320sediment cores collected at three soft bottom stations; two brackish-marine and one freshwater. One of the marine stations was reduced and azoic, whereas the freshwater and the other marine station had well oxygenated conditions in the bottom waters. Positive redox-turnovers, including anaerobic incubation followed by reaeration, were generated in the cores and the supernatant water.In cores from the oxygenated freshwater and marine stations, dissolved phosphate and ferrous ions were released from the sediment during the anaerobic incubation. At the positive redox-turnover, the concentration of dissolved phosphate in the supernatant water decreased sharply due to scavenging by rapidly formed colloidal ferric hydroxide. Dissolved phosphate was also released during the incubation of the marine sediment from the reduced station. However, in these cores the concentration of iron in the supernatant water was low throughout the experiment and after the redox-turnover phosphate remained dissolved. In a parallel experiment in which iron was added to the supernatant water, dissolved phosphate was scavenged by ferric hydroxide at the positive redox-turnover in a similar way as observed for the two oxygenated stations. The low abundance of dissolved iron in the reduced marine system could be due to a rich supply of sulphide.In freshwater systems the concentration of dissolved phosphate is effectively diminished after a positive redox-turnover due to interaction with iron. In marine systems, which have had prevailed reduced conditions in the bottom waters, iron is immobilised. Consequently, a potent retention mechanism for phosphorus is eliminated. Our results imply that the cycling of phosphorus, in this aspect, differs in fresh and saltwater systems. This difference might have large effects on the availability of phosphorus as a nutrient.Manganese showed a consistent redox-dependent behaviour in all systems, but it did not interact with iron or phosphorus.The concentration of dissolved phosphate in sediment box-cores from the Laurentian Trough in the Gulf of St. Lawrence increased sharply across the sediment-water interface from 2.0 #mol/PO 4 /1 in the bottom water to 6 + 3 tmol PO 4 /1 in the top cm, remained almost constant at this value down to 5-15 cm depth, and then increased rapidly with further depth. In the region of constant concentration, phosphate is buffered by sorption equilibria with the sediment. The production rate of phosphate, the sorption capacity of the sediment, and the thickness of the diffusive boundary layer at the sediment-water interface appear to control the shape of the pore water profile. Even though the buffering places an upper limit on the concentration gradient across the sediment-water interface, and hence on the flux, the phosphate flux to the overlying water is controlled by the production rate of phosphate within the sediment. A model is proposed to relate sediment chemistry to phosphorus fluxes.Approximately half of the net sedimentation flux of phosphorus is not buried but is mobil...

A diagenetic reaction-transport model was used to simulate how the sediment redox boundary migrates in response to persistent or episodic changes in the deposition flux of degradable organic matter and the concentration of oxygen in the overlying bottom water. The position of the redox boundary is represented by the depth of oxygen penetration. The simulations reveal that the position of the redox boundary in organic-poor sediments, such as those in the deep basins of the Arctic Ocean, is highly sensitive to the flux of organic matter: relatively small and/or brief increases in that flux can cause the redox boundary to migrate rapidly from deep within the sediment to within a few centimeters of the sediment-water interface. Reoxidation of the sediment column after such an event can take years. Redox fluctuations can redistribute solid-phase manganese within the sediment column and produce multiple concentration peaks in its depth profile on a decadal time scale. Manganese peaks observed in sediment cores from the deep basins of the Arctic Ocean do not necessarily correspond to the position of the redox boundary during previous climatic periods or reflect historical changes in manganese deposition rates. The model supports the hypothesis that the recent decrease in the Arctic ice cover has increased the flux of organic matter to the seafloor and moved the redox boundary close to the sediment-water interface. The presence of iron sulfides at depths significantly below the bioturbated layer suggests that either the Arctic sediments have been anoxic for millennia, or iron and sulfate are reduced at these depths by dissolved organic matter diffusing downward from the bioturbation zone.