Fig. 2). The Pacific fractions differed considerably between the two methods, particularly in the upper halocline region where the DIN:P method returned Pacific fractions 4 0:9 (Fig. 2d). Since we cannot rule out local denitrification on the Lincoln Sea shelf and influence from Siberian shelf waters is a significant possibility in this region (Steele et al., 2004), the POn4 method may be more reliable.

J.M. Jackson et al. / Continental Shelf Research 73 (2014) 83–96

87

Fig. 2. Plots of (a) silicic acid concentrations, (b) dissolved oxygen concentrations, (c) the ratio of NO/PO, (d) Pacific fractions calculated using the POn4 method, (e) Pacific fractions calculated using the DIN:P method, and (f) the difference in Pacific water fractions between the two methods (DIN:P minus POn4 ) for the Lincoln Sea shelf region during annual occupations in 2005 (black circles), 2006 (blue squares), 2007 (red diamonds), 2008 (green triangles), and 2009 (magenta dots). The location of the stations where the nutrient data were collected each year are described in Table 4.

In order to estimate the effect of uncertainty in the assigned endmember values on the calculated water type fractions, the assigned values are randomly varied within the range of uncertainty given in Table 3 and the resulting water type fractions compared. Although we have adopted the POn4 method to estimate Pacific water fractions, the choice of the method for estimating Pacific water influence (DIN:P vs. POn4 ) is the largest source of uncertainty in the calculations. We have therefore assigned the absolute uncertainty in the Pacific and Atlantic water fractions as the median difference between Pacific fractions (7 0.19) calculated using the two different methods. Similarly conservative estimates were assigned as the absolute uncertainties to the MW and SIM fractions (Table 3). To confirm the presence of Pacific water, silicic acid (Fig. 2a) and NO/PO ratios (Fig. 2c) were examined. Previous studies have shown that Pacific water can be characterized by low ratios of NO/PO (Wilson and Wallace, 1990) and high concentrations of silicic acid (McLaughlin et al., 2004) so results from the chemical analyses confirm that Pacific water was present on the North Ellesmere shelf from 2005 to 2009. 2.5. Numerical simulation The global ORCA025 coupled ocean/sea-ice model configuration developed by the Drakkar project (The Drakkar Group, 2007) is used to study circulation in the Lincoln Sea. An overall description of the model and its numerical details is given in Barnier et al. (2006). This model configuration uses a global tripolar grid with 1442
1021 grid points and 75 vertical levels. Vertical grid spacing is finer near the surface (1 m) and increases with depth to 200 m at the bottom. Horizontal resolution is 10–15 km in the Arctic Ocean, and  12 km in the Lincoln Sea and Nares Strait. The ocean/sea-ice code is based on the NEMO framework version 3.2. (Madec, 2008). The sea-ice model is the Louvain-laNeuve model (LIM2), which is a dynamic–thermodynamic model

specifically designed for climate studies. A detailed description is given in Timmerman et al. (2005). The simulation is interannual and runs from 1958 to 2010 with no spin-up. Initialization is done using data from the Polar Science Center Hydrographic T/S Climatology (PHC; see Steele et al., 2001 for details). For the first part of the experiment (1958–1988), the forcing dataset used is DFS4 (Drakkar Forcing Set V4), which is fully described in Brodeau et al. (2010). After 1988, the forcing dataset is based on ERAinterim. Corrections are applied to air temperature and humidity at 2 m, using the POLES climatology, following the method described in Brodeau et al. (2010). Largescale shortwave radiation fields are also corrected towards the GEWEX climatology, and the precipitation fields are corrected towards the GPCP fields. A full description of the simulation and its forcing dataset can be found in a technical report (available on line at http://www.drakkar-ocean.eu). In the model, the mean volume transport through Nares Strait is  1.1 Sv, and thus overestimates the 0.47–0.57 Sv observed transport (Münchow and Melling, 2008; Rabe et al., in press). An overestimated Nares Strait transport of 0.7–0.8 Sv has been found in other models (McGeehan and Maslowski, 2012; Rasmussen et al., 2011; Rudels, 2012). To evaluate the DRAKKAR model for the North Ellesmere shelf, model results were compared to mooring observations collected from 2008 to 2009 (Fig. 3). Since no current data were available from the mooring, temperature and salinity data were used to evaluate the model. The monthly average modeled temperature was similar (less than 0.2 1C different) to the observations from 55 to 85 m and at 255 m. At 110 m and 160 m, the modeled temperature was 0.5–0.9 1C warmer than the observed temperature. Conversely, the modeled salinity was similar to the observed values at 110 m and up to 1.25 salinity units saltier than the observed salinity at 55 m. Thus, the DRAKKAR model overestimated salinity near the surface and heat in the pycnocline. The greatest overestimation of heat and salt occurred at the beginning of 2009, when freshwater was advected to the Lincoln Sea from the Beaufort Gyre (Timmermans

88

J.M. Jackson et al. / Continental Shelf Research 73 (2014) 83–96

Fig. 3. An comparison between results from the DRAKKAR 3-D coupled ocean/sea-ice model and observations from the SY01 mooring located on the North Ellesmere shelf at 83.731N, 65.021W. Here we compare the (a) monthly averaged observed temperature to the modeled temperature at 55 m, 70 m, 85 m, 110 m, 160 m, and 255 m and (b) monthly averaged observed salinity to the modeled salinity at 55 m and 110 m. Solid lines represent observed results and dashed lines represent model output. The monthly averaged difference between the modeled and observed values is shown for (c) temperature and (d) salinity. For (c), positive values indicate that the model was warmer than observations. For (d), positive values indicate that the model was saltier than observations.

et al., 2011). In a recent comparison of 10 models and observations, Jahn et al. (2012) found that the modeled freshwater flux through Nares Strait ranged from 163 to 2463 km3 year  1 (compared to the observed freshwater flux of 788 km3 year  1), with the ORCA025 estimate at 1747 km3 year  1. Thus, while imperfect, the performance of ORCA025 is similar to other models in the region. More details about the representation of the Arctic exports through the CAA and Davis Strait can be found in Lique et al. (2009) and Jahn et al. (2012).

3. Spatial variability on the continental shelves and slope An initial examination of CTD data showed spatial variability throughout the Lincoln Sea. To understand how the different water properties were linked to different regions, profiles from the North Ellesmere shelf, the North Greenland shelf, and continental slope of the Lincoln Sea were compared (Fig. 4). During Switchyard, sampling of the North Greenland shelf commenced in 2009. Thus, only profiles from 2009 to 2012 were examined in this section. Similar to Newton and Sotirin (1997), we observed properties that resembled waters from the Canada Basin in all regions. Waters from the Canada Basin have been described as follows. At the surface is a seasonal mixed layer whose depth is 30–50 m (Coachman and Barnes, 1961; McLaughlin et al., 2004). Below the surface mixed layer is a temperature maximum of Pacific origin modified in the Chukchi Sea during summer (Coachman and Barnes, 1961) called Pacific Summer Water (PSW). The temperature minimum below PSW is water of Pacific origin modified in the Chukchi Sea in winter (Coachman and Barnes, 1961) called Pacific Winter Water (PWW). Below PWW is a temperature maximum of Atlantic origin in the approximate depth

range of 300–900 m (Coachman and Barnes, 1961) called Atlantic water. Below the Atlantic water is colder deep water. From 2009 to 2012, PSW was evident on the continental slope and on the North Greenland shelf. On the continental slope, a temperature minimum at the salinity 34 was observed. This water was likely the cold halocline that forms in the eastern Nansen Basin (Steele and Boyd, 1998) from heat loss to the atmosphere, lateral mixing between two types of Atlantic water, and the admixture of cold, fresh shelf water (Steele and Boyd, 1998; Alkire et al., 2010; Rudels, 2012). The temperature–salinity diagram suggests that the water properties are similar, however the temperature and salinity profiles show that, deeper than about 125 m, water at any given depth was warmer and saltier on the North Greenland shelf than the North Ellesmere shelf. How can these differences be explained? We suggest that Atlantic water was closer to the surface on the North Greenland shelf than the North Ellesmere shelf, implying that Atlantic water is advected at a shallower depth onto the North Greenland shelf. Since Atlantic water is both higher in the water column and warmer in the Eurasian Basin than the Canadian Basin (Rudels, 2012), these results imply that the Atlantic water on the North Greenland shelf from 2009 to 2012 is from the Eurasian Basin while Atlantic water on the North Ellesmere shelf is from the Canadian Basin. From about 50 to 250 m, waters on the North Ellesmere shelf had a smaller range of values than either the North Greenland shelf or the continental slope. This smaller range of values implies that there is less variability in the North Ellesmere waters. Deeper than 200 m, waters on continental shelves were up to 0.5 1C cooler than on the continental slope. As suggested by Melling et al. (1984), the cooling of waters deeper than 200 m on the shelves could be attributed to the continual diffusion of heat from the Atlantic water to shallower waters.

J.M. Jackson et al. / Continental Shelf Research 73 (2014) 83–96

89

Fig. 4. A comparison of (a) temperature profiles, (b) salinity profiles, and (c) temperature–salinity diagrams observed in three different regions (see Fig. 1 for locations of regions). The symbols (green diamonds for the North Ellesmere shelf, red stars for the North Greenland shelf, and blue circles for the continental slope) indicate the total range of properties at given depths (for (a) and (b)) and at given salinities (for (c)). Only data collected from 2009 to 2012 were included in the temperature and salinity ranges because the North Greenland shelf was not sampled prior to 2009. To simplify the plot, only 1 profile from each region in 2009 was shown. The black lines indicate the freezing temperature and the gray lines show the density in 1 kg m  3 increments.

In the next section we examine data from 1991 to 2012 that were collected on the North Ellesmere shelf to explore interannual variability.

4. Interannual variability on the North Ellesmere shelf To investigate interannual variability, the properties at one location on the North Ellesmere shelf were examined from 1991 to 2012. This region was sampled almost every year and was located in the latitude range of 83.1–83.81N and the longitude range of 62.8–66.71W. The bottom depth of this sample region was 220–300 m. This region was sampled sometime between the end of April and the middle of May each year. Since the same location was not sampled each year, we compared nearby stations to ensure that spatial variability was smaller than interannual variability. In this analysis we compared the temperature at the salinity 33.5 and found that the temperature varied by a maximum of 0.22 1C for stations sampled in the same year and 0.45 1C for stations sampled in different years. To show interannual variability, CTD and water-type fractions computed from bottle chemistry data were plotted (Figs. 2 and 5). Beginning at the surface, we find that the surface mixed layer, whose base was defined following Jackson et al. (2012) as the first depth where density was at least 0.01 kg m  3 greater than the density at 8 m, was at the freezing temperature during all years because this region was completely ice-covered when sampled.

The surface mixed layer ranged from a minimum of 25 m in 1993 to a maximum of 60 m in 2004. The salinity of the surface mixed layer ranged from 32.3 in 1991 to 29.5 in 2009. As discussed by Timmermans et al. (2011) and de Steur et al. (2013), a shift in the atmospheric circulation likely transported fresh surface waters from the Beaufort Gyre to the North Ellesmere shelf in 2009. Between the base of the surface mixed layer and 100 m, and similar to Newton and Sotirin (1997), we observe several years (1992, 1993, 1996, 2003, 2004, 2005, 2008, and 2012) when there was a distinct temperature maximum within the Pacific Summer water (PSW) salinity range of 31–33. Similar to Newton and Sotirin (1997), we find that PSW was warmest in 1993. In the years 1991, 2006, 2009, and 2010, no temperature minima or maxima were observed below the surface mixed layer. Despite the lack of temperature minima or maxima, the Pacific water fractions, NO/PO ratios, and silicic acid concentrations derived from the bottle chemistry data (Fig. 2a, d, and e) support the interpretation that Pacific water contributions to the upper  100 m (in the salinity range 32–33.2) were relatively persistent. Therefore, upper halocline waters on the North Ellesmere shelf were primarily Pacific derived at least between 2005 and 2009. Deeper than about 100 m, and in waters that were saltier than 33, we observe three different temperature/salinity structures. The first, observed in 1992, 1993, 1996, 2005, and 2012, was characterized by a temperature minimum at about the salinity 33.1 followed by a temperature increase between salinities 33.1 and 34.8 (Fig. 5c). As shown by McLaughlin et al. (1996), the temperature increase

90

J.M. Jackson et al. / Continental Shelf Research 73 (2014) 83–96

Fig. 5. Water column properties for a station location at about 83.51N, 651W that was sampled in the years 1991–1993, 1996, 2003–2006, and 2008–2010. Each station had a bottom depth of 220–300 m. Here we show the (a) observed temperature, (b) practical salinity, and (c) temperature–salinity diagram with sigma-t contour lines. Red lines indicate stations with temperature extrema that are likely PSW and PWW, green lines show stations with cold halocline water in the salinity range 33.5–34, and blue lines indicate stations with no temperature extrema. The black lines in panels (a) and (c) show the minimum freezing temperature calculated. All stations were sampled sometime between the end of April and the middle of May. See Table 1 for information on the data collection.

between salinities 33.1 and 34.8 is a characteristic of Canada Basin waters. The second, observed in 2003, 2004, and 2008, showed a temperature increase between the salinity range of about 33.8–34.8, and these fresher cold halocline waters are likely from the eastern Nansen Basin (Steele and Boyd, 1998). The third, observed in 1991, 2006, 2009 and 2010, showed that temperature increased from the base of the surface mixed layer to salinity 34.8. In summary, the ICESHELF and Switchyard CTD datasets suggest that three different groups of water types can occupy the North Ellesmere shelf. The first, observed in 1992, 1993, 1996, 2005, and 2012, was typical of water from the Canada Basin with a PSW temperature maximum, a PWW temperature minimum, and Atlantic water at the bottom. The second, observed in 2003, 2004 and 2008, had Canada Basin characteristics (i.e. a PSW temperature maximum and a silicate maximum) in waters shallower than about 100 m and Eurasian Basin characteristics (i.e. a cold halocline and warm Atlantic water) in waters deeper than about 100 m. The third, observed in 1991, 2006, 2009, and 2010, had no obvious temperature extrema below the surface mixed layer. It is likely that at least the upper 100 m of this water is from the Canada Basin since high fractions of Pacific water were observed in 2006 and 2009. How Pacific and Atlantic water could be modified to remove the temperature extrema will be examined in Section 5.

5. Water mass modification on the continental slope of the Lincoln Sea During the years when there were no temperature minima or maxima below the surface mixed layer, water in the 31–33 salinity

range was cooler than PSW, water at the salinity 33.1 was warmer than PWW, and water at the salinity 33.8 was warmer than the cold halocline (Fig. 5), suggesting that heat is lost from PSW and Atlantic water to warm either PWW or the cold halocline. There are three places where this transfer of heat could occur – (i) in the Arctic Basins; (ii) over the continental slope; or (iii) on the continental shelves. The modification of water masses in the Arctic basins (Steele et al., 2004; Jackson et al., 2011) and on the continental shelves (Melling et al., 1984) has been addressed in the literature. Here we discuss the modification of water masses on the Lincoln Sea continental slope. To examine how the water properties could be modified over the continental slope, we analyzed data from ITP15 (Fig. 6). ITP15 was deployed in the Makarov Basin in September 2007 and was first observed in the Lincoln Sea in August 2008. Since ITPs are not stationary, the resulting observations vary in both space and time and can be complex to interpret. In this example, the ITP traveled at speeds of 0.05–0.20 m s  1 (not shown) so likely drifted through, not with, the slower moving water masses (that move at speeds of about 0.01 m s  1 Coachman and Barnes, 1961; Jackson et al., 2011). Periods of rapid warming, at depths of about 120–150 m, by up to 0.4 1C were observed from 27 September to 2 October and from 5 to 10 October as ITP15 drifted over the  1500 m isobath. Three finger-like intrusions of warm water that started on 5th October at 155 m and lasted 60 h to 134 m were observed. At the same time as these intrusions, ITP15 moved at speeds of 0.08–0.12 m s  1. If these features were upward-propagating, and not the advection of the ITP through a pre-existing feature, then we estimate a vertical advection speed of 1
10  4 m s  1, which is much faster than the monthly averaged upwelling rate (from Ekman transport)

J.M. Jackson et al. / Continental Shelf Research 73 (2014) 83–96

91

Fig. 6. Water column properties from ITP15, which collected data in the Lincoln Sea from 8 August to 10 October, 2008. Panel (a) shows the location of ITP15 in colored in 5 day intervals as ITP15 moves from the Lincoln Sea towards the North Ellesmere shelf. The thin black lines show the bathymetry from 200 to 1000 m in 200 m intervals and the thick black line marks the 1500 m isobath. The black ex marks the location of the 2008–2009 mooring SY01. The same colors are used to indicate the 5 day intervals in the temperature–salinity diagram in panel (b). For panel (b), the black line is the freezing temperature relative to a pressure of 0 db and the gray lines are sigma-t at 1 unit intervals. Panel (c) shows temperature in color with salinity represented by black lines. These results indicate upward mixing of warm Atlantic water beginning on September 27, when ITP15 was over about the 1500 m isobath. The black boxes in panel (c) show the two mixing events.

Fig. 7. Results from mooring SY01, which recorded (a) temperature and (b) salinity data from April 27, 2008 to May 16, 2009. Thermistors were located at 6 different depths (55 m, 70 m, 85 m, 110, 160 m, and 255 m) and salinity sensors were located at 2 depths (55 m and 110 m). Panel (c) shows the temperature at 160 m from September 16 to October 16, 2008. This is the same time that ITP15 was located on the Lincoln Sea continental slope (Fig. 6). The vertical gray lines in panel (a) show the time period shown in panel (c).

92

J.M. Jackson et al. / Continental Shelf Research 73 (2014) 83–96

of about 3
10  7 m s  1 found by Yang (2006) for the North Ellesmere shelf in October. Between 17 September and 10 October, water at the salinity 33.5 was warmed by about 0.2 1C, which is about 6–12 times faster than that can be explained by diffusive heat flux (Jackson et al., 2011). These results suggest that heating of PWW or the cold halocline could occur by mechanisms over the continental slope. It is likely that some mixing event caused heat from Atlantic water to be entrained into the shallower temperature minimum. Such an event could be storm-driven upwelling, however ITP data from 27 September to 2 October and from 5 to 10 October were not correlated with strong easterly winds from the NCEP/NCAR reanalysis (not shown). Throughout this time, images from the Cryosphere Today (http://www.arctic.atmos.uiuc. edu/cryosphere) show that the sea-ice concentration of the Lincoln Sea remained at 100% so it is unlikely that the mixing event was caused by ice-edge upwelling. To understand how the fall 2008 mixing event identified by ITP15 could influence waters on the North Ellesmere shelf, data from mooring SY01 were examined (Fig. 7). Of the six different depths where thermistors were deployed, the water at 160 m showed the greatest variability in temperature and this is because water at 160 m is between the cold halocline and the top of Atlantic water so that

any vertical motion can invoke large temperature changes (Fig. 6). Temperature fluctuations of up to 0.5 1C per day were common at 160 m, often occurring several times per month. Results from a power spectral density analysis show that there were energy peaks at periods of 12 h and 9, 10, and 14 days (not shown). The energy peaks at periods of 9 and 10 days were only observed at 160 m. While the energy at 12 h and 14 days can be explained by semidiurnal tides and the spring–neap tide cycle, the energy at periods of 9 and 10 days could be explained by a synoptic events. From September 25 to 30, the mooring temperature at 160 m warmed from  1.15 1C to  0.8 1C and this is similar to the temperature and timing of warming observed at 160 m by ITP15. Thus, it is possible that the event observed along the continental slope of the Lincoln Sea also influenced waters on the North Ellesmere continental shelf. The temperature fluctuations at the mooring at 160 m were least frequent from October to January, implying some seasonality to the events. In summary, mixing events along the continental slope likely cause heat from Atlantic water to be rapidly (speeds of about 1
10  4 m s  1) entrained into the shallower cold water. These mixing events occur several times per month and can explain the warmer Pacific Winter Water or cold halocline water that is often

Fig. 8. An examination of properties during the May 2010 field campaign. Here we show (a) the observed temperature, (b) the observed salinity, (c) a temperature–salinity diagram, and (d) a map of the station locations. Stations and profiles are colored as following – green stations are on the North Ellesmere shelf, red stations are on the North Greenland shelf, black stations are in Nares Strait, purple stations are on the northern end of Robeson Channel, and blue stations are on the continental slope in water that is deeper than 500 m. Data from May 2010 to May 2011 mooring, with its location marked by cyan ex, are also shown – the cyan circles in panels (a) and (b) represent the temperature and salinity range at each depth of the mooring, respectively. The cyan dots in (c) marks the temperature–salinity data points from the one year of mooring data at 34 m and at 94 m. The black dashed line in panels (a) and (c) is the freezing temperature relative to 0 db. The gray horizontal lines in panel (c) represent density at 1 kg m  3 intervals. To simplify the figure, not all profiles from Nares Strait or the continental slope are shown.

J.M. Jackson et al. / Continental Shelf Research 73 (2014) 83–96

observed on the North Ellesmere shelf. Mechanisms that cause these mixing events are an important topic for future research.

6. Transport to Nares Strait To examine the source of waters that are transported to Nares Strait, CTD data from 2010 were analyzed. During this year, Nares Strait data were collected within a few days of data collected from the North Greenland shelf, North Ellesmere shelf, and the continental slope of the Lincoln Sea. In addition, mooring SY04 collected temperature and salinity data year-round from May 2010 to 2011, and model output were available for 1990–2010. By jointly examining water at Nares Strait and the upstream regions, sources and pathways of water to Nares Strait can be assessed. In 2010, six stations were sampled across Nares Strait, one was sampled on the North Greenland shelf, one on the northern end of Robeson Channel, one on the North Ellesmere shelf, and three on the Lincoln Sea continental shelf (Fig. 8). In waters deeper than about 75 m, the temperature and salinity of waters in Nares Strait were most similar to waters on the North Greenland shelf and at the northern end of Robeson Channel. For example, water at five of the six Nares Strait stations were too warm and salty at 150 m to be from the North Ellesmere shelf. Results from mooring SY04, that sampled from May 2010 to May 2011, confirm that water on the North Ellesmere shelf at 94 m was always fresher than the water at Nares Strait at 94 m. In contrast, a temperature–salinity diagram shows that properties on the North Ellesmere shelf, on the North Greenland shelf, and in Nares Strait were similar even though there was a depth mismatch between the North Ellesmere shelf and Nares Strait. The differences between the two shelves appear to be linked to the observation that Atlantic water was colder and deeper on the North Ellesmere shelf than the North Greenland shelf. Thus, if North Ellesmere shelf water flowed to Nares Strait, the Atlantic water would have to be warmed and

93

upwelled to a shallower depth somewhere between the North Ellesmere shelf and Nares Strait. To understand the possible pathway of water to Nares Strait in 2010, the annual average modeled velocity was calculated using the DRAKKAR global 3-D coupled ocean/sea-ice model at four different depths – the surface, 50 m, 100 m, and 200 m (Fig. 9). Model results suggest that, in general, circulation was much weaker on the North Ellesmere shelf than on the North Greenland shelf. Flow on the North Ellesmere shelf was baroclinic and directed away from Nares Strait, with weak ( 5 cm s  1) westward currents at the surface, stronger ( 5–10 cm s  1) northwestward currents at 50 m, and very weak ( 1 cm s  1) currents at 100 m and 200 m. On the continental slope north of Ellesmere Island, currents were very weak ( 1 cm s  1) from the surface to 50 m and stronger ( 10 cm s  1) and southeastward (towards Nares Strait) at 100 m and 200 m. Similar to Lique et al. (2010), Houssais and Herbaut (2011), and McGeehan and Maslowski (2012) we observe that flow on the North Greenland shelf was strong (up to  15 cm s  1), and directed towards Nares Strait. The flow appeared to follow bathymetry, traveling westward north of Robeson Channel and then turning southwestward over Robeson Channel. On the continental slope of the North Greenland shelf, currents shallower than 100 m were relatively weak ( 5 cm s  1) and directed towards the North Greenland shelf (southward) while the current at 200 m was stronger (up to 15 cm s  1) and directed towards Fram Strait (southeastward). Similar to the observations, results from the model suggest that water that flowed through Nares Strait was from the North Greenland shelf in 2010. To understand whether conditions in 2010 were typical on the North Greenland shelf, the DRAKKAR model was used to calculate the monthly average currents at one location (83.51N, 481W) from 1990 to 2010 (Fig. 10). With a few exceptions at 200 m, the currents at the virtual mooring location were always directed westward. Currents at the surface were most variable, and were directed both northwestward and southwestward throughout the study period. Below the surface, the currents were primarily northwestward.

Fig. 9. The average annual circulation calculated from the DRAKKAR 3-D global coupled ocean/sea-ice model in the Lincoln Sea in 2010 at the surface, at 50 m, at 100 m, and at 200 m. The light gray lines indicate the bathymetry. The large black ex marks the location of the virtual mooring whose results are presented in Fig. 10.

94

J.M. Jackson et al. / Continental Shelf Research 73 (2014) 83–96

Fig. 10. The monthly averaged current speed and direction for one location on the North Greenland shelf (83.51N, 481W) that was calculated by the DRAKKAR 3-D ocean/ sea-ice coupled model. Four depths are shown – 0 m, 50 m, 100 m, and 200 m. Each stick represents one monthly average. Here the angle of the stick represents the compass direction so that positive values are northwestward, northward, or northeastward and negative values are southwestward, southward, or southeastward. The location of this model point is marked as a black ex in Figs. 1 and 9.

Thus, results from the virtual mooring from 1990 to 2010 support model results from 2010, which show that water on the North Greenland shelf flows westward around the northern end of Robeson Channel. Once in Robeson Channel, model results from 1990 to 2010 show that the water above 200 m always flows southwestward towards Nares Strait (not shown). Model results at 83.51N, 481W also suggest a shift in the current speed and direction – from 1990 to 1997, currents below the surface were stronger and northwestward while from 1997 to 2010, currents were slower and had a more westward component. At 200 m, there were several current reversals between 1997 and 2010, with the current alternating between northwestward and southeastward. In summary, observational and model results suggest that the water properties and circulation on the North Ellesmere and North Greenland shelves are different. The North Ellesmere shelf is characterized by weak (less than 10 cm s  1), baroclinic flow that is directed away from Nares Strait. The North Greenland shelf is characterized by stronger (up to 15 cm s  1) flow that travels around the northern end of Robeson Channel and then towards Nares Strait. We conclude from observational data and model output that water in Nares Strait is likely from the North Greenland shelf and not the North Ellesmere shelf.

7. Conclusions The oceanography of the Lincoln Sea, Arctic Ocean is complex. Based on CTD, nutrient, mooring, and ITP data along with output

from the DRAKKAR 3-D coupled ocean/sea-ice model, we found that the oceanographic conditions of the North Ellesmere shelf are very different from those on the North Greenland shelf. North Ellesmere shelf currents could be characterized as weak (up to 10 cm s  1), baroclinic, and directed away from Nares Strait. It is likely that no water from the North Ellesmere shelf is advected towards Nares Strait. Three different water types were observed on the North Ellesmere shelf from 1991 to 2012. The first was water from the Canada Basin, where Pacific Summer Water (PSW) and Pacific Winter Water (PWW) were observed in waters shallower than Atlantic water. Water from the Canada Basin was observed in 1992, 1993, 1996, 2005, and 2012. The second water type, observed in 2003, 2004, and 2008, was PSW above cold halocline and Atlantic water that were likely from the Eurasian Basin. The third water type, observed in 1991, 2006, 2009, and 2010, had no definitive temperature maxima or minima below the surface mixed layer. Despite the lack of temperature extrema, nutrient data show a high fraction of Pacific water in waters fresher than 33.5 from 2005 to 2009, suggesting that at least the upper 100 m was from the Canada Basin. CTD, mooring, and ITP results show that the modification of waters that cause the temperature extrema to disappear could occur on the continental shelf. A mixing event along the continental slope was observed from ITP data and we estimate that heat from Atlantic water was entrained into the shallower cold water at a vertical advection rate of 1
10  4 m s  1. Mixing by vertical diffusive heat flux is likely a dominant processes in the basins and this is a much slower process (about 1–3
10  6 m s  1; Jackson et al., 2011) than the vertical advection rated observed on the continental slopes. North Greenland shelf currents can be characterized as stronger (up to 15 cm s  1) and directed towards Nares Strait while currents on the North Ellesmere shelf are weaker, baroclinic, and directed away from Nares Strait. The observational data suggest that water in the upper 100 m on the North Greenland shelf is primarily from the Canadian Basin while water below 100 m is primarily from the Eurasian Basin (Fig. 5). Thus, water that is ultimately advected through Nares Strait originates in different parts of the Arctic Ocean. While it seems clear that water on the North Greenland shelf is advected to Nares Strait, the fate of water on the North Ellesmere shelf is less certain. Model results suggest that water on the North Ellesmere shelf is advected westward, towards the Canada Basin.

Acknowledgments For their help collecting Switchyard CTD data, we would like to thank Roger Andersen, Wendy Ermold, and Nicholas Michel-Hart from the Applied Physics Laboratory at the University of Washington, and Dale Chayes, Richard Perry, Robert Williams, and Ronny Friedrich from the Lamont-Doherty Earth Observatory at Columbia University. For their help collecting Switchyard mooring data, we would like to thank Jason Gobat, Adam Huxtable, and Jim Johnson from the Applied Physics Laboratory. The Ice-Tethered Profiler data were collected and made available by the Ice-Tethered Profiler Program based at the Woods Hole Oceanographic Institution (http://www.whoi.edu/itp). The NCEP Reanalysis data were provided by the NOAA/OAR/ESRL PSD, Boulder, Colorado, USA, from their Web site at http://www.esrl.noaa.gov/psd/. This study uses a numerical experiment carried out within the European DRAKKAR project. The simulation has been run at the IDRIS CNRS-GENCI computer center in Orsay, France, by R. Dussin. M. Steele, C. Lee, W. Smethie, and P. Schlosser acknowledge support from the Office of Polar Programs, National Science Foundation. We would like to thank two anonymous reviewers for their comments that greatly improved this paper.

J.M. Jackson et al. / Continental Shelf Research 73 (2014) 83–96

References Aagaard, K., Carmack, E.C., 1989. The role of sea ice and other freshwater in the Arctic circulation. J. Geophys. Res. 94, 14 485–14 498. Alkire, M.B., Falkner, K.K., Morison, J., Collier, R.W., Guay, C.K., Desiderio, R.A., Rigor, I.G., McPhee, M., 2010. Sensor-based profiles of the NO parameter in the central Arctic and southern Canada Basin: New insights regarding the cold halocline. Deep-Sea Res. I 57, 1432–1444. Alkire, M.B., Falkner, K.K., Boyd, T., MacDonald, R.W., 2010. Sea ice melt and meteoric water distributions in Nares Strait, Baffin Bay, and the Canadian Arctic Archipelago. J. Mar. Res. 68, 767–798. Azetsu-Scott, K., Petrie, B., Yeats, P., Lee, C., 2013. Composition and fluxes of freshwater through Davis Strait using multiple chemical tracers. J. Geophys. Res. 117, C12011, http://dx.doi.org/10.1029/2012JC008172. Barnier, B., Madec, G., Pendruff, T., Molines, J.M., Treguier, A.M., Le Sommer, J., Beckman, A., Biastoch, A., Böning, C., Dengg, J., Derval, C., Durand, E., Gulev, S., Remy, E., Talandier, C., Theetten, S., Maltrud, M., McClean, J., De Cuevas, B., 2006. Impact of partial steps and momentum advection schemes in a global ocean circulation model at eddy permitting resolution. Ocean Dyn. 56, 543–567, http://dx.doi.org/10.1007/s10236-006-0082-1. Bauch, D., Dmitrenko, I.A., Wegner, C., Holemann, J., Kirillov, S.A., Timokhov, L.A., Kassens, H., 2009. Exchange of Laptev Sea and Arctic Ocean halocline waters in response to atmospheric forcing. J. Geophys. Res. 114, http://dx.doi.org/10.1029/ 2008JC005062. Bauch, D., van der Loeff, M.R., Andersen, N., Torres-Vales, S., Bakker, K., Abrahamsen, E.P., 2011. Origin of freshwater and polynya water in the Arctic Ocean halocline in summer 2007. Prog. Oceanogr. 91 (4), 482–495. Bourke, R.H., Addison, V.G., Paquette, R.G., 1989. Oceanography of Nares Strait and Northern Baffin Bay in 1986 with emphasis on deep and bottom water formation. J. Geophys. Res. 94 (C6), 8289–8302. Brodeau, L., Barnier, B., Penduff, T., Treguier, A.M., Gulev, S., 2010. An ERA40-based atmospheric forcing for global ocean circulation models. Ocean Modell. 31, 88–104. Broecker, W.S., Takahashi, T., Takahashi, T., 1985. Sources and flow patterns of deepocean waters as deduced from potential temperature, salinity, and initial phosphate concentration. J. Geophys. Res. 90 (C4), 6925–6939. Coachman, L.K., Barnes, C.A., 1961. The contribution of Bering Sea water to the Arctic Ocean. Arctic 14, 146–161. Cooper, L.W., McClelland, J.W., Holmes, R.M., Raymond, P.A., Gibson, J.J., Guay, C.K., Peterson, B.J., 2008. Flow-weighted values of runoff tracers (δ18O, DOC, Ba, alkalinity) from the six largest Arctic rivers. Geophys. Res. Lett. 35, L18606, http://dx.doi.org/10.1029/2008GL035007. de Steur, L., Steele, M., Hansen, E., Morison, J., Polyakov, I., Olsen, S.M., Melling, H., McLaughlin, F.A., Kwok, R., Smethie, W.M., Schlosser, P., 2013. Hydrographic changes in the Lincoln Sea in the Arctic Ocean with focus on an upper water freshwater anomaly between 2007 and 2010. J. Geophys. Res. 118, 4699–4715, http://dx.doi.org/10.1002/jgrc.20341. Dickson, R.R., Meincke, J., Malmberg, S.A., Lee, A.J., 1988. The “Great Salinity Anomaly” in the northern North Atlantic 1968–1982. Prog. Oceanogr. 20, 103–151. Dickson, R., Lazier, J., Meincke, J., Rhines, P., Swift, J., 1996. Long-term coordinated changes in the convective activity of the North Atlantic. Prog. Oceanogr. 38 (3), 241–295. The Drakkar Group, 2007. Eddy Permitting Ocean Circulation Hind Casts of Past Decades, CLIVAR Exchanges, 42. Ekwurzel, B., Schlosser, P., Mortlock, R.A., Fairbanks, R.G., Swift, J.H., 2001. River runoff, sea ice meltwater, and Pacific water distribution and mean residence times in the Arctic Ocean. J. Geophys. Res. 106 (C5), 9075–9092, http://dx.doi. org/10.1029/1999JC000024. Guay, C.K., Falkner, K.K., 1997. Barium as a tracer of Arctic halocline and river waters. Deep Sea Res. II 44, 1543–1569. Houssais, M.N., Herbaut, C., 2011. Atmospheric forcing on the Canadian Arctic Archipelago freshwater outflow and implications for the Labrador Sea variability. J. Geophys. Res 116, C00D02, http://dx.doi.org/10.1029/2010JC006323. Jackson, J.M., Allen, S.E., McLaughlin, F.A., Woodgate, R.A., Carmack, E.C., 2011. Changes to the near-surface waters of the Canada Basin, Arctic Ocean from 1993–2009: a basin in transition. J. Geophys. Res. 116, C10008, http://dx.doi. org/10.1029/2011JC007069. Jackson, J.M., Williams, W.J., Carmack, E.C., 2012. Winter sea-ice melt in the Canada Basin, Arctic Ocean. Geophys. Res. Lett. 39, L03603, http://dx.doi.org/10.1029/ 2011GL050219. Jahn, A., Aksenov, Y., de Cuevas, B., de Steur, L., Hakkinen, S., Hansen, E., Herbaut, C., Houssais, M.-N., Karcher, M.J., Kauker, F., Lique, C., Nguyen, A.T., Pemberton, P., Worthen, D.L., Zhang, J., 2012. Arctic Ocean freshwater—how robust are model simulations? J. Geophys. Res. 117, C00D16, http://dx.doi.org/10.1029/ 2012JC007907. Jones, E.P., Anderson, L.G., 1986. On the origin of the chemical properties of the Arctic Ocean halocline. J. Geophys. Res. 91, 10759–10767. Jones, E.P., Eert, A.J., 2006. Waters of Nares Strait in 2001. Polarforschung 74 (1–3), 185–189. Jones, E.P., Anderson, L.G., Swift, J.H., 1998. Distribution of Atlantic and Pacific waters in the upper arctic ocean: implications for circulation. Geophys. Res. Lett. 25, 765–768. Koenigk, T., Mikolajewicz, U., Haak, H., Jungclaus, J., 2007. Arctic freshwater export in the 20th and 21st centuries. J. Geophys. Res. 112, G04S41, http://dx.doi.org/ 10.1029/2006JG000274.

95

Krishfield, R., Tole, J., Proshutinsky, A., Timmermans, M.-L., 2008. Automated icetethered profilers for seawater observations under pack ice in all seasons. J. Atmos. Oceanic Technol. 25 (11), 2091–2105. Lique, C., Treguier, A.M., Scheinert, M., Penduff, T., 2009. A model-based study of ice and freshwater transport variability along both sides of Greenland. Clim. Dyn. 33 (5), 685–705, http://dx.doi.org/10.1007/s00382-008-0510-7. Lique, C., Treguier, A.M., Blanke, B., Grima, N., 2010. On the origins of water masses exported along both sides of Greenland: a Lagrangian model analysis. J. Geophys. Res. 115, C05019, http://dx.doi.org/10.1029/2009JC005316. Macdonald, R.W., Carmack, E.C., Paton, D.W., 1999. Using the δ18O composition in landfast ice as a record of arctic estuarine processes. Mar. Chem. 65, 3–24. Madec, G., 2008. NEMO Ocean Engine, vol. 27. Institut Pierre-Simon Laplace (IPSL). McGeehan, T., Maslowski, W., 2012. Evaluation and control mechanisms of volume and freshwater export through the Canadian Arctic Archipelago in a highresolution pan-Arctic ice-ocean model. J. Geophys. Res. 117, C00D14, http://dx. doi.org/10.1029/2011JC007261. McLaughlin, F.A., Carmack, E.C., Macdonald, R.W., Bishop, J.K.B., 1996. Physical and geochemical properties across the Atlantic/Pacific water mass front in the southern Canadian Basin. J. Geophys. Res. 101 (C1), 1183–1197. McLaughlin, F.A., Carmack, E.C., Macdonald, R.W., Melling, H., Swift, J.H., Wheeler, P. A., Sherr, B.F., Sherr, E.B., 2004. The joint roles of Pacific and Atlantic-origin waters in the Canada Basin, 1997–1998. Deep Sea Res. I 51, 107–128. Melling, H., Lake, R.A., Topham, D.R., Fissel, D.B., 1984. Oceanic thermal structure in the western Canadian Arctic. Cont. Shelf Res. 3 (3), 233–258. Münchow, A., Melling, H., Falkner, K.K., 2006. An observational estimate of volume and freshwater flux leaving the Arctic Ocean through Nares Strait. J. Phys. Oceanogr. 36, 2025–2041. Münchow, A., Falkner, K.K., Melling, H., 2007. Spatial continuity of measured seawater and tracer fluxes through Nares Strait, a dynamically wide channel bordering the Canadian Archipelago. J. Mar. Res. 65, 759–788. Münchow, A., Melling, H., 2008. Ocean current observations from Nares Strait to the west of Greenland: interannual to tidal variability and forcing. J. Mar. Res. 66, 801–833. Münchow, A., Falkner, K.K., Melling, H., Rabe, B., Johnson, H.L., 2011. Ocean warming of Nares Strait bottom waters off northwest Greenland, 2003–2009. Oceanography 24 (3), 114–123. Newton, J.L., Sotirin, B.J., 1997. Boundary undercurrent and water mass changes in the Lincoln Sea. J. Geophys. Res. 102 (C2), 3393–3403. Ostlund, H.G., Hut, G., 1984. Arctic Ocean water mass balance from isotope data. J. Geophys. Res. 89, 6373–6381. Rabe, B., Münchow, A., Johnson, H.L., Melling, H., 2010. Nares Strait hydrography and salinity field from a 3-year moored array. J. Geophys. Res. 115, C07010, http://dx.doi.org/10.1029/2009JC005966. Rabe, B., Johnson, H.L., Münchow, A., Melling, H., 2012. Geostrophic ocean currents and freshwater fluxes across the Canadian polar shelf via Nares Strait. J. Mar. Res. 70 (4), 603–640. Rasmussen, T.A.S., Kliem, N., Kaas, E., 2011. The effect of climate change on the sea ice and hydrography in Nares Strait. Atmosphere-Ocean 49 (3), 245–258. Rennermalm, A.K., Wood, E.F., Weaver, A.J., Eby, M., Dry, S.J., 2007. Relative sensitivity of the Atlantic meridional overturning circulation to river discharge into Hudson Bay and the Arctic Ocean. J. Geophys. Res. 112, G04S48, http://dx. doi.org/10.1029/2006JG000330. Rudels, B., 2012. Volume and freshwater transports through the Canadian Arctic Archipelago-Baffin Bay system. J. Geophys. Res. 116, C00D10, http://dx.doi.org/ 10.1029/2011JC007019. Rudels, B., 2012. Arctic Ocean circulation and variability—advection and external forcing encounter constraints and local processes. Ocean Sci. 8, 261–286. Sadler, H.E., 1976. Water, heat, and salt transports through Nares Strait, Ellesmere Island. J. Fish. Res. Board Can. 33, 2286–2295. Samelson, R.M., Agnew, T., Melling, H., Müchow, A., 2006. Evidence for atmospheric control of sea-ice motion through Nares Strait. Geophys. Res. Lett. 33, L02506, http://dx.doi.org/10.1029/2005GL025016. Seibert, G.H., 1968. Oceanographic observations in the Lincoln Sea, June 1967. Baffin Bay-North Water Proj. Arct. Inst. North Am. Sci. Rep., 46, p. 68. Smethie, W.M., Chayes, D., Perry, R., Schlosser, P., Friedrich, R., 2011. A rosette for sampling ice-covered water. Oceanography 24 (3), 160–161. Steele, M., Boyd, T., 1998. Retreat of the cold halocline layer in the Arctic Ocean. J. Geophys. Res. 103 (C5), 10419–10435. Steele, M., Morley, R., Ermold, W., 2001. PHC: a global ocean hydrography with a high quality Arctic Ocean. J. Clim. 14, 2079–2087. Steele, M., Morison, J., Ermold, W., Rigor, I., Ortmeyer, M., Shimada, K., 2004. Circulation of summer Pacific halocline water in the Arctic Ocean. J. Geophys. Res. 109, C02027, http://dx.doi.org/10.1029/2003JC002009. Taylor, J.R., Falkner, K.K., Schauer, U., Meredith, M., 2003. Quantitative considerations of dissolved barium as a tracer in the Arctic Ocean. J. Geophys. Res. 108 (C12), 3374, http://dx.doi.org/10.1029/2002JC0011635. Timmerman, R., Goosse, H., Madec, G., Fichefet, T., Ethe, C., Duliere, V., 2005. On the representation of high latitude processes in the ORCA-LIM global coupled sea ice-ocean model. Ocean Modell. 8, 175–201. Timmermans, M.-L., Proshutinsky, A., Krishfield, R.A., Perovich, D.K., Richter-Menge, J.A., Stanton, T.P., Toole, J.M., 2011. Surface freshening in the Arctic Ocean's Eurasian Basin: an apparent consequence of recent change in the wind-driven circulation. J. Geophys. Res. 116, C00D03, http://dx.doi.org/10.1029/2011JC006975. Toole, J.M., Krishfield, R.A., Timmermans, M.-L., Proshutinshy, A., 2011. The IceTethered Profiler: Argo of the Arctic. Oceanography 24 (3), 126–135 http://dx. doi.org/10.5670/oceanog.2011.64.

96

J.M. Jackson et al. / Continental Shelf Research 73 (2014) 83–96

Wilson, C., Wallace, D.W.R., 1990. Using the nutrient ratio NO/PO as a tracer of continental shelf waters in the central Arctic Ocean. J. Geophys. Res. 95, 22193–22208. Yamamoto-Kawai, M., Tanaka, N., Pivovarov, S., 2005. Freshwater and brine behaviors in the Arctic Ocean deduced from historical data of δ18O and alkalinity (1929–2002 A.D.). J. Geophys. Res. 110, C10003, http://dx.doi.org/ 10.1029/2004JC002793.

Yamamoto-Kawai, M., McLaughlin, F.A., Carmack, E.C., Nishino, S., Shimada, K., 2008. Freshwater budget of the Canada Basin, Arctic Ocean, from salinity, δ18O, and nutrients. J. Geophys. Res. 113 (C1), C01007, http://dx.doi.org/10.1029/ 2006JC003858. Yang, J., 2006. The seasonal variability of the arctic ocean Ekman transport and its role in the mixed layer heat and salt fluxes. J. Clim. 19, 5366–5387.