Search

In the oceans’ high-nitrate–low-chlorophyll regions, such as the Peru/Humboldt Current system and the adjacent eastern equatorial Pacific1, primary productivity is limited by the micronutrient iron. Within the Peruvian upwelling area, bioavailable iron is released from the reducing continental margin sediments2. The magnitude of this seafloor source could change with fluctuations in the extension or intensity of the oxygen minimum zones3, 4. Here we show that measurements of molybdenum, uranium and iron concentrations can be used as a proxy for sedimentary iron release, and use this proxy to assess iron release from the sea floor beneath the Peru upwelling system during the past 140,000 years. We observe a coupling between levels of denitrification, as indicated by nitrogen isotopes, trace metal proxies for oxygenation, and sedimentary iron concentrations. Specifically, periods with poor upper ocean oxygenation are characterized by more efficient iron retention in the sediment and a diminished iron supply to the water column. We attribute efficient iron retention under more reducing conditions to widespread sulphidic conditions in the surface sediment and concomitant precipitation of iron sulphides. We argue that iron release from continental margin sediments is most effective in a narrow redox window where neither oxygen nor sulphide is present. We therefore suggest that future deoxygenation in the Peru upwelling area would be unlikely to result in increased iron availability, whereas in weaker oxygen minimum zones partial deoxygenation may enhance the iron supply.

We present iron (Fe) concentration and Fe isotope data for a sediment core transect across the Peru upwelling area, which hosts one of the ocean’s most pronounced oxygen minimum zones (OMZs). The lateral progression of total Fe to aluminum ratios (FeT/Al) across the continental margin indicates that sediments within the OMZ are depleted in Fe whereas sediments below the OMZ are enriched in Fe relative to the lithogenic background. Rates of Fe loss within the OMZ, as inferred from FeT/Al ratios and sedimentation rates, are in agreement with benthic flux data that were calculated from pore water concentration gradients. The mass of Fe lost from sediments within the OMZ is within the same order of magnitude as the mass of Fe accumulating below the OMZ. Taken together, our data are in agreement with a shuttle scenario where Fe is reductively remobilized from sediments within the OMZ, laterally transported within the anoxic water column and re-precipitated within the more oxic water below the OMZ. Sediments within the OMZ have increased 56Fe/54Fe isotope ratios relative to the lithogenic background, which is consistent with the general notion of benthic release of dissolved Fe with a relatively low 56Fe/54Fe isotope ratio. The Fe isotope ratios increase across the margin and the highest values coincide with the greatest Fe enrichment in sediments below the OMZ. The apparent mismatch in isotope composition between the Fe that is released within the OMZ and Fe that is re-precipitated below the OMZ implies that only a fraction of the sediment-derived Fe is retained near-shore whereas another fraction is transported further offshore. We suggest that a similar open-marine shuttle is likely to operate along many ocean margins. The observed sedimentary fingerprint of the open-marine Fe shuttle differs from a related transport mechanism in isolated euxinic basins (e.g., the Black Sea) where the laterally supplied, reactive Fe is quantitatively captured within the basin sediments

The molybdenum (Mo) stable isotope system has been applied to a variety of geochemical and environmental problems. In the absence of a universally accepted zero-delta reference material, different groups report their data relative to their adopted in-house standards. Rigorous comparison of results generated in different laboratories using different analytical approaches is only possible if the in-house standards are of identical Mo isotope composition. To determine potential isotopic differences among various standards, the δ98Mo (98Mo/95Mo) values of ten Mo standard solutions were measured as part of this study. For six of these solutions, four laboratories carried out an intercalibration. In contrast to previous results, δ98Mo of various in-house standards were found to differ by up to 0.37‰. Renormalisation of our new and published Mo-isotope data available for seawater taken from various sites and the USGS rock reference material SDO-1 relative to NIST-SRM-3134, provides a much better agreement among reported δ98Mo values for these samples. Relative to NIST-SRM-3134, the δ98Mo of SDO-1 is 0.80 ± 0.14‰ (2s), while oxic, open-ocean seawater is characterised by an average δ98Mo of 2.09 ± 0.10‰ (2s). This intercalibration provides a solid platform for comparing and amending existing δ98Mo values. In addition, we recommend that future Mo isotope studies adopt NIST-SRM-3134 as a universal zero-delta reference material.