The Drygalski Ice Tongue is presently the largest ice tongue in Antarctica and it forms the southern boundary of Terra Nova Bay (TNB), Western Ross Sea, Antarctica. It is the floating extension of the David Glacier and plays a crucial role in maintaining the Terra Nova Bay polynya by limiting northward-flowing of sea ice along the Victoria Land Coast . The free floating part of the ice tongue extends ∼ 80–90 km into the Ross Sea, with a width of ∼ 20 km as shown in Fig. , which makes it presently Antarctica's largest free-floating glacier . At the grounding line the glacier is more than 1900 m thick . The glacier then thins to the extent that the tip, located above the edge of the Crary Bank in a water depth of 600–700, is ∼200 m thick . Despite the blocking effect of the ice tongue on sea ice and surface water masses, determined that there is exchange and advection of water masses in the ocean below the floating ice.

Terra Nova Bay is bounded to the south by the Drygalski Ice Tongue and to the west by the Nansen Ice Shelf, which is fed by the Priestly and Reeves Glacier. A key bathymetric feature, the Drygalski Basin, extends north east into the Drygalski Trough and is believed to be the primary export pathway for High Salinity Shelf Water (HSSW). Along the eastern boundary, the Victoria Land Coast Current carries water and sea ice northward from McMurdo Sound . This flow has been inferred from satellite sea ice drift and surface drifters and is reproduced in ocean circulation models such as GLORYS and P-SKRIPS , though its vertical structure and depth extent remain poorly constrained by direct observations between McMurdo Sound and Terra Nova Bay. This current could also bring water masses into Terra Nova Bay below the Drygalski Ice Tongue or along the Drygalski Ice Tongue as replenishing the HSSW outflow.

The polynya that occasionally forms in Terra Nova Bay is relatively small in extent when compared to the nearby Ross Sea polynya but plays an important role in the global climate system. Between autumn and spring, strong katabatic winds blow off the Nansen Ice Shelf, pushing the sea ice off the coast and enabling heat loss from the relatively warm ocean to the cold atmosphere, and producing sea ice and HSSW. Despite its small size in global terms, Terra Nova Bay accounts for approximately 3 %–4 % of Antarctic sea ice production , and up to 10 % of Antarctic Bottom Water (AABW). Sea ice makes ship-based observations very difficult from late autumn when these important water masses are formed and HSSW is an important water mass that contributes to basal melting of floating glaciers and ice shelves . Hydrographic moorings, that observe water masses throughout the year are therefore an essential tool to study the oceanographic processes in Terra Nova Bay polynya. While moorings in the central Terra Nova Bay have provided a long term perspective of water mass transformation , the strong influence of the Drygalski Ice Tongue on the region provided motivation for a mooring array (hereinafter called the “DITx” array) focused around the Drygalski Ice Tongue.

One of the key processes influencing the region are the occasional katabatic-driven polynya events. In such conditions, the fast, cold katabatic winds drive sea ice formation at the ocean surface, which results in the creation of HSSW as a by-product e.g.,. The brine rejection from the sea ice formation first breaks down the stratification of the water column in the polynya before HSSW starts to form. While the production of HSSW begins in late Autumn/early winter, the salinity increase does not increase until September at the bottom in the eastern side of Terra Nova Bay, where DITN is located, and in the Drygalski Basin . Figure shows the seasonal evolution of stratification in Terra Nova Bay, based on a summer CTD profile and Argo float observations from both summer and winter . The circulation pattern in Terra Nova Bay is described by , based on summer CTD observations and three hydrographic moorings between December 2014 and March 2018 and include data from the hydrographic moorings described in this paper (DITN and DITD). They observed a cyclonic circulation in the deeper part of Terra Nova Bay between 400–700 m during the summer, which advects the HSSW created close to the Nansen Ice shelf, which is considered to be the primary location of HSSW production in Terra Nova Bay, towards the DITN, where the salinity increase is observed by the bottom instrument of this mooring. The salinity increases at this depth are observed between September and October, and not at the start of winter when the stratification has broken down. suggested that HSSW is advected from the Nansen Ice Shelf front towards eastern Terra Nova Bay and was confirmed in .

Terra Nova Bay Ice Shelf Water (TISW) is formed when HSSW interacts with basal melt waters from the Drygalski Ice Tongue and the Nansen Ice Shelf during active HSSW production. TISW is characterized by a conserved temperature below the surface freezing temperature around -1.94 °C depending on the salinity and a density larger than 1028 kg m⁻³ (e.g., ; , their Fig. 6). The characteristics of TISW observed in summer in depends on observation period and the outflow from under the Nansen Ice Shelf and is believed to advect in the same cyclonic circulation as HSSW towards the eastern Terra Nova Bay. TISW is also observed in the eastern Terra Nova Bay close to Drygalski Ice Tongue by DITN; further study needs to investigate if this TISW is formed under the Nansen Ice Shelf and advected or is formed when HSSW interacts with basal melt waters from the Drygalski Ice Tongue itself. Antarctic Surface Water (AASW) can be found at the subsurface and originates from the melting of the sea ice. It is a fresh and relatively warm (>-1.7 °C, ) water mass as it warms through solar radiation during Summer . As its density is much lower than the water masses formed during winter, it contributes to the restratification of the water column during the summer months. The stratification driven by the AASW and the breakdown of the stratification driven by the HSSW formation is shown in Fig. . The CTD profiles are also used for instrument calibration, as described in Sect. “Quality control” and Fig. .

The DITx hydrographic moorings distributed around the Drygalski Ice Tongue have been maintained since December 2014 in order to monitor the water mass properties and currents close to the ice tongue. Three moorings were deployed over different time frames, located around the Drygalski Ice Tongue as shown in Fig. . DITN is a long mooring deployed on the northern side of the Drygalski Ice Tongue from December 2014 to the present. It includes instruments near the seafloor, mid-water column, and subsurface, with the shallowest instrument at 75 m depth. The upper instruments are at risk of collisions from icebergs calving off the Drygalski Ice Tongue, which are frequent in the area and can extend to great depths. Despite this risk, the setup is unique and essential for studying subsurface processes. Its proximity to the ice tongue and near-surface coverage make it particularly well suited to investigate subsurface circulation and the influence of meltwater. DITD, a short mooring deployed in February 2017 and still active, is located in the deepest part of the Drygalski Basin. It captures the densest HSSW observed in the area, offering a valuable reference point for comparing water mass properties with other locations. DITS was a long mooring positioned south of the Drygalski Ice Tongue in Geikie Inlet and was maintained from February 2017 until January 2020. This paper presents a comprehensive overview of the decadal dataset collected by the DITx array. The data are freely available via SEANOE in netCDF format, organized by mooring, instrument type, and deployment year . Tables , , and summarize the mooring configurations, and instrument placements are visualized in Figs.  and .

In addition to these moorings, other relevant datasets in the region include the MORSea Moorings D and L and a Lamont-Doherty Earth Observatory (LDEO) mooring located near the front of the Nansen Ice Shelf . Since 2013, Argo floats have also contributed valuable data to the Ross Sea, with four deployed in Terra Nova Bay during 2021, 2022, 2024, and 2025 .

A number of checks are carried out to provide some assurance on data quality. The data collected by the instruments of the moorings are combined into netcdf files for each of the moorings per instrument type and deployment year. An overview of the available data is summarized in Tables , and . The data is also checked on faulty and missing data, which are described in the corresponding section in Sect. 3. In some cases, the instruments started the measurement timeseries whilst still on the ship, this resulted in +20° temperatures, and pressure measurements at sea level. This data is removed in before the data processing.

The SBE37SM(P) instruments are calibrated before deployment by the CSIRO Oceanographic Calibration Facility to a factory calibration. The salinity is calculated through the conductivity and temperature. The accuracy of the salinity measurements have been examined by . The initial accuracy is found to be 0.003 PSU, or in TNB 0.0030 g kg⁻¹. Across different CTD instruments from Seabird, during long-term deployments (>1 year) accuracy drifts to 0.01 PSU (0.01008 g kg⁻¹ in TNB) in regions of strong temperature gradients . These uncertainties are below the range of seasonal and interannual variability and the instrument errors are solely in the range of the winter conditions and these instruments are very stable . The variability during winter is still valuable information as the instrument errors are a shift, rather than noise. In addition to the factory calibration, a full-depth CTD cast was conducted right after each mooring deployment and again prior to recovery to use for the instrument drift calibration. The CTD profiles were obtained within a few hundred meters of the mooring location, as close to the mooring site as the environmental condition allowed, with a safety buffer to avoid entanglement with the mooring lines. The salinity measurements of the moored SBE37SM(P) instruments are post-calibrated with the profile measurements using the depth of the moored instruments. 1ΔSt1=SCTDdep-Smooringt12ΔStf=SCTDrec-Smooringtf3S(t)calibrated=S(t)mooring+ΔSt1+ΔStf-ΔSt1N⋅t, where SCTDdep is the salinity measured by the CTD at depth of the instrument close to the time of deployment, SCTDrec is the salinity measured by the CTD at depth of the instrument close to the time of recovery of the mooring, Smooringt1 is the salinity measured by the mooring just after it is deployed, Smooringtf is the salinity measured by the mooring just before it is recovered and N is the number of observations made by the mooring. The mooring timeseries is only calibrated if ΔSt1 or ΔStf is larger than 0.003 PSU. This value is chosen as it is the lower bound of the uncertainties within SBE37SM(P) CTD instruments. In case that the instrument stopped recording before recovery, the calibration is done by S(t)calibrated=S(t)mooring+ΔSt1. The temperature sensors were not adjusted post-deployment. Sea-Bird temperature sensors are highly stable and exhibit substantially lower drift than salinity, which is derived from conductivity and temperature. Applying a post-deployment temperature calibration to the SBE37SM(P) and SBE56 instruments would therefore likely introduce additional noise rather than improve data quality .

Pressure is recorded by the Aquadopp and the MicroCAT, SBEs; the measurements per depth are plotted in Fig. (DITN: panels a–c , DITS: panels d–f, DITD: panel g) . The depth at which the variables are measured, as described in Tables , and , are the mean values of the pressure-derived depth values. This mean is calculated by the median of the pressure derived depth values, to account for the mooring deflection due to strong wind-induced currents. The mean depth for instruments that do not record the pressure is calculated based on linear interpolation and spacing between two known depths – i.e. between the bed anchor, and an SBE37SM(P). The instrument depths vary between deployments, so combining the datasets requires caution to account for differences in measurement levels.

The pressure derived depth also varies between deployments, as the depth varies, in the order of a few meters, due to the exact location of the mooring anchor, therefore the depth of the mooring anchor changes slightly each time the mooring is retrieved and deployed again. The pressure data shows bigger spikes in the instruments closer to the surface compared to the instruments closer to the bottom. These spikes are not removed from the data as they are caused by a increase in depth when the mooring line gets pulled down due to drag imposed by currents. This has no significant impact on the mean and median of the pressure derived depth, as they are short lived (order of hours) events.

All SBE37SM(P), SBE56 and Aquadopp instruments measure the temperature. The depths of these instruments per mooring can be found in Tables , and . Figure shows the temperatures for each of the instruments (DITN: panels a–c, DITS: panels d–f, DITD: panel g) . The accuracy of the Sea-Bird instrument are significantly better (0.002 °C) than the Aquadopp (0.01 °C) and the RCM9 (0.05 °C) as shown in Table . Neither the SBEs, Aquadopp or RCM9 temperature sensors are calibrated as described in Sect. “Quality control”.

The temperatures measured in the top 400 m, show a clear seasonal signal, decreasing with depth, as shown in Fig. a, b, d and e. The temperature measured by the DITN near the surface at ∼75 m depth (Fig. a increases from November, peaking at the end of February, it decreases during the Autumn months from March until June and during the winter months, the temperatures are ∼-1.9 °C and show the smallest variability at this mooring. In the Geiki inlet where the DITS is located, the temperature in the top 100 m (Fig. d) increases from the end of December, peak in February, and decreases until April where temperatures show low variability throughout the winter and spring. At 275 m depth at the DITN (Fig. b) and at 208 m at the DITS site (Fig. e), the temperature does not increase until late February. During April the highest temperatures of ∼-1.2 to -1.7 °C are recorded and they start to drop again during May and June. The winter period starts in July at the DITN site and in May at the DITS site with temperatures ∼-1.9 °C with super-cooled water reaching <-2.0 °C. Near the bottom (Fig. c, g) temperatures range between -1.87 and -1.91 °C, there is a decrease between July and November at the DITD and DITN locations and no clear seasonal signal at the DITS site (Fig. f). There is a strong temperature increase between end of April and start of May measured by the SBE37SM DITS1803_15239 at 63 m depth (Fig. d) and the SBE56 thermistor DITS1803_4854 at 208 m depth (Fig. e). Because this signal is measured two different instruments at two different depths, it is not removed from the final dataset, but should be treated with caution. Sudden large temperature changes (spikes) that last longer than 1 h are not removed from the dataset as the polynya is a highly active environment and sudden temperature changes can occur.

The temperatures are measured correctly for most of the instruments and moorings. The measurements that need to be treated with caution, are absent, or ignored due to faulty instruments are presented in Tables , and . Aquadopp c1901DITN_9929 at 71 m depth (Fig. a) shows a sudden jump in temperature on 22 August 2019. At the same time as the sudden temperature increase, the pressure increases to 90 dbar at 15:30 NZST and decreases again to 68 dbar at 16:15 (Fig. a), while the temperature does not go back to the values before the sudden increase. This could indicate that the instrument got hit by an iceberg and possibly broke the temperature sensor. The temperature data after 22/8/2019T15:30 is removed in the final dataset.

The salinity is measured in PSU indirectly by the SMBE37SM using conductivity and temperature, and converted to absolute salinity [g kg⁻¹] in this paper. At DITN near the surface at ∼75 m depth (Fig. a), the salinity decreases in the summer from the middle of January from 34.9 to 34.2 PSU. At ∼275 m depth (Fig. b), the salinity oscillates with the highest salinity in September and the lowest salinity in May. Close to the bottom, there is no obvious seasonal signal. In the Drygalski Basin, measured by the DITD (Fig. g), the salinity decreases slowly from the end of Autumn, with a sharp increase between September and October. The SBE37SM instruments on the DITS1803 mooring, measuring the salinity either did not record any data or stopped early. The salinity measurements are plotted in Fig. .

The in-situ density is measured for DITN at 75, 275 and 668 m depth, on DITD at 1220 m, on DITD1803 at 1120 m, on DITS1702 at 190, 390 and 1090 m, and on DITS1803 at 48, 240 and 1050 m. The MicroCAT did not record any density data from the DITN1902. Also there is no density data recorded on the DITN1512 mooring at 275 m (SBE37SM724) and 668 m (SBE37SM5838) depth. The missing density data can be calculated using the TEOS10 function with the salinity, temperature and pressure. The measured in-situ density on the DITN1902 mooring at 668 m depth, is faulty, as it is calculated with the salinity and temperature, which was not recorded correctly. The same applies to the DITS1803 mooring data at 48 and 240 m depth.

The speed and direction of the current is measured with the RCM9 and Aquadopp instruments. On the DITN at 74, 76, 265, 274, 670 & 672 m depth. The RCM9 was installed on the DITN1412 & DITN1512 mooring at 74, 274 & 272 m depth and DITN1702 mooring at 74, 274 & 342 m depth and the Aquadopp at 76, 265 & 670 m depth from the DITN1702 mooring onwards. The DITD measures the currents with the Aquadopp at 1240 m depth. The currents are measured on the DITS1702 at 190, 390 and 1090 m depth with the RCM9, and on the DITS1803 50, 230 and 1060 m depth with the Aquadopp.

Figure shows the hourly means of the current speed for each of the moorings. The current direction is corrected for the magnetic declination at the position of the mooring.

The prevailing direction of the flow shows in Fig. measured by DITN is westerly near the surface at ∼75 m depth (Fig. a) and in the middle of the water column at ∼275 m depth (Fig. b). At the bottom at a depth of 670 m (Fig. c), the current direction is very variable with a slight south westerly tendency. The flow at DITD (Fig. d) has a westward tendency, ranging between north west and south west or in north east direction. The bottom instrument of DITN and DITD show the largest variability, while the subsurface shows the smallest variability. This preferred westerly flow near the surface corresponds with the deflection of the Victoria Land Coastal current deflection observed in satellite data .

The heading, roll and pitch are measured with the Aquadopp on the DITN1702 onwards at 76, 265 & 670 m depth, at 1145 & 1240 m on the DITD moorings and at 50, 230 and 1070 m on DITS1803. The heading is the rotation along the z-axis, the pitch along the x-axis and the roll along the y-axis. The heading, roll and pitch data can be used to study the movement of the mooring. The set-up, lengths between the instrument connections and length of the chains of the mooring can explain the variability in between mooring deployments.

A spectral analysis is performed to assess the dominant timescales of the observed velocities and pressure. The power spectral density (Fig. ) shows two clear peaks at 1 and 2 cycles per day in the pressure data (Fig. e, f) in the bottom two instruments, corresponding to the diurnal and semi-diurnal tidal constituents, with the diurnal tide the dominant tidal frequency in the region. Near the surface (Fig. d) this signal is not as profound, suggesting the pressure is affected by other processes like surface gravity waves induced by wind that operate around the frequency of the tides. The spectral analysis of the speed data only shows a peak at one cycle per day (Fig. a) that corresponds with the tidal motion; at greater depth (Fig. b, c) these peaks are absent. The tidal constituents are not clearly visible in the speed data, likely because eddies are shed from the ice tongue and advected past the instruments, while the background currents driven by tidal motion move around the ice tongue. At greater depth, the diurnal and semi-diurnal tidal constituents are clearly evident in the pressure spectral analysis, reflecting the dominance of barotropic tidal forcing where surface effects and transient processes such as wind or ice interactions are minimal. A distinct secondary peak is observed near 0.2 cycles per day (Fig. a), indicating variability on a 3–5 d timescales. These lower scale fluctuations represent the inertial and sub-inertial variability like eddies and synoptic forcing like storms and katabatic wind events.

The described hydrographic mooring time series, presented in this paper, is a valuable tool to increase our understanding of the highly dynamic polynya system. Comparing this data set with other observations in Terra Nova Bay, helps with the horizontal extent of the polynya dynamics and variability. The length of the observations closer to the surface, provides insight into the drivers of and interannual variability of water mass formation processes. This long-term dataset constrains variability associated with atmospheric conditions, sea ice dynamics, and ocean circulation patterns, that influence the timing, amount, and properties of HSSW rather than allowing direct attribution and supports future investigation of how the Drygalski Ice Tongue itself may influence these processes. We aim to determine the origin of water masses observed near the Drygalski Ice Tongue during the winter months when HSSW is formed while bay-wide CTD observations are impossible. In future work we also aim to useidealised and regional models to complement the observational dataset.

The datasets from DITN and DITS are also valuable for biological applications. Both DITN and DITS include instruments located within the euphotic zone, allowing for the analysis of the mixed layer depth and its variability. Due to close proximity of DITN and DITS to the Drygalski Ice Tongue, these datasets offer an opportunity to investigate oceanographic conditions in the vicinity of the ice tongue that are relevant for understanding ice-ocean interaction, rather than assessing ice-tongue stability or evolution. These moorings can give an in-situ oceanographic perspective that complements changes observed with radar and satellite data. When combined with radar and satellite imagery, the in-situ data provide a complementary oceanographic perspective, enhancing our understanding of ice-ocean interactions and supporting interdisciplinary studies linking ocean physics, glaciology, and marine biology.