The low-temperature geochemical cycle of iron: From continental fluxes to marine sediment deposition

Abstract

Suspended sediments from 34 major rivers (geographically widespread) and 36 glacial meltwater streams have been examined for their variations in different operationally-defined iron fractions; Fe~HR~ (iron oxides soluble in dithionite), Fe~PR~ (iron soluble in boiling HCl but not in dithionite) and Fe~U~ (total iron less that soluble in boiling HCl). River particulates show a close association between Fe~HR~ and total iron (FeT), reflecting the effects of chemical weathering which derive oxide iron from, and retain it in close association with, total iron. Consistent with this, continental-scale average Fe~HR~/FeT ratios vary with runoff ratios (average river runoff per unit area/average precipitation per unit area). By contrast, the diminished effects of chemical weathering produce no recognizable association of Fe~HR~ with FeT in glacial particulates, and instead both Fe~PR~ and Fe~U~ are closely correlated with FeT, reflecting essentially pristine mineralogy. A comparison of the globally-averaged compositions of riverine particulates and marine sediments reveals that the latter are depleted in Fe~HR~, Fe~PR~ and FeT but enriched in Fe~U~. The river and glacial particulate data are combined with estimates of authigenic, hydrothermal, atmospheric and coastal erosive iron fluxes from the literature to produce a global budget for Fe~HR~, Fe~PR~, Fe~U~ and FeT. This budget suggests that the differences between riverine particulates and marine sediments can be explained by; (i) preferentially removing Fe~HR~ from the riverine particulate flux by deposition into inner shore reservoirs such as floodplains, salt marshes and estuaries; and (ii) mixing the resulting riverine particulates with Fe~HR~-depleted glacial particulates. Preliminary measurements of inner shore sediments are consistent with (i) above. Phanerozoic and modern normal marine sediments have similar iron speciation characteristics, which implies the existence of a long-term steady state for the iron cycle. This steady state could be maintained by a glacioeustatic feedback, where Fe~HR~-enriched riverine particulates are *either* more effectively trapped when sealevel is high (small ice masses, diminished glacial erosion), *or* are mixed with greater masses of Fe~HR~-depleted glacial particulates when sealevel is low (large ice masses, enhanced glacial erosion). Further important controls on the steady state for Fe~HR~ operate through the formation of euxinic sediments and ironstones, which also provide sealevel-dependent sinks for Fe~HR~-enriched sediment.