This study presents new concentration measurements of dissolved rare earth elements (dREEs) along a full-depth east?west section across the tropical South Atlantic ($\sim$12?S), and uses these data to investigate the oceanic cycling of the REEs. Enrichment of dREEs, associated with the redox cycling of Fe?Mn oxides, is observed in the oxygen minimum zone (OMZ) off the African shelf. For deeper-waters, a multi-parameter mixing model was developed to deconvolve the relative importance of physical transport (i.e., water mass mixing) from biogeochemical controls on the dREE distribution in the deep Atlantic. This approach enables chemical processes involved in REE cycling, not apparent from the measurements alone, to be distinguished and quantified. Results show that the measured dREE concentrations below $\sim$1000 m are dominantly controlled (\ensuremath>75%) by preformed REE concentrations resulting from water mass mixing. This result indicates that the linear correlation between dREEs and dissolved Si observed in Atlantic deep waters results from the dominantly conservative behavior of these tracers, rather than from similar chemical processes influencing both dREEs and Si. Minor addition of dREEs ($\sim$10% of dNd and $\sim$5% of dYb) is observed in the deep (\ensuremath>$\sim$4000 m) Brazil Basin, resulting from either remineralization of particles in-situ or along the flow path. Greater addition of dREEs (up to 25% for dNd and 20% for dYb) is found at $\sim$1500 m and below $\sim$4000 m in the Angola Basin near the African continental margin. Cerium anomalies suggest that different sources are responsible for these dREE addition plumes. The 1500 m excess is most likely attributed to dREE release from Fe oxides, whereas the 4000 m excess may be due to remineralization of calcite. Higher particulate fluxes and a more sluggish ocean circulation in the Angola Basin may explain why the dREE excesses in this basin are significantly higher than that observed in the Brazil Basin. Hydrothermal venting over the mid-Atlantic ridge acts as a regional net sink for light REEs, but has little influence on the net budget of heavy REEs. The combination of dense REE measurements with water mass deconvolution is shown to provide quantitative assessment of the relative roles of physical and biogeochemical processes in the oceanic cycling of REEs.This is the author accepted manuscript. The final version is available from Elsevier via http://dx.doi.org/10.1016/j.gca.2016.01.01

The Arabian Sea harbours one of the three major oxygen minimum zones (OMZs) in the world's oceans, and it alone is estimated to account for ~10–20 % of global oceanic nitrogen (N) loss. While actual rate measurements have been few, the consistently high accumulation of nitrite (NO₂⁻) coinciding with suboxic conditions in the central-northeastern part of the Arabian Sea has led to the general belief that this is the region where active N-loss takes place. Most subsequent field studies on N-loss have thus been drawn almost exclusively to the central-NE. However, a recent study measured only low to undetectable N-loss activities in this region, compared to orders of magnitude higher rates measured towards the Omani Shelf where little NO₂⁻ accumulated (Jensen et al., 2011). In this paper, we further explore this discrepancy by comparing the NO₂⁻-producing and consuming processes, and examining the relationship between the overall NO₂⁻ balance and active N-loss in the Arabian Sea. Based on a combination of ¹⁵N-incubation experiments, functional gene expression analyses, nutrient profiling and flux modeling, our results showed that NO₂⁻ accumulated in the central-NE Arabian Sea due to a net production via primarily active nitrate (NO₃⁻) reduction and to a certain extent ammonia oxidation. Meanwhile, NO₂⁻ consumption via anammox, denitrification and dissimilatory nitrate/nitrite reduction to ammonium (NH₄⁺) were hardly detectable in this region, though some loss to NO₂⁻ oxidation was predicted from modeled NO₃⁻ changes. No significant correlation was found between NO₂⁻ and N-loss rates (p>0.05). This discrepancy between NO₂⁻ accumulation and lack of active N-loss in the central-NE Arabian Sea is best explained by the deficiency of labile organic matter that is directly needed for further NO₂⁻ reduction to N₂O, N₂ and NH₄⁺, and indirectly for the remineralized NH₄⁺ required by anammox. Altogether, our data do not support the long-held view that NO₂⁻ accumulation is a direct activity indicator of N-loss in the Arabian Sea or other OMZs. Instead, NO₂⁻ accumulation more likely corresponds to long-term integrated N-loss that has passed the prime of high and/or consistent in situ activities.

Dissolved rare earth element (REE) concentration data from intermediate and deep seawater form an array characterized by higher middle‐REE enrichments (MREE/MREE*) in the North Atlantic and a progressive increase in heavy‐to‐light REE ratios (HREE/LREE) as water masses age. The REEs in foraminifera are fractionated toward higher MREE/MREE* and lower HREE/LREE relative to seawater. Calculations based on a scavenging model show that the REE patterns in uncleaned core‐top foraminifera resemble those adsorbed onto calcite, particulate organic material, and hydrous ferric oxides but the full extent of the REE fractionation measured in foraminifera was not reproduced by the model. However, differences in the HREE/LREE and MREE/MREE* ratios and the cerium anomaly between ocean basins are preserved and are in agreement with the seawater REE distribution. Under oxic conditions, the HREE/LREE and MREE/MREE* compositions of uncleaned foraminifera at the sediment/seawater boundary are preserved during burial but the cerium anomaly is sensitive to burial depth. In suboxic sedimentary environments, all uncleaned foraminiferal REE concentrations are elevated relative to core‐top values indicating addition of REEs from pore waters. The HREE/LREE ratio is highest when sedimentation rates were greatest and when high Fe/Ca ratios in the uncleaned foraminifera indicate that Fe was mobile. In sediments that have not experienced suboxic conditions during burial, uncleaned foraminifera preserve the seawater signal taken up at the sediment/seawater interface and are therefore suggested to be a suitable archive of changes in the REE signal of past bottom waters.