4.1 Regeneration incubation experiments
Three incubation experiments to probe for broad, differential effects on the regeneration of trace elements were conducted at the same location in the subantarctic Southern Ocean (Figure 1) within 10 days of each other, with varying experimental design (collection depth and incubation temperatures) and natural variability in the initial particulate regime and elemental concentrations. Only one incubation (Inc. 3) generated significant differences in [Cr] and δ53Cr by the final time point. We focus on this incubation here, with results from Inc. 1 and 2 provided in the supplemental material (Section S1, Figures S1, Tables S3-S4). Initial particulate P (pP), Mn (pMn) and Fe (pFe) from all three incubations are shown in Table S1. Flowcam imaging data indicate that particles in all three incubations were small (mostly ~2 µm, nearly all < 6 µm), with a high abundance of flagellates (data not shown), while regional dust deposition is very low (Mahowald et al., 2005), confirming a predominantly biogenic particulate composition.
There was minimal to no net regeneration of organic-associated macronutrients as indicated by the lack of macronutrient variability (N and P) (Figure 2, Table S2). Chromium concentrations, [Cr(III)] and δ53Cr remained stable over the first three days of the treatments. However, [Cr] increased by an average of 0.46 ± 0.14 and 0.56 ± 0.18 nmol kg-1 (1 SD of the triplicates) by the final sampling for the ambient and cold treatments, respectively (Figure 2), demonstrating that Cr release was independent from organic carbon respiration. Small increases in [Cr(III)] were also observed by the end of the incubation, though the increases in [Cr] were approximately seven times larger. Particulate Cr was probably primarily Cr(III) – Cr(III) dominates the Cr adsorbed to biogenic particles (e.g. Semeniuk et al., 2016) and the Cr content of calcium carbonate, hosting Cr(VI), is very low (Remmelzwaal et al., 2019). Therefore, due to the small changes in [Cr(III)], particulate Cr was either oxidatively released or rapidly oxidized after release.
Particulate Mn oxides are known to oxidize Cr(III), facilitating particulate Cr(III) dissolution (Oze et al., 2007), and incubated particles were enriched in Mn beyond typical cellular quotas (Table S1, Twining & Baines, 2013), indicating the presence of Mn oxides. Therefore, Cr oxidation coupled to the reduction of Mn oxides could explain dissolved Cr release without N and P regeneration or increases in [Cr(III)]. Indeed, the timing of the release of Cr in Inc. 3 matches a decrease in pMn (Figure 2). Differences between the three incubations can also be explained by Mn oxides. Incubations from 150 m had much higher pMn than those from shallower depths (Inc. 1, 100 m & Inc. 2; Table S1), following trends in oceanic pMn distributions (Ohnemus et al., 2019), and explaining the lack of Cr release in shallower treatments. Particulate Fe in Inc. 1 was higher than in Inc. 3 (Table S1), indicating more Fe-rich mixed oxides, which may impact Cr oxidation and explain differences between these incubations. The enriched incubation particle concentrations, possibly in combination with the formation of particle aggregates, would help to increase the probability of oxidative Cr release by particulate Mn oxides relative to ambient seawater conditions. Consequently, oxidative particulate Cr release driven by reduction with Mn oxides would be less likely in natural seawater, suggesting a deeper regeneration cycle for Cr than organic matter respiration influenced by more Mn-rich deeper particles (Ohnemus et al., 2019) and marine sediments.
The increase in [Cr] was accompanied by a statistically significant decrease in δ53Cr in four of six samples (Figure 2, Table S2). δ53Cr decreases were not significant for the remaining two samples, which is consistent with the small increase in [Cr] insufficient to significantly lower δ53Cr. Inter-replicate variability is likely controlled by small initial differences in biological communities and particle compositions in each cubitainer, hence real natural variability, rather than analytical uncertainty, given differences in [Cr] between triplicates were well outside of analytical uncertainty (Figure 2, Table S2). This release of low δ53Cr from particles provides evidence for predictions that Cr adsorbed onto biogenic particles is isotopically light (Scheiderich et al., 2015; Semeniuk et al., 2016; Janssen et al., 2020).
A strong linear relationship is observed between δ53Cr and ln[Cr] in the regeneration incubation samples (r2 = 0.89) with an implied enrichment factor (ε) = -0.66 ‰. This enrichment factor matches the global δ53Cr–[Cr] array (global ε ≈ -0.70 ‰), and the data plot along the global array (Figure 2). This supports release from biogenic particles as an important process in driving global [Cr] and δ53Cr distributions. While proposed earlier (Scheiderich et al., 2015), previous studies have highlighted that biogenic Cr accumulation was lacking strong support in oceanic depth profiles (Scheiderich et al., 2015; Goring-Harford et al., 2018; Moos & Boyle, 2019; Rickli et al., 2019). This study provides the first direct evidence that Cr release from biogenic particles acts as a control on δ53Cr across intermediate and deep waters, and ultimately also supports the inverse of this process – the adsorption of Cr onto biogenic particles in the surface ocean – as a control of δ53Cr and [Cr] (Scheiderich et al., 2015; Semeniuk et al., 2016; Goring-Harford et al., 2018; Janssen et al., 2020).