Sulfur biogeochemical cycling and novel Fe–S mineralization pathways in a tidally re-flooded wetland

Sulfur biogeochemical cycling and associated Fe–S mineralization processes exert a major influence over acidity dynamics, electron flow and contaminant mobility in wetlands, benthic sediments and groundwater systems. While S biogeochemical cycling has been studied intensively in many environmental settings, relatively little direct information exists on S cycling in formerly drained wetlands that have been remediated via tidal re-flooding. This study focuses on a tidal wetland that was drained in the 1970s (causing severe soil and water acidification), and subsequently remediated by controlled re-flooding in 2002. We examine SO42- reduction rates and Fe–S mineralization at the tidal fringe, 7 years after the commencement of re-flooding. The initial drainage of the wetland examined here caused in-situ pyrite (FeS2) oxidation, resulting in the drained soil layers being highly acidic and rich in SO42--bearing Fe(III) minerals, including jarosite (KFe3(SO4)2(OH)6). Tidal re-flooding has neutralized much of the previous acidity, with the pore-water pH now mostly spanning pH 5–7. The fastest rates of in-situ SO42- reduction (up to ∼300 nmol cm−3 day−1) occur within the inter-tidal zone in the near-surface soil layers (to ∼60 cm below ground surface). The SO42- reduction rates correlate with pore-water dissolved organic C concentrations, thereby suggesting that electron donor supply was the predominant rate determining factor. Elemental S was a major short-term product of SO42- reduction, comprising up to 69% of reduced inorganic S in the near-surface soil layers. This enrichment in elemental S can be partly attributed to interactions between biogenic H2S and jarosite – a process that also contributed to enrichment in pore-water Fe2+ (up to 55 mM) and SO42- (up to 50 mM). The iron sulfide thiospinel, greigite (Fe3S4), was abundant in near-surface soil layers within the inter- to sub-tidal zone where tidal water level fluctuations created oscillatory redox conditions. There was evidence for relatively rapid pyrite re-formation within the re-flooded soil layers. However, the results indicate that pyrite re-formation has occurred mainly in the lower formerly drained soil layers, whereas the accumulation of elemental S and greigite has been confined towards the soil surface. The discovery that pyrite formation was spatially decoupled from that of elemental S and greigite challenges the concept that greigite is an essential precursor required for sedimentary pyrite formation. In fact, the results suggest that greigite and pyrite may represent distinct end-points of divergent Fe–S mineralization pathways. Overall, this study highlights novel aspects of Fe–S mineralization within tidal wetlands that have been drained and re-flooded, in contrast to normal, undisturbed tidal wetlands. As such, the long-term biogeochemical trajectory of drained and acidified wetlands that are remediated by tidal re-flooding cannot be predicted from the well-studied behaviour of normal tidal wetlands.