Methodology:

Data and samples were collected from the uncontaminated seawater systems on board the RRS Discovery (DY158, A-76A) and RRS Sir David Attenborough (SD033, A-23A).

Underway temperature and salinity measurements were taken with a Seabird Scientific SBE 45 thermosalinograph with a remote SBE 38 temperature sensor located near the water inlet on both vessels. Salinity was calibrated against in situ samples collected at 3-hourly intervals on DY158 and 4-hourly intervals on SD033, then analysed on a Guildline Autosal 8400B salinometer against IAPSO standard seawater batch P164 (DY158) and P165/P167 (SD033). Salinities were measured on the practical salinity scale. Fluorescence data was collected using a WS3S WETStar fluorometer and was converted to chlorophyll-a using factory calibrations and was not further calibrated against in situ data. After removing spikes in the data streams, underway temperature, salinity, and chlorophyll-a; data were averaged at 5-minute intervals.

Samples for analysis of oxygen isotopes were taken into 50 ml glass bottles, sealed with rubber stoppers, and aluminium crimp seals, and transported in the dark at 4 °C to the British Geological Survey, Keyworth. Samples were then analysed for their oxygen isotope composition using the CO2 equilibration method with an Isoprime 100 mass spectrometer plus Aquaprep device. Measurements were calibrated against international primary standard VSMOW2.

Samples for analysis of dissolved nitrate, nitrite, and phosphate were filtered through an in-line Cytiva AcroPak filter (0.8/0.45 µm) into acid cleaned (10 % reagent grade HCl) 60 ml HDPE bottles. Samples were frozen immediately at -20 °C and transported to Plymouth Marine Laboratory (from DY158) and the National Oceanography Centre (from SD033). At both locations, samples were thawed using international GO-SHIP protocols (Becker et al., 2020). Samples were analysed using a SEAL Analytical AAIII segmented flow colorimetric autoanalyzer following protocols by Woodward and Rees (2001). Sample handling and manipulation followed Becker et al. (2020). Calibration standards were prepared with low-nutrient seawater and the analytical results were quality controlled by reference to analysed certified reference materials (CRMs) suppled from Kanso Ltd. (Japan). Raw data were further corrected to ambient ocean salinity and pH.

Samples for dissolved silicon isotopes were filtered through an in-line Cytiva AcroPak filter (0.8/0.45 µm) into acid cleaned (10 % reagent grade HCl) 250 ml HDPE bottles. Samples were immediately stored in the dark at 4 °C and transported to the British Antarctic Survey. A 2.5 ml aliquot was removed for silicic acid (DSi) concentration analysis using the protocols described above for other macronutrients. DSi was then precipitated from seawater using the 'Magnesium Induced Co-Precipitation' (MAGIC) method (Brzezinski et al., 2003; Cardinal et al., 2005; Reynolds et al., 2006). Samples were purified with AG50W-X12 cation exchange resin (Georg et al., 2006). The AG50W-X12 resin was pre-cleaned and rinsed with acid (3M, 6M, and concentrated HCl) and Milli-Q (18.2 Momega.cm) before samples and standards were passed through the columns and eluted with Milli-Q. Silicon isotopic compositions were determined on a Thermo Scientific Neptune multicollector inductively coupled plasma mass spectrometer (MC-ICP-MS) at the University of Cambridge Department of Earth Sciences.

Silicon isotopic compositions were calculated as the per mille deviation from the primary reference standard NBS28 (NIST RM 8546) measured immediately before and after each sample. 0.1 ppm of an ICP-MS magnesium (Mg) standard was added to the samples to correct for instrumental mass bias by estimating the isotopic fractionation occurring within the machine during measurement through monitoring the 26Mg/25Mg ratio (Cardinal et al., 2003...(12)

Data collection:

Data collection and sample analysis:

- Seabird Scientific SBE 45 thermosalinograph with a remote SBE 38 temperature sensor

- Guildline Autosal 8400B salinometer

- WS3S WETStar fluorometer

- Isoprime 100 mass spectrometer with Aquaprep device

- SEAL Analytical AAIII segmented flow colorimetric autoanalyzer

- Thermo Scientific Neptune multicollector inductively coupled plasma mass spectrometer

Data processing:

- MATLAB R2025a

Data quality:

Quality control and standards:

Salinity: analysed against IAPSO standard seawater batch P164 (DY158) and batches P165/P167 (SD033).

Oxygen isotopes: calibrated against international primary standard VSMOW2.

Dissolved nutrients: quality controlled by reference to certified reference materials from Kanso Ltd. (Japan).

Dissolved silicon isotopes: calibrated against international primary standard NBS28 (NIST RM 8546). Quality controlled by reference to certified reference materials ALOHA 1000 m seawater (a deep Pacific Ocean seawater standard; 1.23 +/- 0.05 per mille (1sigma), n =12), diatomite (a siliceous sediment standard; 1.31 +/- 0.03 per mille, n = 5), and LMG (a siliceous sponge spicule standard; -3.31 +/- 0.04 per mille, n = 7)

Data cleaning:

Data were binned into 30-minute time intervals for each iceberg, taking the mean value for any parameters which had more than one data point during the 30-minute window. Around A-23A (SD033), the ship remained still for a period of time during the sampling transect, and as such the data from this time period has been binned together.

The dataset contains NaN values for instances where a sample for a given parameter was not taken during a given time period, which results from different sampling frequencies being undertaken for different parameters.