Methodology:

ChAOS_1_Pore_Water_DSi: dissolved silicic acid (DSi) concentrations of pore water extracted from sediment cores collected from 6 stations (B03, B13, B14, B15, B16, B17). Three separate Multicorer deployments were carried out at each station for all three cruise years, denoted by the event number. Samples were combined from 4 separate cores of each Multicorer deployment to ensure adequate sample volume was collected. Samples were stored in the dark at 4°C and analysed within 12 hours of collection. Core top water samples were extracted first from sediment cores sampled with the Multicorer, followed by the pore water samples. Pore waters were extracted with Rhizon filters, attached to 30mL plastic syringes using spacers to create a vacuum (resolution of 1 cm from 0.5-2.5 cm below seafloor (cmbsf), 2 cm to 20.5 cmbsf and 5 cm to 35.5 cmbsf), placed in pre-drilled coring tube holes and collected in acid cleaned bottles, which were then acidified with Romil UpA HCl. On-board measurements of DSi concentrations in the pore water samples were carried out with a Lachat Quickchem 8500 flow injection autoanalyzer using certified reference materials for sea water to define the accuracy (KANSO Ltd., Japan), which averaged 2.8% across the three cruises (1.5-5%).

ChAOS_2_Pore_Water_Isotopes: Si isotopic composition of the pore water DSi pool (delta30SiDSi-PW per mil). Samples from one Multicorer deployment from each station and cruise year were sampled for isotopic analysis. B13, B14 and B15 were selected for isotopic analysis as they span the three main hydrographic domains of the Barents Sea (Atlantic Water, Oceanic Polar Front, Arctic Water). Isotopic composition is given in delta notation as a deviation relative to NBS28 Si standard in units of per mil. This sheet also includes pore water NO3- and Fe concentrations, previously published in Freitas et al., (2020) and Faust et al., (2021) respectively. Isotopic analysis was carried out in a clean setting in the Bristol Isotope Group (BIG) laboratory at the University of Bristol. Samples were concentrated using the Mg-induced co-precipitation method (MAGIC) of Karl and Tien (1992) and De Souza et al., (2012). Concentrated samples were then passed through cation exchange columns (Georg et al., 2006) filled with a Bio-Rad AG50W-X12 resin to remove cations. Samples were then doped with H2SO4 (ROMIL-UpA) and 1 M HCl (in-house distilled) prior to isotopic analysis to counteract any potential anionic matrix effects. Isotope data quality was assessed through plotting delta29Si vs delta30Si. Data included in this archive were found to lie along a line of gradient 0.5119, in between that expected of kinetic and equilibrium Si isotopic fractionation. In addition, 71% of samples were measured in duplicate or triplicate (see 'n' for number of replicate measurements of a given sample and ±2 sigma for the standard deviation) and reference Si standards were analysed every 5 samples to determine the long term reproducibility. Measured standards in this study (Diatomite +1.24 ±0.14 per mil (n=116); LMG08 -3.47 ±0.13 per mil (n=46); ALOHA1000 =1.23 ±0.17 per mil (n=30)) agree well with published values (+1.26 ±0.2 per mil; -3.43 ±0.15 per mil; +1.24 ±0.2 per mil respectively) (Hendry et al., 2011; Hendry and Robinson, 2012 ; Grasse et al., 2017).

ChAOS_3_Core_Incubation: results of core incubation experiments carried out on-board (JR18006) for cores collected from stations B03, B13, B14, B15 and B16 to quantify the magnitude of the benthic flux of DSi. Includes DSi concentrations for samples extracted from the core top water at 3 hourly intervals over 24 hours and the Si isotopic composition (delta30SiInc per mil) of such samples at time 0hr, 3hr and 24hr. For B15, 0/6 hr and 21/24 hr samples were combined due to a lack of adequate sample volume. The gradient of the lin...(48)

Data collection:

Megacorer (multicoring device with up to 12 core tubes), National Marine Facilities

Lachat Quickchem 8500 flow injection autoanalyzer

VWR V-1200 spectrophotometer, University of Bristol

Finnigan Neptune Plus High Resolution MC-ICP-MS by Thermo Fisher Scientific, University of Bristol

Agilent Technologies 710 ICP-OES, University of Bristol