Skip to main content
Scientific Reports logoLink to Scientific Reports
. 2025 Jan 10;15:24. doi: 10.1038/s41598-024-81994-8

Subsurface warming associated with Pacific Summer Water transport toward the Chukchi Borderland in the Arctic Ocean

Miaki Muramatsu 1,2,#, Eiji Watanabe 1,✉,#, Motoyo Itoh 1, Jonaotaro Onodera 1, Kohei Mizobata 3, Hiromichi Ueno 2,4
PMCID: PMC11723998  PMID: 39794398

Abstract

Recent rapid sea ice reduction in the Pacific sector of the Arctic Ocean is potentially associated with inflow of Pacific-origin water via the Bering Strait. For the first time, we detected remarkable subsurface warming around the Chukchi Borderland in the Arctic Ocean over the recent two decades (i.e., the early 21st century). A statistically significant decadal trend of 16.6 ± 10.6 MJ m− 2 year− 1 in the subsurface ocean heat content during 1999–2020 was captured by shipboard hydrographic data, and associated with the transport of Pacific Summer Water from Barrow Canyon northwest of the Alaskan coast, where similar warming appeared. Satellite-derived geostrophic ocean velocity indicated that the northwestward ocean current flowing from Barrow Canyon to the Chukchi Borderland became faster in the late 2010s, in association with southeastward shift of the Beaufort Gyre, circulating clockwise around the Canada Basin. Therefore, we suggest that warming of the Pacific Summer Water passing over the Chukchi shelf and intensification of the northwestward ocean current along the shelf–basin boundary both acted to enhance the heat transport and contributed to the positive trend in downstream subsurface ocean heat content. Our findings fill important gaps in the understanding of ocean heat distribution/transport, which is a key factor for sea ice freezing/melting, in the central Arctic.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-024-81994-8.

Subject terms: Physical oceanography, Cryospheric science

Introduction

Sea ice reduction in the Pacific sector of the Arctic Ocean has become remarkable over the past decades, and is associated with inflow of Pacific-origin water through the Bering Strait1,2. The annual mean volume transport of the Bering Strait throughflow increased from 0.7 Sv (1 Sv ≡ 106 m3 s− 1) in 2001 to 1.2 Sv in 20143. The ocean heat transport through the Bering Strait has also increased. The trend in the annual heat transport during 1991–2018 was 8.3 ± 3.9 × 1018 J year− 13. The pathway of Pacific-origin water separates into three routes that follow seafloor topography in the Chukchi Sea4 (Fig. 1a). Most of the water then reaches Barrow Canyon in the northeastern Chukchi Sea57. The Pacific-origin water flowing across the Chukchi Sea shelf in summer is traditionally called Pacific Summer Water (PSW). The PSW temperature is further increased by solar heating following sea ice retreat over the shelf8. The hydrographic properties of the Pacific-origin water in the western Arctic basin vary on multidecadal timescales. The PSW in the basin showed a warming trend between 1997–and 2002 and 2007–20089, and the thickness of its subsurface layer increased between the late 2000s and the early 2010s10. It has also been suggested that the subsurface ocean heat content (OHC) in the Canada Basin doubled from 1987 to 2017, induced by wind-driven northward subduction from the northern Chukchi Sea10. However, in addition to the direct route of PSW into the southern Canada Basin, two other pathways have been suggested.

Fig. 1.

Fig. 1

(a) Seafloor topography in the western Arctic. Gray contours denote 20, 50, 100, 200, 500, 2000, and 3000 m isobaths. Black contour shows 1000 m isobath. Black arrows indicate the western transport routes of the Pacific Summer Water (PSW) in the Chukchi Sea. Red arrow indicates the Alaskan Coastal Current: the primary PSW pathway. Yellow arrow shows the Chukchi Slope Current. Red segment in the northeastern Chukchi Sea corresponds to the Barrow Canyon section. White arrows denote the anticyclonic Beaufort Gyre in the Canada Basin. (b) The subsurface ocean heat content (OHC) obtained at the CTD/XCTD stations of the R/V Mirai cruises for 1999–2020 [MJ m–2]. Only the stations in the southern Chukchi Borderland region (sCBL) are shown. Black contours indicate 200 m and 3000 m isobaths, which were used to define the sCBL. (c) Black dots and dashed line show interannual time series of the subsurface OHC averaged in each year [MJ m–2] and its decadal trend [MJ m–2 year–1], respectively, for 1999–2020 in the sCBL. The trend value is presented in the top of each panel. The spatial variabilities are represented by (black error bars) the standard deviation and (blue error bars) the range between minimum and maximum values in each year. (d) Relationship among the subsurface OHC [MJ m–2], maximum temperature [°C], and thickness (z2z1) [m] in the PSW layer.

After passing through Barrow Canyon, a substantial amount of Pacific-origin water is plausibly transported along the Beaufort/Chukchi Sea shelf slope. For example, the water mass exported from Barrow Canyon has been reported to be transported eastward onto the Beaufort Sea shelf11. The westward transport has also been observed12 and simulated13,14. The northwestward narrow intense flow along the northern slope in the Chukchi Sea is called the Chukchi Slope Current15,16 (Fig. 1a). The vertical structure and seasonality of the Chukchi Slope Current have been investigated using shipboard and mooring-based hydrographic/velocity data collected across the northern slope17. Satellite-tracked drifting floats captured the Chukchi Slope Current reaching Herald Canyon in the northwestern Chukchi Sea18 and the Chukchi Borderland19. It has been estimated that the time of PSW advection between the mouths of Barrow Canyon and Hanna Canyon is less than 2 months, and that the PSW following this route was warmer and fresher in the late 2010s than in the early 2000s20. Further, it has been reported that the subsequent current traces three pathways21. A part of the Chukchi Slope Current continues to flow westward into the Makarov Basin across the Chukchi Plateau. The second pathway turns northward along the western edge of the Beaufort Gyre. The third one exhibits direct intrusion into the Canada Basin. The variability in the Chukchi Slope Current is sometimes modulated by the Beaufort Gyre22.

Comprehensive understanding of the modification and transport of PSW along the Chukchi Sea northern slope would contribute to quantification of the ocean heat budget of the western Arctic basin. Despite knowing that the PSW is transported into the Chukchi Borderland by the slope current, the corresponding heat changes around the Chukchi Borderland on multidecadal timescales are not clearly understood. Because the Chukchi Borderland is adjacent to the Canada Basin (Fig. 1b), the PSW reaching the Chukchi Borderland potentially contributes to subsurface warming in the Canada Basin and other regions of the central Arctic.

Here, we present the decadal variability in the subsurface OHC using shipboard hydrographic data obtained by the R/V Mirai of the Japan Agency for Marine–Earth Science and Technology (JAMSTEC) during 1999–2020 (https://ads.nipr.ac.jp/data/meta/A20241203-001). We then investigate the ocean heat properties in Barrow Canyon using mooring-based time series data5,23. Furthermore, we clarify the relationship among the interannual variabilities in the subsurface OHC around the Chukchi Borderland, the ocean current along the shelf–basin boundary, and the Beaufort Gyre using satellite-derived high-resolution fields of dynamic ocean topography and geostrophic ocean velocity24.

Results

Subsurface ocean heat content in the Chukchi Borderland and Barrow Canyon

In the present study, the southern Chukchi Borderland region (hereafter, sCBL; 0.12 Mkm2) was defined by the area 73°–76.5°N, 155°–170°W, and seafloor depth of 200–3000 m (Fig. 1b). The subsurface OHC averaged in the sCBL increased from 316 MJ m− 2 in 1999 to 571 MJ m− 2 in 2020 (Fig. 1c). A significant positive decadal trend of 16.6 ± 10.6 MJ m− 2 year− 1 (p = 0.01) was detected. Note that the error values of decadal trends correspond to 95% confidence interval and the significance was diagnosed by the Mann-Kendall test throughout this article. The subsurface OHC was 280–350 MJ m− 2 for the earlier 4 years (1999, 2000, 2002, and 2004) and 570–880 MJ m− 2 for the latter 4 years (2017–2020). The maximum OHC of 1543 MJ m− 2 was recorded north of Hanna Canyon on September 10, 2017 (Supplementary Fig. S1).

The relationship among the subsurface OHC, thickness (i.e., z2z1 defined in the Method section), and maximum temperature (Tmax) in the PSW layer shows that both the greater volume and higher temperature of the PSW contributed to the larger OHC (Fig. 1d). Actually, the layer thickness and Tmax of the PSW both increased during 1999–2020: 1.18 ± 1.29 m year− 1 (p = 0.08) and 0.07 ± 0.05 °C year− 1 (p = 0.01), respectively (Supplementary Figs. S2a and S2b). During the same period, Tmax deepening (0.54 ± 0.63 m year− 1, p = 0.09) and freshening (salinity of − 0.01 ± 0.03 year− 1, p = 0.67) were seen (Supplementary Fig. S2c and S2d). A significant correlation with subsurface OHC was detected for only the layer thickness and Tmax of the PSW (r = 0.54–0.66, p < 0.05; Supplementary Figs. S2a and S2b).

The relationship between interannual variabilities in the subsurface OHC in Barrow Canyon and the sCBL was investigated to elucidate PSW transport between the two regions (Fig. 2). The subsurface OHC in Barrow Canyon, with its peak value of 947 MJ m− 2 in 2017, increased from the early 2000s to the late 2010s, and the correlation coefficient of interannual variabilities in the subsurface OHC between the two regions was 0.79 (p = 0.01) (Fig. 2b). Note that decadal trends were removed from the interannual time series for calculation of correlation coefficients throughout this article. The insignificant decadal trend of 9.4 ± 19.2 MJ m− 2 year− 1 (p = 0.32) was smaller than that of the sCBL. The decadal trend and strong correlation suggest that subsurface warming in the sCBL could be attributed to that in Barrow Canyon. However, the different magnitude of the OHC trend between the two regions indicates that additional mechanisms influenced the warming around the Chukchi Borderland.

Fig. 2.

Fig. 2

(a) Location of three moorings deployed along the Barrow Canyon section (Stations BC-W, BC-C, and BC-E). Interannual time series of (b) the subsurface ocean heat content (OHC) [MJ m–2], (c) northeastward heat transport [TW], and (d) northeastward volume transport [Sv] in the Barrow Canyon section for 2002–2020, when the mooring measurements have been conducted. Red dots and dashed lines show the averages for July–September in each year and those decadal trends, respectively. The trend values are presented in the top of each panel. Black dots are the same as those in Fig. 1c. Correlation coefficients between these time series of the Barrow Canyon properties and the subsurface OHC in the southern Chukchi Borderland region are also presented in the right bottom of each panel. Red (blue) values mean positive (negative), and bold fonts show significant values; diagnosed by the Mann-Kendall test for the trend and p < 0.05 for the correlation.

Relationship with Shelf–Basin Boundary Current and Beaufort Gyre

The interannual variabilities in the northwestward ocean current flowing along the shelf–basin boundary and the anticyclonic Beaufort Gyre during 2011–2020 were also analyzed using the satellite-derived datasets of dynamic ocean topography and geostrophic ocean velocity24. First, the velocity across the four boundaries used to define the sCBL was examined (Fig. 3). The northward velocity across the southern boundary (73°N) reveals a strong current along the Chukchi Sea northern slope at approximately 155°–160°W (Fig. 3a). Notably, a horizontal velocity of > 15 cm s− 1 was frequently detected after 2017. The average velocity in the region 155°–160°W in July–September increased by 20% from 7.5 cm s− 1 in 2011–2016 to 9.0 cm s− 1 in 2017–2020. Although the westward flow into the Chukchi Borderland across the southern half of the eastern boundary (73°–74°N, 155°W) captured part of the Beaufort Gyre (Fig. 3d), the weaker current across the eastern boundary (e.g., 2.6 cm s− 1 for 73°–74°N in July–September, 2011–2020) indicates that the heat inflow from the Canada Basin via the Beaufort Gyre was less important in comparison with that of the shelf–basin boundary route.

Fig. 3.

Fig. 3

Hovmöller diagram of the geostrophic ocean velocity [cm s− 1] across the four boundaries around the Chukchi Borderland, which are shown by red lines in a right map, for 2011–2020. (a, b) Positive (negative) values indicate northward (southward) velocity. (c, d) Positive (negative) values indicate westward (eastward) velocity.

Here, the reason for the stronger northwestward ocean current after 2017 is discussed from the perspective of interannual changes in the position and extent of the Beaufort Gyre (Fig. 4; see those definitions in the Method section). While the Beaufort Gyre expanded west of the Chukchi Borderland before 2015, the gyre edge was located east of 165°W after 2017 (Fig. 4a, Supplementary Fig. S3). Simultaneously, the position of the gyre center averaged for July–September shifted southeastward from the early 2010s to the late 2010s (Fig. 4b). Additionally, the maximum dynamic ocean topography (DOTmax; see the definition in the Method section) averaged for July–September increased from 27 cm in 2011 to 31 cm in 2020, and the total extent of the Beaufort Gyre region decreased from 0.39 Mkm2 to 0.30 Mkm2 for the same period (Fig. 4c). The recent stronger ocean current along the shelf–basin boundary was consistent with the enhancement of the DOTmax and the shrinking and southeastward shift of the Beaufort Gyre. Therefore, these decadal changes in the Beaufort Gyre could support our supposition that the stronger ocean current enhanced the PSW transport toward the Chukchi Borderland, thereby causing the larger subsurface OHC in the late 2010s.

Fig. 4.

Fig. 4

(a) Spatial distribution of the dynamic ocean topography (DOT) averaged for July–September (JAS) in 2017 [cm]. Blue contour shows the edge of the Beaufort Gyre (see the definition in the Method section). Purple star shows the location of the maximum DOT (DOTmax) inside the defined western Arctic basin (black-dashed region). Cyan star indicates the weighted center of the Beaufort Gyre. Red rectangle corresponds to the same region as those in Figs. 1b and 3. Purple arrows show the geostrophic ocean current faster than 15 cm s− 1. (b) Interannual variability in the weighted center location of the Beaufort Gyre averaged for JAS 2011–2020. (c) Interannual monthly time series of (red) DOTmax and (blue) the total area inside the edge of the defined Beaufort Gyre.

The outward flows from the Chukchi Borderland across the western (170°W) and northern (76.5°N) boundaries were also examined. It is evident that the westward velocity across the western boundary at 74.5°–76°N decreased or even changed to an eastward velocity (Fig. 3c). For the same period, the northward velocity across the northern boundary, especially west of 165°W, was intensified after 2017 (Fig. 3b). These velocity changes suggest that the direction of PSW transport around the Chukchi Borderland shifted from westward to northward in accordance with the southeastward shift of the Beaufort Gyre.

The Beaufort Gyre properties are primarily driven by the anticyclonic winds around the Beaufort High. The sea level pressure fields obtained from the Climate Forecast System Reanalysis dataset25 for 2001–2010 and the Climate Forecast System version 2 dataset26 for 2011–2020 of the National Centers for Environmental Prediction (hereafter CFSR) show that the center of the Beaufort High was located over the northern Chukchi Borderland in the early 2000s and over the southern Canada Basin in the late 2010s (Supplementary Fig. S4). This relocation would have been favorable for producing the southeastward shift of the Beaufort Gyre. Actually, this wind-driven shift of the Beaufort Gyre in the 2010s was indicated by another approach27 and could account for high content of freshwater owing to its accumulation in the Canada Basin28.

Discussion

For the first time, we detected remarkable subsurface warming around the Chukchi Borderland over the recent two decades. The statistically significant decadal trend of 16.6 ± 10.6 MJ m− 2 year− 1 in subsurface OHC during 1999–2020 was captured by the shipboard CTD/XCTD data and associated with PSW transport from Barrow Canyon, where a similar trend in subsurface OHC has been found. Moreover, the satellite-derived geostrophic ocean velocity indicated that the northwestward ocean current flowing from Barrow Canyon to the Chukchi Borderland became faster in the late 2010s, in association with southeastward shift of the Beaufort Gyre. Therefore, we suggest that warming of the PSW passing over the Chukchi shelf and intensification of the ocean current along the shelf–basin boundary both enhanced the northwestward ocean heat transport and contributed to the positive trend in downstream subsurface OHC.

The relationship between the interannual variabilities in subsurface OHC in the sCBL and northeastward heat transport along Barrow Canyon was also analyzed. The subsurface heat transport averaged for July–September in 2002–2020 was in the range of approximately 1–8 TW (Fig. 2c). The correlation coefficient between those time series was 0.58 (p = 0.10); however, this heat transport exhibited no significant decadal trend (2.6 ± 16.3 × 10− 2 TW year− 1, p = 0.75). During the same period, the northeastward volume transport along Barrow Canyon gradually decreased (− 5.0 ± 7.3 × 10− 3 Sv year− 1, p = 0.17; Fig. 2d). The weakening of the Barrow Canyon throughflow could have been associated with the southeastward shift of the Beaufort Gyre, which potentially blocked the coastal current. Even when the top depth for the vertical integration was changed from 30 m to 0, 20, and 40 m, the interannual variability and decadal trend in the OHC and heat/volume transports for 2002−2020 were confirmed as similar (Supplementary Fig. S5). A key point here is that the PSW pathway turned from northeastward through Barrow Canyon to northwestward along the shelf–basin boundary. It was previously reported that weakening of the former along-canyon current and intensification of the latter boundary current simultaneously occur under the same wind pattern14. Thus, the observed subsurface warming in the sCBL could not be explained directly by the northeastward heat transport along Barrow Canyon. The subsequent northwestward transport of the PSW that reached the Barrow Canyon mouth should be paid attention to account for the OHC variability around the Chukchi Borderland. The larger OHC trend in the sCBL than that in Barrow Canyon would be attributed to the accelerating boundary current from the canyon mouth.

The shortwave absorption at the ocean surface averaged in the sCBL and integrated for July–September (see the data procedure in the Method section) shows significant positive decadal trend (7.2 ± 6.1 MJ m− 2 year− 1, p = 0.02) according to gradual sea ice retreat (Supplementary Figs. S6a and S6b). Here, definition of the PSW layer (see the Method section and Supplementary Fig. S7) is considered to be a factor for uncertainties on our story, because deepening of the Near-Surface Temperature Maximum29 might modulate the subsurface OHC. To address this potential bias from a viewpoint of Ekman pumping, we calculated the wind stress curl using the CFSR dataset25,26, without considering sea ice and ocean surface currents for simplicity. The magnitude of the Ekman downwelling averaged in the sCBL and integrated for July–September in each year was smaller than 5 m (Supplementary Fig. S6c). The sea ice coverage would damp the Ekman contribution to deepening of the surface warm water, especially under the convergent wind stress fields (i.e., the negative curl). Besides the mixed layer depth in the sCBL was mostly shallower than the PSW layer as shown in vertical salinity profiles (Supplementary Fig. S9), probably owing to massive sea ice meltwater and riverine water. Thus we could not find confident mechanisms for penetration of local solar heat to the PSW layer (located mostly deeper than 30 m). Moreover, the OHC integrated from the ocean surface to the top depth of the PSW layer (i.e., z1) showed no significant decadal trend (2.7 ± 5.5 MJ m− 2 year− 1, p = 0.32), although the positive trend of 19.7 ± 12.8 MJ m− 2 year− 1 in the total OHC for the surface and PSW layers was still significant (p = 0.01) (Supplementary Fig. S8). These results imply that a substantial part of solar heat absorbed locally in the surface layer was released to atmosphere as the upward longwave radiation and sensible heat flux before sea ice freezing (i.e., minor impact on subsurface warming in the PSW layer). It is rather plausible that solar heating over the Chukchi Sea could warm up the PSW flowing in the shallower layer of the upstream region8. In addition, when the PSW has experienced freshening and shoaling as revealed in recent studies3, the PSW should have the larger decadal trend in its OHC than that assuming the top depth adopted in the present analyses (see the Method section). Besides, we suggest that the maximum temperature near the defined top depth of the PSW layer in 2018 and 2019 was also derived from intrusion of a shallower part of PSW because of no underlying temperature peak and no substantial local Ekman downwelling for these periods (Supplementary Figs. S6c and S9). Therefore, definition of the PSW layer would not alter major conclusions in the present study.

We recognize the difficulty to evaluate how the northwestward ocean current along the shelf–basin boundary detected by the satellite-derived DOT at the ocean surface reflects the Chukchi Slope Current15. The spatial resolution of the DOT dataset analyzed in the present study is approximately 20 km (0.5° × 0.2°)24, which is higher than that of similar products (0.75° × 0.25°)30,31. However, application of the Gaussian filter of 80 km for noise reduction would unable the corresponding geostrophic ocean velocity to explicitly resolve the Chukchi Slope Current, whose width was approximately 50 km estimated by previous mooring measurements22.

The potential gap from the subsurface ocean current controlling the PSW transport was then addressed using the acoustic Doppler current profiler (ADCP) datasets from four mooring sites around the Chukchi Borderland (Chukchi Abyssal Plain, CAP, 2010‒2012; Northwind Abyssal Plain, NAP, 2013‒2014 and 2018‒2019; North of Hanna Canyon, NHC, 2015‒2019; North of Barrow Canyon, NBC, 2015‒2018; see the exact locations in Supplementary Fig. S10). The detailed information of these mooring measurements was reported in our previous publications14,32 and a data archive (https://ads.nipr.ac.jp/dataset/A20191217-002). Notably, the ADCP profiles in the surface 10 m have low accuracy owing to interference with the sea ice keel. Here, we confirmed that the current direction and speed were similar between the depths of approximately 15 and 40 m (i.e., mean Tmax depth in the PSW layer; Supplementary Fig. S2), except on two occasions (e.g., subsurface southward current in December 2010 at Station CAP, and surface northward current in November‒December 2016 at Station NBC). Thus, potential bias of velocity attributable to the different depths was considered minor for most of our findings.

It was indicated that the Chukchi Slope Current was sometimes intensified under the positive wind stress curl in the northeastern Chukchi Sea on weekly time scales17. We checked the decadal variability to address this relationship. The wind stress curl averaged in the northeastern Chukchi Sea (here, enclosed by 155°–170°W, 69°N, and seafloor depth of 100 m) and integrated for July–September reached a positive peak for 2016–2017 (Supplementary Fig. S6c). On the other hand, it was also assumed that the northwestward ocean current along the shelf–basin boundary was modulated by wind patterns in the wider area including the Beaufort Sea. The wind stress curl in the southern Canada Basin (enclosed by 75°N and seafloor depth of 3000 m) was decreasing from the mid-2000s to the early 2010s (Supplementary Fig. S6c). Whereas no significant decadal trends in the wind stress curl were seen, combination of the positive and negative anomalies relative to mean states in two regions may be favorable for intensification of the northwestward ocean current.

Actually, the intensification of the northwestward ocean current along the shelf–basin boundary for the recent decades was consistent with previous mooring and modeling analyses. The ADCP data at the mooring station of the Chukchi Sea northern slope captured the faster current in the late 2010s than that in the early 2000s20. The interannual experiment for 2001–2020 using an eddy-resolving sea ice–ocean model simulated a remarkable positive trend in the volume transport for the depth of 0–200 m along this pathway, providing a substantial amount of the resuspended seafloor sediment toward the Chukchi Borderland14.

On a broader scale, the stepwise reduction of the Arctic sea ice was explained by the shortened residence time according to acceleration of the Transpolar Drift Stream toward the Fram Strait33. Moreover, the enhanced role of ocean heat transport through the major Arctic gateways (i.e., the Barents Sea Opening, Fram Strait, and Bering Strait) and a poleward shift of its greater impact on sea ice variability were simulated by ensemble historical and future experiments using climate models34,35. It was also reported that the anomalously warm PSW delayed the southward advance of the sea ice area on the Pacific side36. Our findings further elucidate the ocean heat transport from the Chukchi shelf to the western Arctic basin, especially from Barrow Canyon to the Chukchi Borderland, which is one of the key processes influencing change in sea ice on a regional scale. However, the PSW transport to regions further downstream in the central Arctic is still unknown. Therefore, more in-depth investigations on the detailed pathways and heat transport of PSW to the deeper basins should be conducted. Heat budget analyses, including vertical mixing with the surface layer, are also necessary for more quantitative evaluation of the impact of PSW on the freezing/melting of overlying sea ice. Numerical model experiments would be important in achieving this objective. The target region in the present study is regarded as a potential fishable area where “the agreement to prevent unregulated high seas fisheries in the central Arctic Ocean” was acknowledged in 2018 and entered into force in 2021 (https://www.fao.org/faolex/results/details/en/c/LEX-FAOC199323/). The detected subsurface warming would influence the habitat shift and biomass of fishery resources37.

Methods

Hydrographic data

The CTD/XCTD hydrographic data obtained during the cruises of the R/V Mirai in the sCBL during 1999–2020 were collected to calculate the subsurface OHC (Fig. 1b, Supplementary Table S1). The original datasets with more detailed information are available at the JAMSTEC website: Data and Sample Research System for Whole Cruise Information (http://www.godac.jamstec.go.jp/darwin/). The R/V Mirai cruises in the Arctic Ocean did not operate in 2001, 2003, 2005, 2007, and 2011, and the cruise in 2006 did not cover the sCBL. Profiles recording salinity with values of < 20 or > 36 were eliminated as errors. Note that practical salinity with no unit (https://www.teos-10.org/) is presented throughout this article. In most target years, the CTD/XCTD observations were conducted at approximately 160°W and 74°N over the base of the Northwind Ridge (Supplementary Fig. S1). The stations in 2002, 2004, and 2008–2010 were extensively distributed over the sCBL, and those in 2012, 2013, 2015, and 2017 were located along seafloor depth of 500 m. Most of the stations in 2015 and 2018 were located near the shallow Chukchi Sea shelf. Only one profile was obtained in 2019. In 2010, relatively large OHC was produced by the existence of a distinct warm eddy (temperature: >6 °C) over the southern Northwind Ridge (Supplementary Fig. S1), which has been examined in earlier studies38,39. We excluded the period of this warm eddy in the evaluation of the subsurface OHC.

Ocean heat content

Subsurface OHC was calculated using in situ temperature and salinity data obtained at each CTD/XCTD station (Supplementary Fig. S9). The vertical resolution of the CTD/XCTD data was 1 m. To focus on the PSW transported to the subsurface layer around the Chukchi Borderland, the Near-Surface Temperature Maximum (NSTM)29 and Atlantic Water were excluded in the OHC calculation. In most vertical profiles, two types of temperature peaks were detected (Supplementary Fig. S7). The shallow peak for salinity of 24–29 was derived from the NSTM, which was not the target of the present study. The deep peak for salinity of > 29 was produced by the PSW. Therefore, either the depth of water with salinity of 29 or the depth of 30 m was defined as the top of the PSW layer (z1), following a previous study18. The depth of the temperature minimum located deeper than 100 m was defined as the bottom of the PSW layer (z2). The subsurface OHC was then evaluated through vertical integration in the PSW layer as follows:

graphic file with name M1.gif 1

where the seawater density ρ and the specific heat of seawater C were estimated using the CTD/XCTD hydrographic and pressure data; T is in situ temperature; and Tf is the freezing point depending on salinity at each station and depth. It was indicated that absolute values of the OHC are arbitrary because of its dependency on the reference temperature40. However, we chose the freezing point of seawater Tf as the reference temperature to quantitatively provide a potential impact of the subsurface ocean heat on sea ice melting. This choice has usually been adopted in polar seas and enables the direct comparison of the estimated OHC with previous publications41. The difference between a constant temperature (e.g., − 1.8 ºC) or values depending on seawater salinity hardly affects uncertainty of the defined OHC. Besides, amplitude of the interannual variabilities and multidecadal trends are independent of the reference temperature.

Shortwave absorption

Shortwave absorption at the ocean surface (SWabs) was calculated as follows:

graphic file with name M2.gif 2

where SWdw and SIC are downward shortwave radiation and sea ice concentration, respectively, obtained from the CFSR dataset25,26. αo is ocean surface albedo set to 0.1. We confirmed that SWabs integrated for July–September accounts for a most part of the annual total value in the sCBL (Supplementary Fig. S6a).

Dynamic ocean topography

The satellite-derived fields of DOT and geostrophic ocean velocity (https://ads.nipr.ac.jp/data/meta/A20210401-001) were used to investigate the spatiotemporal variabilities in major ocean currents including the Beaufort Gyre. The DOT was estimated by an altimeter onboard CryoSat-242. The spatial resolution of the latest gridded monthly composite is approximately 20 km (0.5° × 0.2° longitude–latitude)24. This high-resolution DOT dataset enables detection of narrow intense geostrophic ocean currents along the shelf‒basin boundary.

In the present study, the DOTmax was defined by the maximum DOT, which was recorded in the western Arctic basin defined by 170°E, 80°N, 130°W, and seafloor depth of 100 m (Fig. 4a, Supplementary Fig. S3). The closed contour where the DOT anomaly from DOTmax is − 20 cm was regarded as the edge of the Beaufort Gyre. The weighted center of the Beaufort Gyre was evaluated following a previous study43:

graphic file with name M3.gif 3

where lonc (latc) is the longitude (latitude) of the weighted center. Here, loni, lati, and DOTi are the longitude, latitude, and DOT in each grid of the dataset inside the edge of the Beaufort Gyre. The Beaufort Gyre region has sometimes been defined using geographical longitude–latitude and bathymetry44. A couple of studies adopted the lowest closed contour of DOT (i.e., the largest area) around its maxima43,45,46. In another case, multiplication of the maximum DOT value and a factor of 0.6 was used to obtain the Beaufort Gyre extent27. Based on various trials, we judged that the definition in the present study was more suitable to evaluate the DOT gradient influencing ocean currents toward the Chukchi Borderland in the southwestern Beaufort Sea.

Barrow Canyon properties

Barrow Canyon is the primary gateway of PSW transport from the Chukchi shelf to the western Arctic basin. JAMSTEC has conducted mooring measurements at three stations in Barrow Canyon since October 20015,23. To quantify the canyon properties, the Barrow Canyon section was defined by the line from Station BC-W (71°48’N, 155°21’W) to the point of 71°36’N, 154°49’E north of Point Barrow via Stations BC-C (71°44’N, 155°10’W) and BC-E (71°40’N, 155°00’W). The horizontal interval between each mooring station is approximately 10 km, and the total section distance is 28.7 km. The seafloor depth is approximately 100 m at Stations BC-W and BC-E, and approximately 300 m at Station BC-C. The OHC at the depth of 30–100 m in the section was averaged for July−September in each year, assuming the advection time of less than 2 months from Barrow Canyon to the 73°N line over the northern slope of the Chukchi Sea20. The layer with depth of 0–30 m was excluded because surface ocean heat could be lost by autumn cooling before reaching the Chukchi Borderland20. It was also known that the layer deeper than 100 m is occupied by the cold Pacific Winter Water and Atlantic Water23,47.

The northeastward heat and volume transports (HT and VT) across the Barrow Canyon section were calculated following previous studies5,23:

graphic file with name M4.gif 4
graphic file with name M5.gif 5

where ρ, C, T, and Tf are the same as in Eq. (1), υ is the northeastward velocity measured by the ADCP, x is the location along the Barrow Canyon section, where x1 (x2) is the eastern (western) end point. Here, it should be kept in mind that the seawater freezing temperature was often referenced to quantify the ocean heat transport3, although uncertainty of the absolute values is derived from this choice as well as the OHC40. Besides, the Barrow Canyon properties had to be calculated using the hydrographic data of discrete moored sensors, whereas the OHC in the sCBL could be obtained from the shipboard continuous vertical profiles. Hence we decided to use the depth levels for the vertical integration.

The time series data of the transport from October 2001 to August 2008 were previously constructed5. The OHC and heat/volume transports for September 2008 to September 2010 and for September 2013 to September 2014 could not be calculated because of the lack of mooring deployments. The volume transport until July 2019 was recently analyzed14, and the dataset of the OHC and heat/volume transports was then extended to September 2020 in the present study. The detailed estimation procedure was described in previous publications5,23, where potential uncertainties of these Barrow Canyon properties were estimated to be less than 10% based on comparison with various high-resolution shipboard CTD and ADCP data.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1 (4.7MB, docx)

Acknowledgements

This study was supported by the Arctic Challenge for Sustainability II (ArCS II: JPMXD1420318865) project and the Grant-in-Aid for Scientific Research of Japan Society for the Promotion of Science (JSPS) (KAKENHI 22221003, 15H01736, and 18H03368) funded by the Ministry of Education, Culture, Sports, Science and Technology (MEXT). We are greatly indebted to the captains, crews, scientists, and technicians aboard the R/V Mirai, R/V Araon, USCGS Healy, CCGS Sir Wilfrid Laurier, and Amundsen for cruise operations and data collections. Profs. Yutaka Watanuki and Akihide Kasai at Hokkaido University gave us many constructive comments. We thank James Buxton MSc, from Edanz (https://jp.edanz.com/ac), for editing a draft manuscript. Finally, thoughtful and constructive comments from two anonymous referees were highly valuable to enhance quality of this article.

Author contributions

M.M. performed major analyses and wrote the draft manuscript with E.W. and H.U. in her PhD study. E.W. is responsible for the final version of manuscript as the equally first and corresponding author. The mooring and satellite-derived datasets were provided by M.I., J.O., and K.M. All authors contributed to the interpretation of the results and improve the manuscript.

Data availability

The dataset of the subsurface OHC and the related variables in the southern Chukchi Borderland region is available at the Arctic Data archive System (ADS) (https://ads.nipr.ac.jp/data/meta/A20241203-001). The DOT and geostrophic ocean velocity datasets for 2011–2020 were also registered to the ADS (https://ads.nipr.ac.jp/data/meta/A20210401-001). The CFSR (https://rda.ucar.edu/datasets/ds093.0/) and CFSv2 (https://rda.ucar.edu/datasets/ds094.0/) datasets are provided by the NCAR Research Data Archive.

Declarations

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Miaki Muramatsu and Eiji Watanabe have contributed equally to this work.

References

  • 1.Shimada, K. et al. Pacific Ocean inflow: influence on catastrophic reduction of sea ice cover in the Arctic Ocean. Geophys. Res. Lett.33, L08605. 10.1029/2005GL025624 (2006). [Google Scholar]
  • 2.Woodgate, R. A., Weingartner, T. & Lindsay, R. The 2007 Bering Strait oceanic heat flux and anomalous Arctic sea-ice retreat. Geophys. Res. Lett.37, 01602. 10.1029/2009GL041621 (2010). [Google Scholar]
  • 3.Woodgate, R. A. & Peralta-Ferriz, C. Warming and freshening of the Pacific inflow to the Arctic from 1990–2019 implying dramatic shoaling in Pacific Winter Water ventilation of the Arctic water column. Geophys. Res. Lett. 48, e2021GL092528. (2021). 10.1029/2021GL092528
  • 4.Weingartner, T. et al. Circulation on the north central Chukchi Sea shelf. Deep Sea Res. Part. II. 52 (24–26), 3150–3174. 10.1016/j.dsr2.2005.10.015 (2005). [Google Scholar]
  • 5.Itoh, M., Nishino, S., Kawaguchi, Y. & Kikuchi, T. Barrow Canyon volume, heat, and freshwater fluxes revealed by long-term mooring observations between 2000 and 2008. J. Geophys. Res. Ocean.118, 4363–4379. 10.1002/jgrc.20290 (2013). [Google Scholar]
  • 6.Gong, D. & Pickart, R. S. Summertime circulation in the eastern Chukchi Sea. Deep Sea Res. Part. II. 118, 18–31. 10.1016/j.dsr2.2015.02.006 (2015). [Google Scholar]
  • 7.Pickart, R. S. et al. Circulation of winter water on the Chukchi shelf in early summer. Deep Sea Res. Part. II. 130, 56–75. 10.1016/j.dsr2.2016.05.001 (2016). [Google Scholar]
  • 8.Tsukada, Y. et al. Interannual variation in solar heating in the Chukchi Sea, Arctic Ocean. Polar Sci.17, 33–39. 10.1016/j.polar.2018.06.003 (2018). [Google Scholar]
  • 9.Bourgain, P. & Gascard, J. C. The Atlantic and Summer Pacific waters variability in the Arctic Ocean from 1997 to 2008. Geophys. Res. Lett.39, L05603. 10.1029/2012GL051045 (2012). [Google Scholar]
  • 10.Timmermans, M. L. et al. Mechanisms of Pacific summer water variability in the Arctic’s central Canada basin. J. Geophys. Res. Ocean.119, 7523–7548. 10.1002/2014JC010273 (2014). [Google Scholar]
  • 11.Nikolopoulos, A. et al. The western arctic boundary current at 152°W: structure, variability, and transport. Deep Sea Res. Part. II. 56, 1164–1181. 10.1016/j.dsr2.2008.10.014 (2009). [Google Scholar]
  • 12.Brugler, E. T. et al. Seasonal to interannual variability of the pacific water boundary current in the Beaufort Sea. Prog. Oceanogr.127, 1–20. 10.1016/j.pocean.2014.05.002 (2014). [Google Scholar]
  • 13.Watanabe, E., Onodera, J., Itoh, M., Nishino, S. & Kikuchi, T. Winter transport of subsurface warm water toward the Arctic Chukchi Borderland. Deep Sea Res. Part. I. 128, 115–130. 10.1016/j.dsr.2017.08.009 (2017). [Google Scholar]
  • 14.Watanabe, E., Onodera, J., Itoh, M. & Mizobata, K. Transport processes of seafloor sediment from the Chukchi shelf to the western Arctic basin. J. Geophys. Res. Ocean.127, e2021JC017958. 10.1029/2021JC017958 (2022). [Google Scholar]
  • 15.Corlett, W. B. & Pickart, R. S. The Chukchi Slope Current. Prog Oceanogr.153, 50–65. 10.1016/j.pocean.2017.04.005 (2017). [Google Scholar]
  • 16.Spall, M. A. et al. Transport of Pacific water into the Canada Basin and the formation of the Chukchi Slope Current. J. Geophys. Res. Ocean.123, 7453–7471. 10.1029/2018JC013825 (2018). [Google Scholar]
  • 17.Li, M. et al. Circulation of the Chukchi Sea shelfbreak and slope from moored timeseries. Prog Oceanogr.172, 14–33. 10.1016/j.pocean.2019.01.002 (2019). [Google Scholar]
  • 18.Stabeno, P., Kachel, N., Ladd, C. & Woodgate, R. Flow patterns in the eastern Chukchi Sea: 2010–2015. J. Geophys. Res. Ocean.123, 1177–1195. 10.1002/2017JC013135 (2018). [Google Scholar]
  • 19.Boury, S. et al. Whither the Chukchi Slope Current? J. Phys. Oceanogr.50, 1717–1731. 10.1175/JPO-D-19-0273.1 (2020). [Google Scholar]
  • 20.Muramatsu, M., Ueno, H., Watanabe, E., Itoh, M. & Onodera, J. Transport and heat loss of the Pacific Summer Water in the Arctic Chukchi Sea northern slope: mooring data analysis. Polar Sci.29, 100698. 10.1016/j.polar.2021.100698 (2021). [Google Scholar]
  • 21.Lin, P., Pickart, R. S., Våge, K. & Li, J. Fate of warm Pacific water in the Arctic basin. Geophys. Res. Lett.48, e2021GL094693. 10.1029/2021GL094693 (2021). [Google Scholar]
  • 22.Leng, H., Spall, M. A., Pickart, R. S., Lin, P. & Bai, X. Origin and fate of the Chukchi Slope Current using a numerical model and in-situ data. J. Geophys. Res. Ocean.126, e2021JC017291. 10.1029/2021JC017291 (2021). [Google Scholar]
  • 23.Itoh, M. et al. Interannual variability of Pacific Winter Water inflow through Barrow Canyon from 2000 to 2006. J. Oceanogr.68, 575–592. 10.1007/s10872-012-0120-1 (2012). [Google Scholar]
  • 24.Mizobata, K. Reconstruction of Bering Strait volume transport suggesting the contribution of Bering Sea continental shelf to the pressure head forcing. Polar Sci.27, 100560. 10.1016/j.polar.2020.100560 (2021). [Google Scholar]
  • 25.Saha, S. et al. The NCEP climate forecast system reanalysis. Bull. Am. Meteorol. Soc.91, 1015–1057. 10.1175/2010BAMS3001.1 (2010). [Google Scholar]
  • 26.Saha, S. et al. The NCEP Climate Forecast System Version 2. J. Clim.27, 2185–2208. 10.1175/JCLI-D-12-00823.1 (2014). [Google Scholar]
  • 27.Lin, P. et al. Recent state transition of the Arctic Ocean’s Beaufort Gyre. Nat. Geosci.16, 485–491. 10.1038/s41561-023-01184-5 (2023). [Google Scholar]
  • 28.Proshutinsky, A. et al. Analysis of the Beaufort Gyre freshwater content in 2003–2018. J. Geophys. Res. Ocean. 124, 9658–9689. 10.1029/2019JC015281 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Jackson, J. M., Carmack, E. C., McLaughlin, F. A., Allen, S. E. & Ingram, R. G. Identification, characterization, and change of the near-surface temperature maximum in the Canada Basin, 1993–2008. J. Geophys. Res. Ocean.115, C05021. 10.1029/2009JC005265 (2010). [Google Scholar]
  • 30.Armitage, T. W. K., Bacon, S. & Kwok, R. Arctic sea level and surface circulation response to the Arctic Oscillation. Geophys. Res. Lett.45, 6576–6584. 10.1029/2018GL078386 (2018). [Google Scholar]
  • 31.Doglioni, F. et al. Sea surface height anomaly and geostrophic current velocity from altimetry measurements over the Arctic Ocean (2011–2020). Earth Syst. Sci. Data. 15, 225–263. 10.5194/essd-15-225-2023 (2023). [Google Scholar]
  • 32.Onodera, J. et al. Interannual variation of settling particles reflects upper-ocean circulation in the southern Chukchi Borderland, 2010–2014. J. Geophys. Res. Ocean.126, e2021JC017431. 10.1029/2021JC017431 (2021). [Google Scholar]
  • 33.Sumata, H., de Steur, L., Divine, D. V., Granskog, M. A. & Gerland, S. Regime shift in Arctic Ocean sea ice thickness. Nature615, 443–449. 10.1038/s41586-022-05686-x (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Dörr, J., Årthun, M., Eldevik, T. & Sandø, A. B. Expanding influence of Atlantic and Pacific Ocean heat transport on winter sea-ice variability in a warming Arctic. J. Geophys. Res. Ocean.129, e2023JC019900. 10.1029/2023JC019900 (2024). [Google Scholar]
  • 35.Oldenburg, D. et al. The respective roles of ocean heat transport and surface heat fluxes in driving Arctic Ocean warming and sea ice decline. J. Clim.37, 1431–1448. 10.1175/JCLI-D-23-0399.1 (2024). [Google Scholar]
  • 36.Kodaira, T., Waseda, T., Nose, T. & Inoue, J. Record high Pacific Arctic seawater temperatures and delayed sea ice advance in response to episodic atmospheric blocking. Sci. Rep.10, 20830. 10.1038/s41598-020-77488-y (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Alabia, I., García-Molinos, J., Hirata, T., Mueter, F. J. & David, C. L. Pan-arctic marine biodiversity and species co-occurrence patterns under recent climate. Sci. Rep.13, 4076. 10.1038/s41598-023-30943-y (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Nishino, S., Itoh, M., Kawaguchi, Y., Kikuchi, T. & Aoyama, M. Impact of an unusually large warm-core eddy on distributions of nutrients and phytoplankton in the southwestern Canada Basin during late summer/early fall 2010. Geophys. Res. Lett.3810.1029/2011gl047885 (2011).
  • 39.Kawaguchi, Y., Itoh, M. & Nishino, S. A detailed survey of a large baroclinic eddy with extremely high temperature in the Western Canada Basin. Deep Sea Res. Part. I. 66, 90–102. 10.1016/j.dsr.2012.04.006 (2012). [Google Scholar]
  • 40.Schauer, U. & Beszczynska-Möller, A. Problems with estimation and interpretation of oceanic heat transport – conceptual remarks for the case of Fram Strait in the Arctic Ocean. Ocean. Sci.5, 487–494. 10.5194/os-5-487-2009 (2009). [Google Scholar]
  • 41.Timmermans, M. L., Toole, J. & Krishfield, R. Warming of the interior Arctic Ocean linked to sea ice losses at the basin margins. Sci. Adv.4 (8), eaat6773. 10.1126/sciadv.aat6773 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Mizobata, K., Watanabe, E. & Kimura, N. Wintertime variability of the Beaufort gyre in the Arctic Ocean derived from CryoSat-2/SIRAL observations. J. Geophys. Res. Ocean.121, 1685–1699. 10.1002/2015JC011218 (2016). [Google Scholar]
  • 43.Regan, H. C., Lique, C. & Armitage, T. W. K. The Beaufort Gyre extent, shape, and location between 2003 and 2014 from satellite observations. J. Geophys. Res. Ocean. 124, 844–862. 10.1029/2018JC014379 (2019). [Google Scholar]
  • 44.Meneghello, G., Marshall, J. C., Campin, J. M., Doddridge, E. & Timmermans, M. L. The ice-ocean governor: ice-ocean stress feedback limits Beaufort Gyre spin-up. Geophys. Res. Lett.45, 11293–11299. 10.1029/2018GL080171 (2018). [Google Scholar]
  • 45.Proshutinsky, A. Y. & Johnson, M. A. Two circulation regimes of the wind-driven Arctic Ocean. J. Geophys. Res.102, 12493–12514. 10.1029/97JC00738 (1997). [Google Scholar]
  • 46.Armitage, T. W. K. et al. Arctic Ocean surface geostrophic circulation 2003–2014. Cryosphere11, 1767–1780. 10.5194/tc-11-1767-2017 (2017). [Google Scholar]
  • 47.Hirano, D. et al. Winter water formation in coastal polynyas of the eastern Chukchi Shelf: Pacific and Atlantic influences. J. Geophys. Res. Ocean. 123, 5688–5705. 10.1029/2017JC013307 (2018). [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material 1 (4.7MB, docx)

Data Availability Statement

The dataset of the subsurface OHC and the related variables in the southern Chukchi Borderland region is available at the Arctic Data archive System (ADS) (https://ads.nipr.ac.jp/data/meta/A20241203-001). The DOT and geostrophic ocean velocity datasets for 2011–2020 were also registered to the ADS (https://ads.nipr.ac.jp/data/meta/A20210401-001). The CFSR (https://rda.ucar.edu/datasets/ds093.0/) and CFSv2 (https://rda.ucar.edu/datasets/ds094.0/) datasets are provided by the NCAR Research Data Archive.


Articles from Scientific Reports are provided here courtesy of Nature Publishing Group

RESOURCES