Processes in the Kitikmeot Sea estuary constraining marine life

The Kitikmeot Region of Nunavut covers the southern central part of the Northwest Passage in southern Inuit Nunangat (Canadian Arctic Archipelago, CAA) and is home to five Inuit communities (Kugluktuk, Iqaluktuuttiaq, Gjoa Haven, Taloyoak, and Kugaaruk) and two small seasonal communities in Bathurst Inlet (Umingmaktok and Bathurst Inlet). The word kitikmeot roughly translates as “the center of something” (Kitikmeot Heritage Society, 1996); for the purposes of this paper we define the Kitikmeot Sea within the Kitikmeot Region as that segment of the Northwest Passage bounded by the shallow sills of Dolphin and Union Strait in the west and Victoria Strait in the northeast (Figures 1 and 2). The Kitikmeot Sea is a series of basins and inlets connected by narrow straits with shallow sills (Figure 2). There are two large east-west basins, Coronation Gulf and Queen Maud Gulf, connected by Dease Strait, and one large north-south inlet, Bathurst Inlet (Figure 2, with areas given in Table 1).

Inuit have lived in the area long before European explorers came along, and today approximately 6,500 Inuit are tied intrinsically to the region through their history and culture and depend on the land and ocean for travel, hunting, and livelihood. Much later followed a history of 19^th century exploration by European settlers in search of the fabled Northwest Passage from Europe to Asia (Delgado, 1999). These explorers and their patrons gave geographical names for the gulfs, islands, and straits of the region that made it onto settler-made maps instead of Inuit place names. For example, Coronation Gulf was named by Sir John Franklin on his overland journey in 1821 in honor of the coronation of King George IV. Queen Maud Gulf was crossed in 1839 by Peter Dease and Thomas Simpson, whose names remain in the connecting straits, and Queen Maud Gulf was named by Roald Amundsen during his passage through the Northwest Passage in 1905 in honor of the Norwegian Queen Maud of Wales. Today, many of the names associated with colonization are returning to names used by Indigenous residents, which are far more representative of place (Keith et al., 2014): examples include Cambridge Bay, named for Prince Adolphus, Duke of Cambridge, which is now again Iqaluktuuttiaq, meaning “place of good fishing”; Coppermine is now again Kugluktuk, meaning “place of running water”; and Bay Chimo is again named Umingmaktok, meaning “like a musk ox.” Iqaluktuuq, meaning “place of many Arctic char,” is the name of a 3 km stretch of the Ekalluk River that drains Tahiryuaq into Wellington Bay. Tahiryuaq, meaning the big lake, is the Inuit name for Ferguson Lake, on Victoria Island. Clearly, more information of “place” describing the Kitikmeot Region is contained in Indigenous names than in using the names of absent aristocrats, explorers, and wealthy sponsors.

The Kitikmeot Sea has complex bathymetry, which is critical to shaping ocean circulation and mixing but is poorly mapped outside of the main shipping channels. Available Canadian Hydrographic Service charts and ship tracks confirm the region’s very complex bathymetry. Rough, irregular bathymetry is found at the eastern end of Queen Maud Gulf and in Victoria Strait, and sharp, “sawtooth” bathymetry in Bathurst Inlet and Coronation Gulf (Figure 2). This “sawtooth” bathymetry is due to south-to-southeast facing escarpments in inclined rock layers and likely shaped during the last glaciation by scouring from the Laurentide Ice Sheet (Clark et al., 2009). Consistent with Hannah (2009) and Melling (2015), we will show that this bathymetry, especially the shallow sills and narrow passageways, plays a dominant role in shaping circulation in terms of limiting water exchange with adjacent areas and supporting tidal dissipation; thus the major straits are shown in detail in Figure S1a–c. We will also show that the Kitikmeot Sea is oceanographically and ecologically distinct from adjoining regions of the CAA because it receives disproportionately large river inflow from the large mainland watersheds to the south and a very restricted inflow of salty oceanic water due to the shallow sills of its bounding straits.

A small number of studies have focused specifically on the oceanography of the Kitikmeot Sea, mainly in its connection to the larger CAA. For example, at the larger scale, Carmack et al. (2010) and Carmack and McLaughlin (2011) described water properties along an oceanographic transect that encircled northern North America as part of the International Polar Year, Canada’s Three Oceans (C3O) project, and argued that local structures and processes within the CAA cannot be understood in isolation from large scale oceanographic structures and advection. Following this work, Yamamoto-Kawai et al. (2013) showed how external processes and inputs affected aragonite saturation within the CAA, and noted that low aragonite saturation was due to inputs of Pacific-origin waters from the North Pacific and Canada Basin. Carmack et al. (2015) proposed a contiguous, boundary-trapped flow of waters freshened by continental discharge extending clockwise around northern North America, and Yamamoto-Kawai et al. (2010) and Brown et al. (2020b) used geochemical distributions to elucidate the roles of river discharge. Within the Kitikmeot Sea, Arfeuille (2001) examined the joint roles of wind and buoyancy forcing on transports through the southwest CAA from Amundsen Gulf to Larsen Sound, noting very low salinity waters within the Kitikmeot Sea. Rotermund et al. (2021) described tidal flows within the Kitikmeot Sea with an emphasis on the roles of mixing in shallow straits and modifications associated with seasonal ice cover.

These physical and geochemical descriptions of the system provide the context for understanding the generally low, but heterogeneous marine productivity of the region. Varela et al. (2013) and Wyatt et al. (2013) measured biogenic particle distributions, size-fractionated chlorophyll-a, and primary productivity along the C3O transect and noted the relative importance of siliceous phytoplankton in the CAA. They also observed intermediate values of chlorophyll-a and total primary production between the adjoining Canada Basin to the west (lower) and Baffin Bay and the Labrador Sea to the east (higher); within the Kitikmeot Sea a subsurface chlorophyll maximum lay near 40 m. Ice algal biomass also appears to be low (Kim et al., 2020), but can be enhanced in areas of high tidal currents where ice is thin (e.g., Dease Strait; Dalman et al., 2019). Ice algae provide a distinct signal in the vertical flux of particulate organic matter during ice melt (e.g., in Queen Maud Gulf; Dezutter et al., 2021). In particular, Campbell et al. (2016) found co-limitation of sea-ice algal production by irradiance and nitrogen during spring in Dease Strait, in an area of enhanced tidal mixing. Zooplankton composition within the Kitikmeot Sea is dominated by Arctic taxa, with clear community gradients related to freshwater influence indicated by presence of brackish species (Nelson et al., 2019), and characterized by lower biomass compared to other regions of the North American Arctic (Darnis et al., 2024). Bluhm et al. (2022) provided an initial inventory of kelp within the Kitikmeot Sea, noting a heterogeneous distribution and joint importance of water flow and seabed composition. Benthic faunal composition has been characterized as a transition between the adjacent Beaufort Sea and Baffin Bay, with regionally high biodiversity and measurable influence of terrestrial and Pacific Water inputs (Dumais et al., 2022). Published benthic biomass data are sparse, but values may be low (Roy et al., 2014) to high; Balsom (2003) collected macrobenthic samples along a transect between the Gulf of Alaska and Larsen Sound and noted that while abundance was generally low within the Kitikmeot Sea, a single benthic “hotspot” was present in Queen Maud Gulf south of Hat Island. Fish abundance is also low in comparison with adjacent regions, and Arctic cod (Boreogadus saida), the most abundant Arctic fish in many Arctic seas, is scarce (Bouchard et al., 2018). Instead, large populations of the anadromous Arctic char (Salvelinus alpinus) migrate annually from their feeding grounds in the ocean to freshwater lakes and rivers to escape sub-zero temperatures (Dubos et al., 2023). The region supports a commercial char fishery in Cambridge Bay (Day and Harris, 2013; Harris et al., 2020a).

Marine mammal sightings occur, but the area is generally characterized as a “cold spot” in that regard (Roff et al., 2020), presumably as a result of low primary productivity and low abundance of prey. A report based on a survey of Indigenous knowledge documents different species of seals and fish from those found outside the Kitikmeot Sea (NCRI, 2010, 2011, 2014, 2015). This observation is consistent with Wong et al. (2014) who likewise noted the low abundance and diversity of marine birds within the Kitikmeot Sea in comparison to other regions along an oceanographic transit around northern North America. Food supply for birds on land, in contrast, appears to be more plentiful, inferred from the internationally important Queen Maud Gulf Migratory Bird Sanctuary located on the mainland to the south of Queen Maud Gulf (compare to Didiuk and Ferguson, 2005).

The exchanges of water into, within, and out of the Kitikmeot Sea through the straits remain largely unquantified, and thus we know relatively little about the role that the Kitikmeot Sea plays in modifying the flows of water through the CAA. Here, our goal is to investigate the resulting estuarine circulation within the Kitikmeot Sea and how it may constrain internal biological productivity. To address this goal, we (1) describe the oceanographic setting of the Kitikmeot Sea, (2) describe the large, regional river inflows and pathways, (3) discuss physical and chemical property distributions in terms of the estuarine circulation, and (4) speculate on the region as a setting for marine life. We then conclude with a conceptual model of summer and winter estuarine circulation and how it influences the ecosystem.

To provide a holistic perspective of the environmental conditions in the study area we compiled data from a suite of different sources, cruises, and research programs. Inherently, data coverage therefore does not always coincide in time and space.

• Bathymetry: The bathymetry used in the maps, hydrographic sections, and calculation of basin volume is a combination of gridded data made available by the General Bathymetric Chart of the Oceans (GEBCO Compilation Group, 2023), soundings obtained from the Canadian Hydrographic Service, and soundings collected from the CCGS Sir Wilfrid Laurier during its 2007 oceanographic expedition to Coronation Gulf and Bathurst Inlet as part of the International Polar Year’s C3O expeditions. All maps were made using the M_Map mapping package (Pawlowicz, 2020). Certain areas are still poorly charted or uncharted, leaving errors in the map and in models depending on bathymetry.

• Sea ice: Ice type and concentration data for the Kitikmeot Sea region are made available by the Canadian Ice Service, from which we used the 1991–2020 median sea-ice concentration. Ice thickness data are from the Canadian Ice Service (using 2002–2023), Ocean Networks Canada (2015–2016), and the Canadian Rangers Ocean Watch (2011–2020).

• Tides: Estimates of tidal flows were obtained from 2 models: a high-resolution package of Webtide developed for the Kitikmeot Sea (Collins et al., 2011) and from the tidal model analyzed by Rotermund et al. (2021). Both of these barotropic, finite-element models have sufficient horizontal resolution to resolve the larger straits and passageways in the Kitikmeot Sea. Being barotropic they do not contain stratification of salinity, temperature, or density.

• River inflow: Estimates of river water inflow were based on the watersheds that drain into Coronation Gulf, Dease Strait, Bathurst Inlet, and Queen Maud Gulf (Figure 1), with drainage basins selected based on the direct draining of rivers defined in the HydroRIVERS v1.0 and associated drainage basin polygons selected from the global HydroSHEDS database at 15 arc-second resolution (Lehner and Grill, 2013). Average annual discharges for the Kitikmeot drainage regions, as well as the Simpson River and the Rae River (see Figure 2 for river mouth locations), were determined from the relationship between CAA river discharge and drainage basin area defined in Brown et al. (2020b) based on gauged river data from Environment and Climate Change Canada (ECCC). Average annual discharge for the Coppermine River, Burnside River, Ellice River, Hood River, and Freshwater Creek come from available Environment and Climate Change Canada data sets up to 2014.

• Precipitation: Precipitation data were obtained from Environment and Climate Change Canada. The seasonal cycle of precipitation, rain, and snow is the average of the Cambridge Bay and Kugluktuk monthly climate data for 1981–2010.

• Wind: We used the 6-hourly ERA5 reanalysis data from the European Centre for Medium Range Weather Forecasts (Hersbach et al., 2023) for the years 2001–2020. Wind-roses were generated for the open water season (mid-July to mid-October) and ice-covered seasons across the Kitikmeot Sea using M_Map (Pawlowicz, 2020). We followed the convention for ocean currents and show the direction that the winds are blowing to, rather than from.

• Temperature and salinity profiles: We use water column data collected during summer in 1995 aboard the Arctic Ivik, and from the summer/fall in 1996, 1999, 2000, 2007, and 2008 aboard the CCGS Sir Wilfrid Laurier. All CTD sensors were calibrated pre- and post-cruise at Sea-Bird Scientific and salinity was validated by comparison to water samples collected by Niskin bottle and analyzed as described in McLaughlin et al. (2012). In addition, a limited number of CTD profiles in winter were collected by C3O and the Canadian Rangers Ocean Watch in years 2011–2014. All CTD data used here are available from Fisheries and Oceans Canada’s Marine Environmental Data Section Archive. Practical salinity from CTD sensors was converted to absolute salinity using McDougall and Barker (2011).

  Relatively few CTD profiles have been collected in the Kitikmeot Sea. The datasets chosen here are from a number of different programs and, taken together, are unique in their broad coverage of the Kitikmeot Sea from Larsen Sound in the east, to Amundsen Gulf in the west and Bathurst Inlet in the south. The locations of the CTD profiles used are shown along with the data in the Results section. Variations due to different years, months, and environmental conditions during sampling are inherent to this (as any) data compilation.

• Dissolved nutrients: Water samples for the determination of nitrate + nitrite (NO₃⁻ + NO₂⁻, herein referred to as N or nitrate), orthophosphate (H₃PO₄, herein referred to as P or phosphate), and silicic acid (Si(OH)₄, herein referred to as Si or silicate) were collected from Niskin bottles and analyzed following the methods described in McLaughlin et al. (2012). Nutrient samples were collected during opportunistic transits of the region in 1995 by the Arctic Ivik and in 1996, 1998–2010 by the CCGS Sir Wilfrid Laurier and CCGS Louis S. St-Laurent. The dissolved nutrient data are also available from Fisheries and Oceans Canada’s Marine Environmental Data Section Archive.

The Kitikmeot Sea is forced by freshwater input, tides, and winds. It is also forced by inflow of saline, nutrient-enriched, waters from the Arctic Ocean, via Amundsen Gulf in the west and from Larsen Sound in the northeast. We will examine each of these factors in turn and discuss how they vary by season.

Walker (1977) and Vuglinsky (1998) estimated the total river discharge to the CAA to be 219 and 211 km³ yr⁻¹, respectively, while Alkire et al. (2017) and Brown et al. (2020b) estimated inflows to be slightly higher at 257 and 250 km³ yr⁻¹, respectively. On average, the Kitikmeot Sea receives approximately 50 km³ yr⁻¹, or approximately 0.8 m of river water added to its surface each year. This amount is disproportionately large compared to further north in the CAA due to the large mainland watersheds that are almost 5 times the area of the sea itself (see Figure 1 and estimates in Tables 1 and 2; Brown et al., 2020b, supplementary material). In total, the Kitikmeot Sea is only approximately 2% of the entire marine area of the CAA, but receives more than 20% of the total river inflow.

There is significant variation of river inflow across the Kitikmeot Sea, dependent on locations of the seven of the largest rivers (Figure 2) which form 45% of the total inflow. Coronation Gulf, Bathurst Inlet, and Dease Strait together receive about 33 km³ yr⁻¹, or 1.12 m, of which about 22 km³ yr⁻¹ (or 66%) flows into Coronation Gulf and about 11.5 km³ yr⁻¹ (or 34%) flows into Bathurst Inlet. The three largest rivers flowing into Coronation Gulf are the Coppermine River (approximately 8.1 km³ yr⁻¹) and Rae River (approximately 2.5 km³ yr⁻¹) in western Coronation Gulf and the Tree River (approximately 1.1 km³ yr⁻¹) in southern Coronation Gulf. Bathurst Inlet has the Hood River (approximately 2.6 km³ yr⁻¹) in northern Bathurst Inlet, and the Burnside River (approximately 4.2 km³ yr⁻¹) in southern Bathurst Inlet. Queen Maud Gulf receives about 15 km³ yr⁻¹ (or 0.5 m) mostly from many smaller rivers that drain through the migratory bird sanctuary to the shore. The largest of these are the Ellice River (approximately 2.7 km³ yr⁻¹) and the Simpson River (approximately 1.2 km³ yr⁻¹).

Queen Maud Gulf also receives significant freshwater from the north from the terminus of a “river” of multi-year sea ice that slowly moves southward through the CAA and ultimately melts in the gulf (Melling, 2002). This southward flow is augmented by thick seasonal ice in Victoria Strait, formed by the strong tidal flows in the strait that push the ice into rough deformed ridges (e.g., Hass and Howell, 2015; Rotermund et al., 2021). Rotermund et al. (2021) reported ice ridges 6–8 m thick in Victoria Strait with an average ice thickness of 3.32 m. The influx of sea ice to Queen Maud Gulf in summer has not been measured, but a recent modeling study found this southward flow of ice to be approximately 24 km³ yr⁻¹ with inflow split between the early summer melt periods and the fall (Xu et al., 2021). Here we conservatively estimate 14 km³ yr⁻¹, or 0.5 m, of freshwater from sea ice. Using this ice-derived freshwater value doubles the annual freshwater loading of the gulf, and the total loading is then similar to Coronation Gulf (Table 2).

Precipitation over the ocean as rain and snow also contributes to the annual freshwater loading of the Kitikmeot Sea. Using Kugluktuk and Cambridge Bay weather station data, we estimated that about 0.2 m of precipitation falls each year, which adds a further 20% to the annual freshwater input.

The seasonal cycle of freshwater loading in the Kitikmeot Sea has a strong peak in June, dominated by snow melt on the large southern mainland watersheds, and then returns to wintertime values by the end of November (Figure 3). A small amount of flow, approximately 10% of summertime flow, then continues in winter as the large mainland rivers continue to be fed by numerous small and medium lakes. Of the 0.2 m of precipitation per year that falls on the ocean noted above, approximately half falls as rain from June to September and the other half falls as snow from October to May. The snow on the sea ice then melts concurrent with the ice in June/July, coinciding with peak river inflow.

Sea-ice growth removes approximately 1.4 m of freshwater from November to May that is returned to the water column in June when the ice melts (Figure 3, Table 2, Text S1), although numerical modeling by Xu et al. (2021) shows some export of sea ice through Dolphin and Union Strait during break-up. This addition of freshwater during melt doubles the annual freshwater loading of the Kitikmeot Sea from precipitation and river inflow combined, and thus has a large impact on summertime stratification and surface salinity. Conversely, the removal of freshwater by the growth of ice over November to May causes the mixed layer depth to deepen to approximately 20–50 m, the winter depth limit of brine convection (see Section 3.2). The seasonal cycle of sea ice is similar in the adjacent Amundsen Gulf and Larsen Sound, though ice may persist in Larsen Sound all summer.

The sea-ice phenology of the CAA has been described recently by Dauginis and Brown (2020). The Kitikmeot Sea is the first region within the Northwest Passage to clear of ice in July to early August and has the longest ice-free period (Figures 4 and 5). While dates of break-up will likely change to earlier melt and later freeze-up as the climate continues to warm, the pattern of south-to-north melt and freeze-up described below will likely remain (e.g., Derksen et al., 2019).

Owing to the warming effects of north-flowing rivers and the southerly location, areas of open water appear first in early July in southern Bathurst Inlet, at the mouths of the Burnside and Western Rivers, and at the mouth of the Coppermine River in western Coronation Gulf. From this starting point, the open water generally expands northward and eastward across Coronation Gulf, and then progresses eastward along Dease Strait to Cambridge Bay near the end of July. Queen Maud Gulf is then ice-free by early-August. Ice often lingers all season around Victoria Strait. This ice is typically the thick, multiyear ice that has flowed south from the northern CAA and jammed in the narrows of Victoria Strait (Melling, 2002; Haas and Howell, 2015) and thick ridged ice formed by strong tidal currents in the strait. Freeze-up progresses in reverse order: beginning in mid-October in Queen Maud Gulf (Dezutter et al., 2021) with ice formation progressing westward across Dease Strait into eastern Coronation Gulf and Bathurst Inlet. Freeze-up is complete by the first week of November, and ice growth then continues over winter and spring. In Cambridge Bay, ice grows to approximately 1.9 m thick by late May/early June and melts to open water over the next month (Cambridge Bay Canadian Ice Service and Ocean Networks Canada ice profiling sonar data; Figure 5). The “open water” season in the Kitikmeot Sea is thus approximately the 3 months from mid-July to mid-October, while the “ice-covered” season is the 9 months from mid-October to mid-July.

Tidal currents and their associated mixing process are especially important within the Kitikmeot Sea for their roles in creating winter polynyas and thin ice where tidal currents are amplified over the narrow and shallow sills of both the interior and bounding straits (Stirling, 1997; Williams et al., 2007; Hannah et al., 2009; Melling et al., 2015). Tides enter the Kitikmeot Sea from the adjacent Amundsen Gulf and Larsen Sound through the bounding straits. While tidal models for the full CAA show the Kitikmeot Sea to be a region of minimum tidal elevation (Hannah et al., 2009; Collins, et al., 2011), a high resolution tidal model of the Kitikmeot Sea shows that the dominant M2 and K1 tidal currents of the region are amplified in Dolphin and Union and Victoria Straits which, in turn, produce high tidal dissipation as the tide enters the Kitikmeot Sea (Rotermund et al., 2021) with strong spring-neap variation (shown in Figure 6). This dissipation is especially strong in the Victoria Strait region where 80% of the incoming 0.74 GW of M2 tidal energy flux and 40% of the 0.033 GW K1 tidal energy flux are dissipated in summer (Rotermund et al., 2021). In addition, over 50% of the M2 tide that enters Queen Maud Gulf turns eastward around an amphidromic point to dissipate in the relatively shallow waters of the eastern gulf (see Figure 6 and bathymetry in Figures 2 and S2). As bathymetric data of the region improves, we expect the locations of this dissipation within the eastern gulf to be refined. Rotermund et al. (2021) also investigated the strong damping effect of thick ice in Victoria Strait, showing that the partial blockage of Victoria Strait in winter by rough and thick sea ice is responsible for the 50% reduction in wintertime tidal velocities found in the Kitikmeot Sea. At still smaller scales, enhanced tidal dissipation is also observed and modeled over shallow and narrow sills within the Kitikmeot Sea (Williams et al., 2018).

During the 3-month-long summer-fall open water period, wind forcing will strongly influence circulation patterns both interior to the Kitikmeot Sea and through wind-forced exchanges with the adjacent Amundsen Gulf and Larsen Sound (Xu et al., 2021). A thorough analysis of available wind data and response of the Kitikmeot Sea is beyond the scope of this paper; however, wind-roses for the open water season (mid-July to mid-October) are shown in Figure 7a across the Kitikmeot Sea for the years 2001–2020. This 20-year average does not show the large interannual variation (Xu et al., 2021) but does reveal significant differences across the region. Wind in Coronation Gulf and Dolphin and Union Strait is bi-directional, roughly aligned between W/NW and E/SE. These winds are aligned with Dolphin and Union Strait and so are expected to drive surface waters through the strait between Coronation Gulf and Amundsen Gulf. We also expect the Coronation Gulf winds to transport the fresher surface waters of the gulf efficiently toward the north or south coasts, generating cyclonic coastal boundary currents that transport surface waters toward the east on the southern coast and toward the west on the northern coast (Carmack et al., 2015). Drifter tracks during wind events in Dease Strait/eastern Queen Maud Gulf illustrate this effect (Figure 7b). In contrast to Coronation Gulf, Queen Maud Gulf, and Victoria Strait have a prevailing wind from the north, blowing toward SSW. These winds are expected to contribute to the southward flow of sea ice into the gulf from the north during summer-fall (see Section 3.1.1) and will tend to move the surface waters to the eastern gulf. During the 9-month-long ice-covered season, winds have the same overall pattern (see Figure S3) but the ice is land-fast so there is no stress on the ocean due to the wind.

Owing to large river inflows, the upper layer waters (0–40 m) within the Kitikmeot Sea are fresher than those in the adjoining basins, Amundsen Gulf and Larsen Sound (Figure 8). In the surface waters, nutrient concentrations are low in summer and fall (Figure 9), with nitrate becoming the limiting macronutrient (approximately 0 mmol m⁻³) in the surface mixed layer. In Bathurst Inlet, phosphate concentrations also approach zero at the surface (Figure 9), reinforcing the generally oligotrophic nature of the Kitikmeot Sea. The few existing CTD profiles in winter (Figures 10d and 11d) show the magnitude of surface water seasonality. Wintertime mixed layers deepen to 20–50 m, from the brine-rejection of sea-ice formation, which would lead to a limited seasonal replenishment of surface nutrients prior to the spring bloom. For example, nitrate concentrations are approximately half that of Amundsen Gulf or Larsen Sound so mixed layer deepening entrains less nitrate to the surface mixed layer (compare to Figure 9a).

As shown by a section of salinity extending from Amundsen Gulf to Larsen Sound (Figure 12), the inflow of the subsurface Pacific origin water mass from the west is constrained by the shallow sill of Dolphin and Union Strait (10–15 m) and from the northeast by the sill in Victoria Strait (20–30 m; see Section 4 for details of the straits). Only the lower-salinity, uppermost Canada Basin waters enter the Kitikmeot Sea over these shallow bounding sills, so that the bottom waters within the Kitikmeot Sea (which begin at about 50 m depth and extend to the sea-floor) have a salinity of approximately 29 g kg⁻¹, or 4–5 g kg⁻¹ fresher than those in the adjacent basins (Figures 12, 13, and 14). In eastern Coronation Gulf, Bathurst Inlet lies behind an 80–100 m sill at its northern end and a 30 m sill just north of the Burnside River (Figure 2) and has bottom waters with even lower salinity. Bottom salinities are approximately 29–30 g kg⁻¹ for Queen Maud Gulf, 28.8–29 g kg⁻¹ for Coronation Gulf South, 28.5 g kg⁻¹ in northern Bathurst Inlet, and 27.8 g kg⁻¹ in southern Bathurst Inlet (Table 1).

The deep and bottom waters within the Kitikmeot Sea correspondingly have lower dissolved nitrate and phosphate concentrations (nitrate of approximately 8 mmol m⁻³, phosphate of approximately 1.5 mmol m⁻³; Figure 9) that are approximately half the nutrient maximum of the Pacific-origin water in the adjacent basins (Figure 9). Deep water silicate concentrations, on the other hand, increase in the deep waters of Coronation Gulf and Bathurst Inlet compared with Pacific-origin source waters in Amundsen Gulf and Larsen Sound. This increase, visible in Figure 9e, indicates de-coupling between nitrate and silicate cycles in Kitikmeot deep waters.

Profiles of salinity (Figure 10), temperature (Figure 11), and temperature/salinity correlations (Figure 15) are grouped to show differences in water properties from basin to basin across each of the three major sills, Victoria Strait, Dease Strait, and Dolphin and Union Strait. Salinities outside of the Kitikmeot Sea, in the adjoining Amundsen Gulf and Larsen Sound, are 2 to 3 g kg⁻¹ greater that those of the Kitikmeot Sea and show winter-water temperature minima near 32.5 g kg⁻¹ (Larsen Sound) and 31.5 g kg⁻¹ (Amundsen Gulf). Noting from the isolines of density shown on Figure 15 that density in deep waters is mainly controlled by salinity, we observe that: (1) deep waters in Larsen Sound are denser than those in Queen Maud Gulf and are freshened while flowing over the 20–30 m sills of Victoria Strait; (2) deep waters in Queen Maud Gulf are slightly denser than waters in Coronation Gulf and are freshened while flowing though Dease Strait, with a 45–55 m sill; (3) deep waters in Amundsen Gulf are denser than those in Coronation Gulf and are freshened while passing over the 10–20 m sill of Dolphin and Union Strait; (4) deep waters in Coronation Gulf North Basin reveal a deep temperature minimum with warming toward the bottom; and (5) the freshest deep waters are found in Bathurst Inlet.

Near-bottom waters in Coronation Gulf North Basin are as much as 1–2°C above the freezing point owing to tidal mixing within Dolphin-Union Strait and entrainment of near-surface waters, thus providing a significant reservoir of deep, sensible heat year-round (Figures 11 and 15). Importantly, these waters have lower oxygen concentrations (not shown), higher nutrient concentrations and slightly lower salinities (Figures 10 and 15) than those flowing westward through Dease Strait. Thus, being slightly less dense, they interflow as intermediate waters (80–240 m) above the slightly denser deep waters entering through Dease Strait and into the Coronation Gulf South Basin (Figure 14).

Salinity and temperature profiles in the sill regions of Victoria Strait, Dease Strait, and Dolphin and Union Strait show evidence of strong mixing (Figures 10 and 11), with (1) much lower top-to-bottom stratification than over the same depth in adjoining basins and (2) stepped or irregular profiles illustrating that both surface and bottom waters are modified as they transit the sills. With the estuarine-like circulation (see Section 4), surface waters become saltier as they flow outward, and sill depth waters become less salty as they flow inward. Waters in Dolphin and Union Strait are more mixed than Victoria Strait, which is consistent with its shallower sill.

The CAA is a major oceanic gateway enabling transport of freshwater and low-salinity Pacific-origin water from the Arctic Ocean to Baffin Bay and the North Atlantic. This “throughflow” has been found, via observations and models, to occur predominantly through the northern straits of the CAA (for review and locations of the throughflow, see McLaughlin et al., 2004; Beszczynska-Moeller et al., 2011; McGeehan and Maslowski, 2012). The Kitikmeot Sea in the southern Northwest Passage is not a significant part of the throughflow and is viewed here as an additional southern source of freshwater to the CAA and adjacent seas, with a significant influence on the oceanography of the region, including its near-surface stratification.

River water that flows into the Kitikmeot Sea combines with direct precipitation and sea-ice melt to form a low salinity, estuarine circulation that exits the region into Amundsen Gulf to the west and Larsen Sound to the northeast. From Amundsen Gulf, water can either head north through the CAA or continue northwest and become entrained into the anticyclonic Beaufort Gyre circulation. From Larsen Sound in the center of the southern CAA, it is expected to join the CAA throughflow heading north and east to Baffin Bay (McLaughlin et al., 2004). The partitioning of this low salinity outflow between Dolphin and Union Strait and Victoria Strait requires further study.

The large volume of freshwater added to the Kitikmeot Sea in June, July, and August will raise sea level relative to the adjoining basins and force a barotropic outflow through the bounding straits. We estimated that (1) this barotropic outflow easily “keeps-up” with the freshwater inflow, so that little sea level rise occurs (see Text S2) and (2) only approximately 10% of the annual freshwater input exits this way (see Text S3). The remaining 90% of the freshwater exits the Kitikmeot Sea through the bounding straits via other processes. While episodic wind-driven flow through the bounding straits may be important, we focus here on the slower estuarine-like exchange through the bounding straits.

In the estuarine-like circulation, the input of freshwater from rivers, advected ice and precipitation, and the inflow of salt water through the bounding straits mix together within the Kitikmeot Sea to produce a larger volume of low-salinity surface water that exits through the bounding straits (compare to Geyer and Cannon, 1982; MacCready and Geyer, 2010; MacCready and Banas, 2011; Geyer and MacCready, 2014; Thomson et al., 2020). Internal to the Kitikmeot Sea, this circulation and mixing is modified by many processes including wind forcing, surface heating and cooling, sea-ice formation, locally enhanced tides, and buoyancy driven flows. As freshwater progresses outward across the Kitikmeot Sea, these processes entrain salty water from below resulting in a gradual increase in surface salinity from the river mouths to the bounding straits. This gradual increase is evident for our 2007 sections along Bathurst Inlet (Figure 13) and from southern Coronation Gulf through Dolphin and Union Strait (Figure 14).

Overall annual conservation of salt and volume for the Kitikmeot Sea allows us to estimate the exchange flow though the bounding straits, based on the volume of freshwater input to the Kitikmeot Sea, the salinity of low-salinity water flowing out of the Kitikmeot Sea through the bounding straits, S_out, and the salinity of high-salinity water entering the Kitikmeot Sea from the bounding straits, S_in (see Text S4). We took the Kitikmeot Sea bottom water salinity of approximately 29 g kg⁻¹ as S_in and roughly estimate S_out to be approximately 25 g kg⁻¹. Using these values, the annual estuarine exchange through Dolphin and Union and Victoria Straits is estimated to be about 300 km³ yr⁻¹, or approximately 6 times the freshwater inflow. This inflow rate leads to a residence time of approximately 13 years for deep waters in the Kitikmeot Sea (volume below 50 m is about 4,000 km³).

Across the bounding straits, salinity at sill depth outside of the Kitikmeot Sea is larger than the bottom salinities found inside the Kitikmeot Sea (Figure 10). This suggests that mixing within the bounding straits lowers the salinity of inflowing high salinity water to close to the salinity of the bottom waters of the Kitikmeot Sea.

The sill regions of Victoria Strait and Dolphin and Union Strait are about 40 km wide (coast to coast) and 15 km long with maximum sill depths of 20–30 m and 10–15 m, respectively. Using an effective sill depth of 10 m, we estimated that estuarine exchange has a flow speed of 2–3 cm s⁻¹ (about 2 km day⁻¹) within the straits, yielding a transit time of about 7 days. While this slow exchange is required for salt balance in the Kitikmeot Sea, tidal flows over the sills of the bounding straits are much larger, with maximum tidal speeds of 20–30 cm s⁻¹ during neap tides and 60–70 cm s⁻¹ during spring tides (e.g., Webtide; Collins et al., 2011; Rotermund et al., 2021). These tidal speeds are 10–30 times larger than our estimate of the annual average estuarine exchange speed and yield tidal excursion length scales of 4–13 km (Webtide; Collins et al., 2011), similar to, or less than, the length of the sills. We thus consider tidal mixing and friction over the sills as the dominant factors that modify and control the estuarine exchange flow and contribute to the very low overall salinity of the Kitikmeot Sea (Figure 12). In support of this approach, we note the very rough or lumpy bathymetry of the sill regions and weakened vertical gradients in the few temperature and salinity profiles taken near the sills (Figures 10 and 11).

As mixing over the sills of the bounding straits will raise the salinity of the outflowing low-salinity water and lower the salinity of the inflowing high-salinity water, a lower bound for bottom salinity in the Kitikmeot Sea can then be estimated by considering complete mixing over the sills. For complete mixing, we simply estimated (S_in)_low = (S_out + S_AG,LS)/2 where S_out is the salinity of the outflowing surface waters entering the strait and S_AG,LS is the salinity of water entering the bounding sill regions at sill depth from outside the Kitikmeot Sea in either Amundsen Gulf or Larsen Sound. Using S_o = 25 g kg⁻¹ and S_AG,LS = 32 g kg⁻¹ gives (S_in)_low = 28.5 g kg⁻¹, which is lower than, but similar to, the bottom waters of the Kitikmeot Sea, providing some evidence for strong mixing over the bounding sills.

As discussed in Section 3.1.2, modeled barotropic tidal dissipation is highest above the bounding sills (see Rotermund et al., 2021). We wished to assess whether this dissipation over the sills is strong enough to vertically mix the whole water column against the stratifying estuarine inflow to the sill regions. An often used and successful parameter for evaluating the likelihood of tidal mixing is $H/UT3$⁠, where H is the water depth and U_T is the tidal velocity amplitude (Simpson and Hunter, 1974), and empirical results suggest a critical value of $H/UT3$ ∼ 100 m⁻² s³ for full water column mixing. This parameter has been used previously in the CAA to note the potential for tidal mixing over the major sills (McLaughlin et al., 2004; Hannah et al., 2009; Melling, 2015). Figure 6 shows estimates of $H/UT3$ for spring and neap tides from the tidal model used in Rotermund et al. (2021) for open water conditions. Estimates of $H/UT3$ for Dolphin and Union Strait and Victoria Strait show a large range over the spring neap tidal cycle, owing to the dependence on $UT3$ and predictions that spring tides are 3–4 times faster than neap tides. During spring tides $H/UT3$ is O(1) for Victoria Strait and Dolphin and Union Strait, suggesting strong tidal mixing. During neap tides $H/UT3$ is O(10³) suggesting weaker tidal mixing and that stratification may persist, especially in Victoria Strait.

Following Simpson and Hunter (1974), $H/UT3$ is part of the ratio, R, between the rate of energy input required to mix the water column against a stratifying influence to the rate of work done against bottom friction by the tide. We applied their approach for the sill regions of the bounding straits and estimated the full ratio that includes the stratifying estuarine flow and bottom friction (see Text S5). We then used R to compare tidal mixing in the bounding straits during the ice-free and ice-covered seasons. During ice cover we found that: (1) landfast sea ice adds surface drag, nominally doubling the area over which drag acts, and reduces depth over the sills by the thickness of the ice; and (2) the salinity difference across the strait is at least halved due to 20–50 m deep wintertime mixed layers driven by sea-ice formation. These factors reduced R by at least a factor of 4, but to estimate R during ice cover we also needed estimates of the reduction in tidal speeds due to the additional drag under sea ice. For these we could either use the model results of Rotermund et al. (2021) or assume that the tidal dynamics in the strait are dominated by friction (see Text S6). Both approaches produced the same result: tidal speeds are approximately halved in winter so that there is surprisingly little change in R from summer to winter, and the spring neap cycle continues to dominate.

Our use of $H/UT3$ and R is a simplistic approach inviting further study. For example, it does not consider the efficiency of the mixing in raising potential energy, times scales of tidal boundary layer growth, or the complexity of the flows interacting with the rough and lumpy bathymetry, or heavily ridged sea ice and complex coastlines. These additional significant factors are beyond the scope of this study.

The progress of inflowing salty deep water across the sills and basins of the Kitikmeot Sea is consistent with salinity and temperature profiles and temperature-salinity correlations (Figures 10, 11, and 15). We begin outside of the Kitikmeot Sea in Larsen Sound and progress inward from there across sills and basins to the southern end of Bathurst Inlet. Profiles in the Victoria Strait region are consistent with inflowing salty water from Larsen Sound becoming less salty as it transits the 20–30 m deep sill, producing water within the range of bottom salinities of Queen Maud Gulf (Figures 10a and 15). Profiles from Queen Maud Gulf through Dease Strait to the southern basin of Coronation Gulf (Figures 10b and 15) have a similar pattern, consistent with salty water in Queen Maud Gulf above the sill depth of 55 m becoming slightly fresher as it transits Dease Strait and producing water within the range of bottom salinities of Coronation Gulf (Figures 10c and 15). This pattern repeats along Bathurst Inlet with an 80–100 m sill at its northern end and a 30–45 m sill just north of the Burnside River. This route for deep water formation, starting with Victoria Strait is the deepest route for waters to flow into southern Coronation Gulf. Dolphin and Union Strait is shallower, with a 15 m deep sill, but salinity and temperature profiles from eastern Amundsen Gulf though the strait to the northern basin of Coronation Gulf share the same pattern as Victoria Strait. Inflow of salty water over Dolphin and Union Strait likely produces the deep water in the northern basin of Coronation Gulf, and there is evidence of this process in Figure 11: the bottom waters of the northern basin of Coronation Gulf warm slightly toward the bottom, and these warmer waters have spread laterally into the southern basin forming a weak temperature maximum there at about 150 m deep. This pattern is consistent with the saltiest water inflowing over Dolphin and Union Strait in summer.

The Kitikmeot Sea is an oceanographically distinct region that has limited connection to the greater Arctic Ocean and CAA via the shallow Dolphin and Union Strait and Victoria Strait (Figure 16a). Large river inflow from mainland watersheds delivers an equivalent of 80 cm of freshwater annually to the 60,000 km² Kitikmeot Sea. Horizontal circulations (Figure 16b) within Coronation Gulf and Queen Maud Gulf are proposed to be cyclonic, due to both the wind response (Xu et al., 2021) and the rotational constraint on incoming rivers (Carmack et al., 2015; see Section 4.5.2). The horizontal scale of these features also depends on the baroclinic Rossby radius which we estimate to be only 2–5 km during the open-water season and <2 km during ice cover (compare to approximately 6 km outside of the Kitikmeot Sea; Nurser and Bacon, 2014). The small Rossby radii in the Kitikmeot Sea are due to its lower bottom salinity, resulting in lower overall stratification (during open water) that is then eroded mostly by brine rejection during wintertime ice growth.

The shallow sills of the straits limit the salinity of inflowing water from Amundsen Gulf and Larsen Sound, and lead to a maximum salinity of the bottom waters of the Kitikmeot Sea of only approximately 29 g kg⁻¹ in Queen Maud Gulf and Coronation Gulf. For the annual river inflow to leave the Kitikmeot Sea, estuarine-like circulation is required in which low-salinity water leaves at the surface of the straits and water with salinity of approximately 29 g kg⁻¹ enters. Bottom water properties suggest that inflow through Dolphin and Union Strait forms the bottom waters of the adjacent northern basin of Coronation Gulf, while inflow though Victoria Strait forms the bottom waters of Queen Maud Gulf and the deeper southern basin of Coronation Gulf. Studies of the flows through the two gateway straits are required to fully quantify the estuarine inflow and outflow and its forcing.

Our estimate of the estuarine circulation assumes the inflow has nitrate concentrations of 6–8 mmol m⁻³, which are the bottom nitrate concentrations in the Kitikmeot Sea. Combining this nitrate concentration with our estimated inflow rate and applying a standard Redfield ratio N:C of 106:16 leads to an estimate of annual new pelagic production of only about 2.5 g m⁻² yr⁻¹. This estimate is similar to previous field-derived estimates (Varela et al., 2013) and also similar to our estimate based on wintertime replenishment of nitrate in the surface waters prior to the spring bloom. For comparison, annual new primary production at Arctic inflow shelves, the Chukchi and Barents Sea, can be over an order of magnitude higher (Codispoti et al., 2013). Observations of seaweeds (Bluhm et al., 2022) and green-brown mats on under-water imagery in coastal waters (unpublished) warrant further study on whether higher nutrient concentrations in sub-surface layers facilitate comparatively higher primary production at the seafloor, as seen in other highly stratified areas such as Young Sound in NE Greenland where pelagic primary production is low (Glud et al., 2002). This production could then benefit benthic communities. Interestingly, vertical particulate organic carbon flux measured from long-term sediment traps in Queen Maud Gulf yielded estimates comparable to other sites in the Canadian Arctic, though total particulate matter flux was higher than in the adjacent regions reflecting additional inputs of terrestrial matter (Dezutter et al., 2021).

Low pelagic primary production also appears to be expressed in the higher trophic levels of the Kitikmeot Sea marine ecosystem: this region is where local residents predominately find the planktivorous, low-salinity preferring Arctic char. These fish only feed seasonally in productive patches of the ocean, and in the Kitikmeot Sea preferably in estuarine habitats, and then migrate up rivers to overwinter in lakes (Harris et al., 2020b). Residents mostly travel outside of the region to find higher densities of polar bears and ringed seals (Canadian Rangers from communities in the Kitikmeot region, personal communication) that are largely dependent on seals and Arctic cod, respectively, and greater biological production (Roff et al., 2020).

Potentially elevated vertical mixing in the vicinity of shallow sills might affect regional ecosystems. We have shown that sill depth exerts a fundamental control over the water column structure by regulating the balance between buoyancy and tidal mixing in the straits that determines bottom salinities and thus stratification in the Kitikmeot Sea. By extension, the sill also wields strong influence over the regional ecosystem. Near the surface, stratification regulates the entrainment of nutrients into the surface mixed layer, limiting the supply of new nutrients to the euphotic zone through summer months. The strong salinity stratification also limits the vertical extent of convection from sea-ice brines, as the upper mixed layer does not typically deepen beyond about 40 m depth in winter. Because brine convection does not reach the seafloor over the deeper basins, near-bottom temperatures in Coronation Gulf and Bathurst Inlet remain 1–2°C above the freezing point. This reservoir of sensible heat is a potential source for driving ice melt if it can be upwelled or mixed into the surface layer. Near-bottom temperatures that are close to, but still above, the freezing point would exclude many temperate and sub-Arctic species, but even cold-tolerant organisms often thrive as well or even better in waters that are above the freezing point (Renaud et al., 2019). As such, above-freezing temperatures may provide an important control on the benthic community species composition and respiration rates and, we speculate, on nutrient remineralization rates. In contrast, Queen Maud Gulf, with near freezing point bottom temperatures, may behave differently.

While basin-scale geomorphology and dynamic processes determine the overall physical structures and ecological functions of the Kitikmeot Sea, numerous processes interior to the system determine localized settings for marine life. While much more research is required to identify and quantify such processes, we highlight two here: first, the role of incoming rivers on interior transport; and second, the role of tidal currents that are amplified over shallow sills and/or narrow channels (Figure 17).

Riverine coastal domain. As noted above, the numerous rivers that surround the perimeter of the Kitikmeot Sea supply nutrients, organic matter, and sediments, affect the carbonate cycle and the timing and strength of stratification (Brown et al., 2020a). Near the river mouths, the low-salinity river plumes and their coastal buoyancy boundary currents mediate these factors. Additionally, individual river plumes at times may join to form a contiguous riverine coastal domain that is expected to extend generally 2–5 km offshore (the scale of the baroclinic Rossby radius) and can be enhanced or degraded by wind (Figure 7b; Carmack et al., 2015). This domain may be important via creating a pathway for longshore transport and possible migration habitat for anadromous fish such as char.

Tidal mixing. Two distinct tidal mixing phenomena can be identified. First, as noted in Section 3.1.2 we expect increased M2 dissipation in the relatively shallow waters of eastern Queen Maud Gulf (Figure 6) and thus the region may be one with increased nutrient supply to the upper ocean. In addition, these shallower waters may allow sufficient sunlight at depth to support a microphytobenthos community (Attard, 2024). Second, throughout the Kitikmeot Sea, tidal currents accelerate over sills and through narrow channels, creating turbulence and locally enhanced vertical mixing. Where resolved in the barotropic model of Rotermund et al. (2021), these channels are “hotspots” of tidal dissipation (Figure 6). As tidal flow accelerates though a channel, its total transport (T) remains constant as the minimum in width (W) and depth (H) is approached. Writing $T=UT WH$ we note that $H/UT3=W3H4/T3$⁠, highlighting the expected strong reduction in $H/UT3$ and increase in mixing at the sill. We propose that in winter the increased vertical mixing of heat upward leads to conditions of thin ice or open waters (polynyas) and, similarly, in summer the increased vertical flux of nutrients upward leads to increased potential for primary production. We refer to this scenario as the winter holes and summer gardens hypothesis, as shown schematically in Figure 17. Support for this concept comes from Dalman et al. (2019), who documented local thinning of ice and increased algal growth in areas of stronger currents, and from Bluhm et al. (2022), who noted that kelp sites away from the shore were characterized by strong tidal currents in passages where the necessary hard substrates prevail. Filter-feeding communities of, for example, soft corals, sponges, and suspension-feeding sea cucumbers prevailing at sills (Williams et al., 2018) suggest that adequate food particle densities are available in tidal passage to sustain these large-bodied, long-lived organisms. In slower-flow areas away from sills, densities of brittle stars comparable to several other Arctic regions (Fredriksen, 2018) may suggest that additional carbon sources beyond phytoplankton feed patchy seafloor communities, or that the hypothesized enhanced production in tidal passages settles out as food for benthos down-stream where flow rates decrease.

In summary, the Kitikmeot Sea in the south-central CAA is a complex system with positive estuarine circulation throughout the year due to the joint effects of disproportionately large freshwater discharge from the Canadian mainland and shallow, tidally mixed sills to the east and west that retain freshwater largely in the system. The resulting circulation and stratification limit the lateral advection of nutrients from adjoining areas of the CAA and constrain the vertical supply of nutrients to the sunlit layer of the sea thus creating oligotrophic surface waters. Local hotspots for pelagic and benthic algal production do occur in narrow and shallow regions of amplified tidal mixing in spring and summer, and analogous mixing in winter favors thin ice with sympagic algal formation or open water in these locations.

In this paper we have focused on fundamental, present-day oceanographic processes in the Kitikmeot Sea estuary by linking its internal structures and circulation to external forcing by freshwater inputs, tides, winds, and sea ice. This approach, we hold, is the essential first step toward anticipating future change in the Kitikmeot Sea estuarine region as a setting for marine life. However, all components of the Arctic Ocean system are changing rapidly (Landrum and Holland, 2020; Overland et al., 2020), and there is general agreement that the future Arctic Ocean will see warmer water temperatures, reduced sea-ice thickness and extent, increased precipitation, and potentially stronger storm winds (but see Chemke and Coumou, 2024). Changes may be further influenced through the collective effects of borealization from encroachment of sub-Arctic seas (Carmack and McLaughlin, 2011; Polyakov et al., 2020; Carmack, 2021) and rapid changes in their adjacent sub-Arctic draining watersheds (compare to Tank et al., 2023). Central questions emerge for the Kitikmeot Sea. How will such changes alter the estuarine circulation and associated stratification? How will such changes alter the light and nutrient climate? How will such changes constrain marine primary productivity? To address these questions we must inquire how the Kitikmeot Sea estuary currently responds to forcings, and how it might respond to future forcings.

While time-series observations of regional climate and oceanographic variables across the Kitikmeot Sea are sparse, we can extrapolate from observed trends across the CAA for insight into on-going and potential future change. Steiner et al. (2015) projected CAA time series of key climate variables into the future using the earth system models in the fifth assessment report under both the RCP8.5 and RCP4.5 scenarios of the United Nations Intergovernmental Panel on Climate Change, and compared projected changes for 2066–2085 relative to the 1986–2005 period. Under the RCP8.5 warming scenario sea-ice concentration is projected to reduce by almost 20% across the CAA by 2065, precipitation is projected to increase up to 25%, Arctic river discharge is projected to increase by 9.6%, with an earlier maximum runoff (freshet) by −12.6 days century⁻¹. Overall Steiner et al. (2015) predict a warmer and fresher surface ocean in the CAA under both the RCP8.5 and RCP4.5 scenarios.

Arctic change will alter the phenology of seasonal timing and patterns. Earlier break-up and later freeze-up will expose surface waters to increased solar radiation, near the time of summer solstice, and increased wind stress, driving vertical mixing of nutrients near the time of strong fall storms. Will these factors lengthen the growing season? Will mixing over the sills inject more surface-origin heat to depth, providing an amplifying feedback on the generation of winter holes? Will enhanced open water during fall allow storms to drive stronger boundary current flows, affecting ice mobility south through Victoria Strait? Will the expected thinning of sea ice lead to an increased number of localized areas of open water sites, currently present only as so-called hidden polynyas (e.g., Williams et al., 2007, their figure 9; Melling et al., 2015)?

When such changes noted above are applied to the estuarine dynamics of the Kitikmeot Sea, key ecosystem consequences emerge related to the amount of solar radiation and nutrient supply reaching the euphotic zone. On the one hand, increased freshwater loading combined with surface layer warming will increase stratification, thus further constraining the vertical resupply of nutrients. On the other hand, increases in wind strength combined with sea-ice retreat could result in greater wind mixing, particularly in the narrow, shallow straits where tidal mixing is already prominent. Regional effects may also come into play, manifest in differing impacts on the drivers of stratification in Coronation Gulf (more river input) versus Queen Maud Gulf (less sea-ice input). A related unknown surrounds the transfer of surface productivity to depth, particularly as benthic communities are important food sources for certain higher trophic levels of the area such as bearded seals. How sensitive are the Kitikmeot Sea benthic food webs to changes in lower trophic fluctuations in the surface; in other words, how strong is the pelagic-benthic coupling in this region?

Collectively, these anticipated changes will impact the lives and well-being of indigenous residents (Arctic Council, 2016; Falardeau et al., 2020). Inuit across the Kitikmeot have been observing changes in local weather and sea ice, noting in particular that both are becoming much less predictable, which makes safe travel on the land increasingly uncertain (e.g., McLennan et al., 2022). Kitikmeot communities are experiencing warmer winters and shorter snow seasons (Derksen et al., 2019); shore-fast first-year ice near communities is forming later in the fall and melting earlier in the spring (Stroeve et al., 2014), with increased roughness due to windier conditions during freezing (Segal et al., 2020). On land, warming is contributing to permafrost thaw and active layer deepening (Arctic Monitoring and Assessment Programme, 2017); coincident with increased river flows (Dery et al., 2016; Holmes et al., 2018), bank destabilization and slumping on land can impact water quality and deliver more sediments to marine system (Chin et al., 2016). In order to better prepare and adapt for these changes, Kitikmeot communities and management teams require new information about the effects of ongoing climate change on the connected land-ocean system. This need includes a more comprehensive characterization of the structure of the Kitikmeot marine ecosystem, extending from the physical controls on nutrient availability and primary productivity, through to zooplankton, fish, and marine mammals. A more comprehensive understanding of these components will create the necessary framework from which to predict the impacts of our changing climate on this unique Arctic region.

Data access is as follows:

Bathymetry:

• GEBCO grid: https://www.gebco.net/data_and_products/gridded_bathymetry_data/

• Uncertified Canadian Hydrographic Service soundings: https://data.chs-shc.ca/dashboard/map

Sea ice:

• Canadian Ice Service median sea-ice concentration: https://iceweb1.cis.ec.gc.ca/30Atlas/page1.xhtml?grp=Guest&

• Canadian Ice Service sea-ice thickness data: https://www.canada.ca/en/environment-climate-change/services/ice-forecasts-observations/latest-conditions/archive-overview/thickness-data.html

• Ocean Networks Canada Cambridge Bay ice profiling sonar data: https://data.oceannetworks.ca/DataSearch?locationCode=CBYIP&deviceCategoryCode=ICEPROFILER

Tides:

• Webtide: https://www.bio.gc.ca/science/research-recherche/ocean/webtide/index-en.php

River inflow:

• HydroSHEDS: https://www.hydrosheds.org/hydrosheds-core-downloads

• HydroRIVERS: https://www.hydrosheds.org/products/hydrorivers

• Environment and Climate Change Canada river discharge data:

https://www.canada.ca/en/environment-climate-change/services/water-overview/quantity/monitoring/survey.html

Precipitation:

• Environment and Climate Change Canada climate normals: https://climate.weather.gc.ca/climate_normals/index_e.html

Wind:

• European Centre for Medium Range Weather Forecasts ERA5 10 m wind: https://cds.climate.copernicus.eu/datasets/reanalysis-era5-single-levels?tab=overview

Temperature and salinity profiles:

• Available from Fisheries and Oceans Canada’s Marine Environmental Data Section Archive

(https://meds-sdmm.dfo-mpo.gc.ca) by request at:

https://isdm.gc.ca/isdm-gdsi/request-commande/form-eng.asp

Dissolved nutrients:

• Available from Fisheries and Oceans Canada’s Marine Environmental Data Section Archive

(https://meds-sdmm.dfo-mpo.gc.ca) by request at:

https://isdm.gc.ca/isdm-gdsi/request-commande/form-eng.asp

The supplemental files for this article can be found as follows:

Text S1. Estimate of freshwater contained in ice melt.

Text S2. Barotropic outflow timescale.

Text S3. Estimate of freshwater carried with barotropic outflow through the bounding straits.

Text S4. Magnitude of estuarine circulation.

Text S5. Formulation of R for the sill regions of the bounding straits.

Text S6. Estimate of change R from open water to ice cover.

Figure S1. Detailed maps of the major straits of the Kitikmeot Sea.

Figure S2. A map of the shallow waters and complex bathymetry of eastern and southern Queen Maud Gulf.

Figure S3. 2001–2020 ECMWF ERA5 winds during the mid-October to mid-July ice-covered season.

We would like to thank the Arctic Research Foundation and Canadian High Arctic Research Station for their ongoing extensive support of, and interest in, the oceanography of the Kitikmeot Sea under the Kitikmeot Sea Science Study. We would also like to thank Fisheries and Oceans Canada (DFO) for their support of the Kitikmeot Sea Science Study and for their collection of oceanographic data in the Kitikmeot Sea between 1995 and 2007. In particular, we thank DFO’s Canada’s Three Oceans project for extensive collection of data in Coronation Gulf and Bathurst Inlet in 2007. We would also like to thank the Captains and crews of the Arctic Ivik, CCGS Sir Wilfrid Laurier, and CCGS Louis S. St-Laurent. Graphics were prepared by Patricia Kimber of Tango Graphics.

This work was supported by Fisheries and Oceans Canada (DFO), with contribution from their Arctic Science Fund, a Postdoctoral Scholar Award from Woods Hole Oceanographic Institution, and an NSERC Visiting Fellowship to DFO. Winter data collection were supported by the Polar Continental Shelf Project and the Department of National Defense’s Canadian Rangers via the Canadian Rangers Ocean Watch.

The contact author has declared that none of the authors has any competing interests.

Contributed to conception and design: WJW, KAB, LMR, BAB, SLD, MD, FAM, SV, ECC.

Contributed to acquisition of data: WJW, KAB, MD, FAM, SV, ECC.

Contributed to analysis and interpretation of data: WJW, KAB, LMR, BAB, SLD, FAM, SV, ECC.

Drafted and/or revised the article: WJW, KAB, LMR, BAB, SLD, MD, FAM, SV, ECC.

Approved the submitted version for publication: WJW, KAB, LMR, BAB, SLD, MD, FAM, SV, ECC.

How to cite this article: Williams, WJ, Brown, KA, Rotermund, LM, Bluhm, BA, Danielson, SL, Dempsey, M, McLaughlin, FA, Vagle, S, Carmack, EC. 2025. Processes in the Kitikmeot Sea estuary constraining marine life. Elementa: Science of the Anthropocene 13(1). DOI: https://doi.org/10.1525/elementa.2024.00031

Domain Editor-in-Chief: Jody W. Deming, University of Washington, Seattle, WA, USA

Associate Editor: Stephen F. Ackley, Department of Geological Sciences, University of Texas at San Antonio, San Antonio, TX, USA

Knowledge Domain: Ocean Science