Sediment and pore water collection:

Short sediment cores were collected using a Bowers & Connelly megacorer, a multiple coring device that can collect ~20-40 cm long sediment cores with undisturbed sediment surfaces. At two sites (stations 41 and 64) longer cores (up to ~2 m) were also collected with a Kasten corer.

Megacorer cores were either sectioned for solid phase analysis, profiled with polarographic micro-electrodes to determine dissolved O2 concentrations, or sectioned in a cold van under N2 for pore water sample extraction (for details see, Komada et al., 2016). Samples for solid phase analyses were placed in Whirl-Pak plastic bags and frozen for solid phase Fe speciation analyses at ODU.

Kasten cores were brought into a large cold room on-board ship, laid on their side and one side of the core box removed to expose the sediment in the core. A plastic block was placed against the top of the core to prevent slumping of the sediment during processing, and pore waters were collected from these cores using Rhizon samplers (Seeberg-Elverfeldt et al., 2005) inserted directly into the cores at measured intervals. After pore water sampling was complete, samples for solid phase analyses were removed from the cores with plastic spoons and again placed in Whirl-Pak plastic bags (Fe speciation analyses) and immediately frozen.

While it is possible to recover cores with intact sediment-water interfaces using a megacorer, loss of surface sediments is typical during Kasten coring, making it not possible to directly determine absolute depths below the sediment-water interface in a Kasten core. We therefore determined the absolute depths of pore water and solid phase sample intervals from Kasten cores by aligning Kasten core profiles of pore water alkalinity to megacore alkalinity profiles from the same site (Berelson et al., 2005; Komada et al., 2016).

Sediment iron speciation:

This was determined using sequential extraction techniques (Goldberg et al., 2012; Poulton and Canfield, 2005). Sediments were freeze-dried and homogenized before use, and in each step a 10 ml extraction volume was used (except where noted) starting with 200-300 mg of sediment. Samples were shaken during all extractions, except when heated during extractions. At the end of each extraction step, the samples were centrifuged, the extract solution was removed, and the sediments were then rinsed twice with distilled, deionized water before moving on to the next extraction. Except where noted all extracts were analyzed for iron by flame AAS (Atomic Absorption Spectrometry).

Sediments were initially treated with 0.5 M HCl for 1 h to remove highly reactive, poorly crystalline iron oxides such as ferrihydrite and lepidochrocite, as well as any unsulfidized Fe(II) produced during early diagenesis. Fe(II) released by the 0.5 M HCl extraction (termed Fe_II_HCl) was determined immediately by the ferrozine method (Viollier et al., 2000) without the addition of hydrolyamine HCl using an aliquot of the extract solution. Subtraction of the Fe concentration determined by the Fe_II_HCl measurement from the total Fe present in the 0.5 M HCl extract (Fe_HCl) yields the concentration of highly reactive, poorly crystalline Fe oxides (Fe_ox1),

Fe_ox1 = Fe_HCl – Fe_II_HCl (1)

Next, the sediment was extracted for 6 hr with a citrate-dithionite solution (50 g /l sodium dithionite buffered to pH 4.8 with 0.35 M acetic acid/0.2 M sodium citrate) to extract less reactive crystalline iron oxides such as goethite and hematite (Fe_ox2). After this, the sediment was extracted with ammonium oxalate (0.2 M ammonium oxalate/0.17 M oxalic acid) at pH 3.2 for 6 hr to dissolve iron in the mineral phase magnetite (Fe_mag). Finally, the remaining sediment was placed in boiling 12N HCl (5 ml) for 1 min to extract Fe found in poorly reactive sheet silicates (Fe_prs; i.e., “structural” Fe(III) in clays).

Total sediment iron (Fe_T) was determined in a separate sediment aliquot by ashing the sediment at 450°C for 8 h followed by extraction for 24 h in near boiling 6 N HCl. Finally, iron in the sulfide-containing phases AVS and pyrite (termed Fe_pyr) was determined in a separate sediment aliquot by acidic chromium reduction/distillation and colorimetric analysis of the sulfide liberated by the process. This procedure was based on that described in Canfield et al. (1986) with the exception that we used 150-200 mg sediment samples, and collected the sulfide produced by the distillation process in three sequential traps (trap volumes of 30, 30 and 20 ml) containing 5 mM each ZnCl2 and NaOH (final concentrations; Ingvorsen and Jørgensen, 1979). The distillation was done using a sparging rate of 250 ml N2/min for 45 – 60 min. Sulfide in the traps was determined colorimetrically (Cline, 1969), and the concentration of iron in this pool was calculated by assuming that all of the sulfide liberated by this procedure is pyrite-S (i.e., that there is no AVS in these sediments) with a 1:2 Fe:S molar ratio in the pyrite.

Finally, we also defined a pool of unreactive iron (Fe_U) whose concentration is given by

Fe_U = Fe_T – (Fe_ox1 + Fe_ox2 + Fe_II_HCl + Fe_mag + Fe_prs + Fe_pyr ) (2)

This iron is presumably found in mineral phases that are even less reactive towards reductive dissolution than iron in any of these other extracts.