Local and remote processes' impacts on dense water formation (Deliverable D1.10)

Ocean Cryosphere Exchanges in Antarctica: Impacts on Climate and the Earth System Local and remote processes’ impacts on dense water formation Deliverable D1.10 UK Partners are funded by UK Research and Innovation (UKRI) under the UK government’s Horizon Europe funding Guarantee. https://ocean-ice.eu/ OCEAN ICE – GA 101060452 Deliverable D1.10 About this document Deliverable: D1.10 Local and remote processes’ impacts on dense water formation Work Package: WP1 Subpolar circulation, heat delivery, and water mass export Delivery date: 24 October 2025 Type of document: Report Dissemination level: Public Lead beneficiary and author: PP3: Alfred-Wegener-Institut Helmholtz-Zentrum Für PolarUnd Meeresforschung (AWI): Markus Janout Contributors: PP3: Alfred-Wegener-Institut Helmholtz-Zentrum Fur PolarUnd Meeresforschung (AWI): Mathias Campos van Caspel, Ralph Timmermann, Markus Janout Review: PP1: Danish Meteorological Institute (DMI), Chiara Bearzotti ([email protected]) Cover sheet: Circumpolar map showing the main dense shelf water (DSW) formation regions: Weddell Sea (WS), Ross Sea (RS), Prydz Bay (PB), and Adélie Land (AL). The red lines indicate the regional limits used to calculate the DSW volume. The background colors indicate the DSW thickness (right). Disclaimer: Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or European Research Executive Agency (REA). Neither the European Union nor the granting authority can be held responsible for them. 2 OCEAN ICE – GA 101060452 Deliverable D1.10 Table of contents 1. Publishable summary.....................................................................................................................................4 2. Work performed and main achievements..................................................................................................... 5 2.1 Description of the work performed.......................................................................................................5 2.1.1 Model description........................................................................................................................5 2.1.2 Dense Shelf Water definition....................................................................................................... 5 2.1.3 Results.......................................................................................................................... 6 2.2 References...........................................................................................................................................13 2.3 Open Science.......................................................................................................................................14 3. Results..........................................................................................................................................................14 4. Impact.......................................................................................................................................................... 14 3 OCEAN ICE – GA 101060452 Deliverable D1.10 1. Publishable summary The dense shelf water (DSW) formed around Antarctica is a source of the coldest waters spreading through the ocean's abyss, the Antarctic Bottom Water (AABW) (e.g., Orsi et al., 1999). The formation is associated with recurring polynyas in four main areas of the continental shelf, including the southern Weddell Sea, the Ross Sea, Prydz Bay, and Adélie Land (Solodoch et al., 2022). Polynyas open up when sea ice is pushed away from the coast by winds or ocean currents. In cold months, the exposed ocean surface freezes again, and salt is rejected into the ocean as brine. The large brine injection related to intense sea ice production (SIP) increases the ocean salinity, forming High Salinity Shelf Water (HSSW), one type of DSW with surface freezing temperatures. In the Southern Ocean, most polynyas occur over the continental shelf where HSSW sinks and follows the bathymetry along the bottom. HSSW can flow offshore and slide down the continental slope, mixing with ambient waters along its path until it reaches the equilibrium depth as AABW. HSSW can also flow into the ice shelf cavity, where the freezing temperature is lower than at the surface due to the pressure effect. There, it causes melting, and the mixture with Glacial Melt Water (GMW) generates the supercooled (colder than surface freezing temperature) and slightly fresher Ice Shelf Water (ISW), a second flavor of DSW. We used previously validated model results (van Caspel et al., 2024; D1.7) to show that DSW in the Weddell and Ross Seas persists year-round, whereas it is absent in summer at the other two formation sites. In the Weddell Sea and Ross Sea, the maximum volume of DSW available during the formation season is correlated with the minimum volume of the following year. This finding indicates that, in these areas, the annual formation rate plays an important role in the DSW reservoir. The annual increase in DSW volume correlates with the local accumulated SIP in polynyas in the Weddell Sea and Prydz Bay. In most cases, the salinity of the surface and/or subsurface waters entering each region from remote areas has a significant correlation with the salinity of the DSW formed, but only in Prydz Bay the correlation is higher than 0.5. 2. Work performed and main achievements 2.1 Description of the work performed 2.1.1 Model description “The Finite volumE Sea Ice-Ocean Model (FESOM2) is a multi-resolution ocean general circulation model that solves the equations of motion describing the ocean and sea ice using finite-volume methods on unstructured computational grids. The model is developed and operated by the Alfred Wegener Institute, Helmholtz Centre for Polar and Marine Research (AWI), in Bremerhaven, Germany.” (quoted from the FESOM2 documentation, https://fesom2.readthedocs.io/en/latest/index.html, Danilov et al. 2017, Danilov et al. 2015). The interaction with the ice shelf is based on the three-equation approach proposed by Hellmer and Olbers (1989). Heat exchange is applied to temperature changes, while freshwater input from ice shelf melt and water freezing are handled as salinity changes. Temperature and salinity are modified only in the layers in direct contact with the ice shelf base. "The mesh used for this study was generated using the FESOM Mesh Generator (fmesh - https://github.com/FESOM/fmesh) modified to include the ice shelf cavities. The global mesh contains 4 OCEAN ICE – GA 101060452 Deliverable D1.10 1,261,056 triangles (elements) formed by 648,389 vertices (nodes). About 36% of those are located south of 60 °S where the resolution (square root of element area) varies between 39.3 and 2.2 km with higher resolution in the ice shelf cavities and on the continental shelves and lowest resolution in areas with greater depth and few bathymetric features. The bathymetry and ice shelf draft are derived from RTopo-2 (Schaffer et al., 2016). The vertical discretization was made with 148 horizontal “z” levels. The layer thickness is smaller near the surface and is kept constant or slightly increased with depth. The layer below has a maximum thickness increase of 10 %, for example, if the layer on top is 20 m thick, the layer below would be between 20 and 22 m thick. The ocean initial temperature and salinity are from the “Global Ocean Hydrography with a High Quality Arctic Ocean Version 3.0” (Steele et al., 2001), a climatology derived from the 1998 version of the World Ocean Atlas (Antonov et al., 1998; Boyer et al., 1998). All the initial velocities are 0. The atmospheric forcing for the model is obtained from the Japanese 55-year Reanalysis, also known as JRA-55 (Kobayashi et al., 2015; Harada et al., 2016). The simulation starts in 1958 and goes until 2020 with 2-minute time steps; the outputs are provided as monthly means." (van Caspel et al. 2024, D 1.7) . Here we use the results from 1991 to 2020. 2.1.2 Dense Shelf Water definition Previous comparison with climatological data (van Caspel et al. 2024 - D1.7) shows that the model performs well over the continental shelf; the surface water, winter water, and the inflowing modified Circumpolar Deep Water (mCDW), as well as the processes that convert these water masses into DSW in the main formation areas, are accounted for in the model. The validation report also indicated that the model performed best onshore of the 1000 m isobath, the same area used here for circumpolar values, in addition ,the four main formation sites are assessed individually (Fig. 1). A temperature-salinity (T-S) diagram of the circumpolar climatological value (Fig. 2) shows that the model forms both types of DSW, i.e., the saline HSSW and the supercooled ISW. Most of the ISW is found below the ice shelves, but also on the continental shelves outside of the cavities. Based on the T-S diagram, the model DSW was defined as waters with potential temperature lower than –1.7 °C and density higher than 1027.8 Kg/m3. According to this definition, the DSW occupies 30 % of the ocean volume onshore of the 1000 m isobath, taking into account the volume inside the ice shelf cavities; excluding the cavities, it decreases to 17 %. 2.1.3 Results The mean circumpolar volume of DSW is 190 x 103 Km3. 57.5% of this DSW is found on the continental shelf of the southern Weddell Sea (WS-DSW), 40.7% on the Ross Sea continental shelf (R-DSW), while the Prydz Bay and Adelie Land areas account for only 1.0% and 0.6%, respectively (Table 1). The average potential temperature of the WS-DSW is -1.91 °C, which is lower than the surface freezing point temperature, -1.90 °C, derived from the average salinity of 34.65, indicating the influence of ISW. The R-DSW has the highest average salinity, 34.67, and, in agreement with the literature, can produce the saltiest AABW (e.g. Purkey et al, 2018). 5 OCEAN ICE – GA 101060452 Deliverable D1.10 Fig. 1: Circumpolar map showing the main DSW formation regions: Weddell Sea (WS), Ross Sea (RS), Prydz Bay (PB), and Adélie Land (AL). The red lines indicate the regional limits used to calculate the DSW volume. The background colors indicate the depth (left) and DSW thickness (right). The black thick line represents the 1000 m isobath, the magenta dashed line represents the 600 m isobath, the dashed black line represents the 500 m isobath, and the cyan line represents the ice shelf front. Fig. 2: T-S diagram for the region onshore of the 1000 m isobath with (left) and without (right) cavities. The black contours represent isopycnals, the dashed line indicates the surface freezing point, and the magenta lines denote the defined DSW limits. In the Weddell (Fig. 3) and Ross Seas (Fig. 4), the DSW is present on most of the continental shelf and below the ice shelves. The thickness of the DSW layer is larger in the deeper regions, where dense water can accumulate before it is being exported. The Ronne Polynya, located north of the Ronne Ice Shelf front near the Antarctic Peninsula, is evident in the climatological SIP of the southern Weddell Sea (see Fig. 3) and forms a pool of DSW that enters the Ronne Ice Shelf cavity. The inflowing DSW initiates a cavity circulation 6 OCEAN ICE – GA 101060452 Deliverable D1.10 through the cavity, during which part of the inflowing HSSW transforms into ISW, and then exits into Filchner Trough and finally migrates down the continental slope. Besides Filchner Trough as the main DSW pathway, smaller quantities of shelf-formed DSW (mainly HSSW) are present in the Central Trough. In situ observations show the presence of DSW on both sides of the Central Trough, namely in Ronne Trough and on Berkner Bank (Fig. 3 in Janout et al. 2021), indicating that the presence of DSW in the Central Trough is possible besides the lack of direct observations. Table 1: DSW climatological volume, salinity, and potential temperature in different regions Region Volume (103 Km3) Volume (%) Salinity Potential Temperature (°C ) Circumpolar 190 100 34.66 -1.90 Weddell Sea 109 57.5 34.65 -1.91 Ross Sea 77 40.7 34.67 -1.88 Prydz Bay 2 1.0 34.63 -1.86 Adelie Land 1 0.6 34.57 -1.81 In the Ross Sea, polynyas are present along the ice shelf front and in Terra Nova Bay, in both model and observations (Fig. 4 e.g., Silvano et al. 2023). The DSW layer is more prominent in the western sector of the continental shelf, with two branches extending northward towards the shelf break on the Drygalski Trough, and part of the DSW entering the cavity and leaving via the Glomar Challenger Trough, in the center of the ice shelf front (e.g., Silvano et al. 2023). The DSW layer is less pronounced in Prydz Bay (Fig. 5) and Adelie Land (Fig. 6) and, on average, does not flow under the Amery Ice Shelf or the Merz Glacier Tongue. In Prydz Bay, the Cape Darnley Polynya and Mackenzie Bay Polynya are present in the model, as well as a third polynya to the east of the Amery Ice shelf, all consistent with observations (e.g., Silvano et al. 2023). In the model, DSW accumulates only around the Cape Darnley Polynya. In Adélie Land, the Mertz Polynya is located west of the ice tongue in both the model and observations (e.g., Silvano 2023). This polynya produces DSW that drains along the Adelie Depression towards the shelf break. In all regions, the DSW volume changes in a seasonal cycle (Figs. 3-6). Between autumn and spring, the SIP drives DSW formation, and the volume increases. During summer, very little DSW forms, but the export either into the deep ocean or into ice shelf cavities continues, and the volume diminishes. In the Weddell and Ross Seas, the volume never drops to zero, and the largest volume during any given year correlates with the minimum volume in the following year (Table 2). The average annual cycle (monthly climatology) reaches its maximum volume in October in the larger reservoirs and in September in the two smaller formation areas; the minimum volume occurs in April in all regions. We use October and April as reference months to estimate annual DSW formation. For each year, we calculate the volume difference between October and April (Figs. 3 to 6), which provides an approximation of the annual DSW production. This production is primarily driven by local sea ice formation, and to assess their relationship, we computed the annual accumulated Sea Ice Production (SIP) both within polynyas and across the entire region. Polynyas are defined here as areas where SIP exceeds 5 m yr⁻¹. The annual increase in regional DSW volume correlates with the accumulated SIP in polynyas in the Weddell Sea (correlation coefficient r = 0.8, with significant correlation at 95% confidence, p < 0.05) and Prydz Bay (r =0.68) (Table 2). In both cases, the correlation is slightly lower if the SIP in the whole region is considered. The low 7 OCEAN ICE – GA 101060452 Deliverable D1.10 correlation in the Ross Sea and Adélie Land could be due to lower HSSW formation or to the rapid export of DSW from the continental shelf. Table 2: Correlation between the annual maximum DSW volume (Max Vol) and the minimum volume (Min Vol) in the following year; the DSW formed and the Polynya SIP; the DSW formed and the region SIP; and the DSW salinity in October (Sal) and the Polynya SIP. Significant correlations at 95% confidence, p < 0.05 are shown with bold letters. Region Max Vol X Min Vol DSW Formed X Polynya DSW Formed X SIP DSW Sal X Polynya Weddell Sea 0.80 0.68 0.60 0.74 Ross Sea 0.73 0.19 0.40 0.42 Prydz Bay 0.24 0.68 0.65 0.29 Adelie Land 0.57 0.10 0.12 0.27 Not only the volume, but also the properties of the DSW change over time. In the Weddell Sea the salinity of DSW in October correlates well with the polynya SIP (r = 0.74) (Table 2), where periods of low polynya SIP are associated with low DSW salinity (Fig. 3). In the Ross Sea this correlation is significant but smaller (r = 0.42) and the SIP alone is not enough to explain the distinct but temporary freshening after 2007. The modeled Ross Sea freshening period is shorter and steeper (Fig. 4) compared with observations (e.g., Jacobs et al., 2022). The 2007-2014 period is characterized by lower DSW salinity, likely associated with lower SIP, followed by a rebound thereafter related to increasing SIP, which again is synchronous with observations. Several studies associate these salinity changes with a larger input of glacial meltwater upstream (e.g., Castagno et al., 2019; Silvano et al., 2020; Jacobs et al., 2022), modifying the properties of the surface waters and the Circumpolar Deep Water (CDW) carried, for instance, from the Amundsen Sea towards the Ross Sea DSW formation sites. To assess the impact of this remote source on the DSW properties, we computed the average salinity of the surface and subsurface waters entering each region. The average salinity of waters carried westward in the Coastal Current (CC) and in the Antarctic Slope Current (ASC) is computed for the layer between the surface and 200 m depth, hereafter referred as surface waters, and from 200 m depth down to 600 m or the bottom, referred to as the subsurface layer. In the Ross Sea, the lower salinity of the subsurface waters carried by the CC and ASC in 2007 and 2008 (Fig. 4) may have played a role in triggering the DSW freshening, although based on our analyses, a freshening of inflowing waters needs to be synchronized with local SIP changes to impact the DSW formation and volume. The overall correlation between DSW salinity and the salinity of inflowing water is low in all regions except for Prydz Bay (Table 3), where the salinity in the CC and the surface waters carried by the ASC correlate well with the DSW salinity. The subsurface waters carried by the ASC, mainly CDW, in the Ross Sea, Weddell Sea, and Adelie Land have low but significant correlation with the DSW salinity formed. Previous assessment of the model results indicates that the model does not reproduce the basal melt increase in the Bellingshausen and Amundsen Seas (Fig. 6 in van Caspel et al. 2024 - D1.8), therefore, it is not a surprise that the long-term freshening of the Ross DSW is not accounted for. On the other hand, the basal melt interannual variability superimposed over the trend (e.g., Fig. 2 in Jacobs et al. 2022) is present in the model, and those fluctuations correlate with the Ross Sea DSW salinity (r = 0.49) better than the salinity of the inflowing waters or SIP. This suggests that, at least in the model, the influence of the upstream basal melt occurs by modulating the current strength rather than via direct changes in salinity. 8 OCEAN ICE – GA 101060452 Deliverable D1.10 Table 3: Correlation between the DSW salinity with the polynya SIP, and the salinity of surface (Surf) and subsurface waters (SubSurf) carried by the Coastal Current (CC) and by Antarctic Slope Current (ASC). Significant correlations at 95% confidence, p < 0.05 are shown with bold letters. Region CC Surf CC SubSurf ASC Surf ASC SubSurf Weddell Sea -0.02 0.17 0.07 0.21 Ross Sea -0.01 0.06 -0.09 0.30 Prydz Bay 0.61 0.53 0.61 0.04 Adelie Land -0.02 0.01 0.06 0.19 In summary, our correlation analysis indicates that, in the model, the DSW on the Weddell Sea is primarily influenced by local SIP rather than by anomalies advected from upstream regions. In the Ross Sea, the difference between local and remote impact in the DSW is smaller, with the DSW salinity correlation of r = 0.42 with the polynya SIP and r = 0.3 with the CDW salinity (ASC subsurface), and the largest correlation with the upstream basal melt, r = 0.49. In Prydz Bay, the SIP controls the volume of DSW formed, but the DSW salinity is regulated by waters carried towards the area by the CC and ASC. In contrast to the other formation sites, the DSW formed in Adelie Land does not have a direct correlation with SIP nor the properties of inflowing waters. Overall our model reproduces the circumpolar water mass distribution, as is largely observed by various observing systems around the continent, with 98 % of the DSW volume in the Weddell and Ross Seas but clear presence of DSW also on the other two main DSW formation sites. Besides the good geographical distribution, the model is also capable of producing, storing, and exporting DSW. Here we present results for the first two steps of the DSW cycle, while DSW export will be addressed in more detail in a manuscript that is currently being prepared. 9 OCEAN ICE – GA 101060452 Deliverable D1.10 processes. Our work was based on the 1979-2020 hindcast simulation using FESOM2, which was verified against observations that were assimilated within WP1. O5: Assess how global ocean circulation is impacted by freshwater discharge from the northern and southern ice sheets. In this report, we analyzed the major dense water formation processes and pathways and investigated the impact of ice shelf meltwater on the major dense water formation regions of the Weddell and Ross Seas. These processes are directly relevant to the global ocean circulation, as these shelf-formed dense waters are the precursors to Antarctic Bottom Water. 16