Figure 6: Sediment organic carbon metrics differed significantly in association with oxygen on a multiannual scale. Metrics assessed include iron-bound organic carbon (Fe-OC) (a), total sediment organic carbon (OC; b), and Fe-OC as a percentage of OC (c) in Falling Creek Reservoir (FCR) and Beaverdam Reservoir (BVR). Blue color indicates the reservoir-year with the highest mean oxygen (2019 in FCR) and letters delineate groups that are significantly different (p < 0.05; Table S1, S2, S3).

3.4 Experimental microcosm incubations revealed rapid effects of hypoxia on Fe and OC 

Experimental microcosm incubations successfully established four distinct oxygen regimes. DO concentrations increased rapidly when hypoxic microcosms were unsealed and decreased rapidly when microcosms were sealed (Figure 7). At the transition from hypoxic-to-oxic conditions, DO concentrations increased to approximately the same level as the continuous oxygen treatment (~7 mg/L) within one day. At the transition from oxic-to-hypoxic conditions, DO concentrations decreased below 1 mg/L within one day and declined to 0 mg/L by the end of the experiment.

Changes in oxygen conditions were associated with clear but asynchronous changes in aqueous OC and Fe. As microcosms switched from hypoxic-to-oxic conditions, TOC, DOC and total Fe decreased near synchronously, while dissolved Fe decreased below detection within one day of oxygen exposure. At the transition from oxic-to-hypoxic conditions, DOC and TOC rapidly increased to the same level as microcosms that had experienced continuous hypoxia (~10 mg/L; Figure 7). However, concentrations of both dissolved and total Fe only began to increase after three weeks of hypoxia (Figure 7). Measured DOC and TOC were strongly and linearly correlated, with DOC representing a mean of 96±14% of TOC (Figure S5); thus, we focus our discussion on DOC hereafter, but the same trends apply to TOC.

At the end of the experiment, sediment OC differed significantly among treatments (one-way ANOVA: F_3,20=9.09, p<0.001). Sediment OC was significantly higher in microcosms that started under oxic conditions (oxic: µ=4.6±0.3, oxic-to-hypoxic: µ=4.5±0.3) than microcosms that started under hypoxic conditions (hypoxic: µ=4.0±0.0, hypoxic-to-oxic: µ=4.1±0.2; Figure 8). Fe-OC did not differ significantly between treatments as a proportion of sediment mass (F_3,20=0.51, p=0.683) or as a proportion of sediment OC (F_3,20=2.40, p=0.098).

Speciation calculations (Table S5) based upon observed solution chemistry (Figure S6) suggest that oxygen conditions had primary control over Fe speciation, with a lesser impact on Fe-OC. The experiments that were maintained under hypoxic conditions had dominant Fe species of Fe(II), FeHCO₃⁺, FeCO₃, and FeSO₄ (all of these species contained Fe in its reduced state, Fe(II)). For all of the microcosms that were exposed to oxygen at any time (hypoxic-to-oxic, oxic-to-hypoxic, oxic), the dominant Fe species were Fe(OH)₃, Fe(OH)₂⁺, Fe(OH)₄^-, FeOH₂⁺ and FeHumate⁺ (all of these species contained Fe in the oxidized state, Fe(III)). pH remained circumneutral across all treatments (Figure S7). These results indicate that 1) exposure to oxic conditions at any time in the experiment shifted the dominant oxidation state to Fe(III); 2) under oxic conditions, and to a lesser extent, hypoxic conditions, Fe complexed with DOC.