Description: The low surface nitrate concentration and high atmospheric iron input in the tropical eastern North Atlantic provide beneficial conditions for N-2 fixation. Varying abundances of diazotrophs have been observed and an Fe- and P-colimitation of N-2 fixation was reported in this ocean region. It is however unclear, how different limiting factors control the temporal variability of N-2 fixation and what the role of Fe-limitation is in a region with high fluxes of dust deposition. To study the environmental controls on N-2 fixation, an one-dimensional ecosystem model is coupled with a physical model for the Tropical Eastern North Atlantic Times-series Station (TENATSO), north of the Cape Verde Islands. The model describes diazotrophy according to the physiology of Trichodesmium, taking into account a growth dependence on light, temperature, iron, dissolved inorganic (DIP) and organic phosphorus (DOP). The modelled total Chl a is compared with satellitederived total Chl a and modelled Trichodesmium (Tri) compared with satellite-derived cyanobacterial Chl a as well as with High Performance Liquid Chromatography data. Model results show a complex pattern of competitive as well as mutually beneficial interactions between diazotrophs and non-diazotrophic phytoplankton. High DOP availability after spring blooms of non-diazotrophic phytoplankton and the ability of Trichodesmium to take up DOP are crucial for allowing a maximal abundance of Tri in autumn. Part of the reactive nitrogen newly fixed by diazotrophs is directly excreted or released through mortality, significantly fuelling the growth of non-diazotrophic phytoplankton in autumn and winter. Fe consumption by non-diazotrophic phytoplankton earlier in the year makes Fe limitation of Tri in late summer more acute, whereas Tri growth in surface waters reduces phytoplankton abundance deeper in the water column by light limitation. Overall, the atmospheric iron input at the TENATSO site is required to enable diazotrophic growth and to support the observed abundance of non-diazotrophic phytoplankton

Description: We analyzed 214 new core-top samples for their CaCO3 content from shelves all around Antarctica in order to understand their distribution and contribution to the marine carbon cycle. The distribution of sedimentary CaCO3 on the Antarctic shelves is connected to environmental parameters where we considered water depth, width of the shelf, sea-ice coverage and primary production. While CaCO3 contents of surface sediments are usually low, high (〉 15%) CaCO3 contents occur at shallow water depths (150–200 m) on the narrow shelves of the eastern Weddell Sea and at a depth range of 600–900 m on the broader and deeper shelves of the Amundsen, Bellingshausen and western Weddell Seas. Regions with high primary production, such as the Ross Sea and the western Antarctic Peninsula region, have generally low CaCO3 contents in the surface sediments. The predominant mineral phase of CaCO3 on the Antarctic shelves is low-magnesium calcite. With respect to ocean acidification, our findings suggest that dissolution of carbonates in Antarctic shelf sediments may be an important negative feedback only after the onset of calcite undersaturation on the Antarctic shelves. Macrozoobenthic CaCO3 standing stocks do not increase the CaCO3 budget significantly as they are two orders of magnitude lower than the budget of the sediments. This first circumpolar compilation of Antarctic shelf carbonate data does not claim to be complete. Future studies are encouraged and needed to fill data gaps especially in the under-sampled southwest Pacific and Indian Ocean sectors of the Southern Ocean.

Description: Most dissolved iron in the ocean is bound to organic molecules with strong conditional stability constants, known as ligands that are found at concentrations ranging from 0.2 to more than 10 nmol L− 1. In this work we report the first mechanistic description of ligand dynamics in two three-dimensional models of ocean biogeochemistry and circulation. The model for ligands is based on the concept that ligands are produced both from organic matter remineralization and phytoplankton processes, and that they are lost through bacterial and photochemical degradation, as well as aggregation and to some extent in the process of phytoplankton uptake of ligand-bound iron. A comparison with a compilation of in-situ measurements shows that the model is able to reproduce some large-scale features of the observations, such as a decrease in ligand concentrations along the conveyor belt circulation in the deep ocean, lower surface and subsurface values in the Southern Ocean, or higher values in the mesopelagic than in the abyssal ocean. Modeling ligands prognostically (as opposed to assuming a uniform ligand concentration) leads to a more nutrient-like profile of iron that is more in accordance with data. It however, also leads to higher surface concentrations of dissolved iron and negative excess ligand L⁎ in some ocean regions. This is probably an indication that with more realistic and higher ligand concentrations near the surface, as opposed to the traditionally chosen low uniform concentration, iron modelers will have to re-evaluate their assumption of low scavenging rates for iron. Given their sensitivity to environmental conditions, spatio-temporal variations in ligand concentrations have the potential to impact primary production via changes in iron limitation.

Description: The excretion of siderophores and the reduction of organic iron-complexes at the cell surface are common reactions of terrestrial plants, fungi and bacteria in response to low availability of iron. However, there is much less evidence for the use of these strategies by marine phytoplankton. It has been argued that siderophore excretion is inefficient in an aquatic environment due to rapid diffusion. This study examines how diffusion and chemical reactions in the microenvironment of a phytoplankton cell influence the efficiency of both strategies to increase the bioavailability of iron and to reduce iron stress. A numerical model of the cell surroundings is presented that calculates the concentration distribution for different iron species and allows to study the effect of siderophores or surface reductases. It calculates the efficiency of these mechanisms, defined as the quotient between the increase in iron uptake rate and the excretion rate of siderophores or electrons, needed to obtain this increase. The dependence of this efficiency on rates of iron coordination reactions, on diffusivity, and on the kinetics of iron uptake is discussed with the aid of some analytical calculations.

Description: Harde (2017) proposes an alternative accounting scheme for the modern carbon cycle and concludes that only 4.3% of today's atmospheric CO2 is a result of anthropogenic emissions. As we will show, this alternative scheme is too simple, is based on invalid assumptions, and does not address many of the key processes involved in the global carbon cycle that are important on the timescale of interest. Harde (2017) therefore reaches an incorrect conclusion about the role of anthropogenic CO2 emissions. Harde (2017) tries to explain changes in atmospheric CO2 concentration with a single equation, while the most simple model of the carbon cycle must at minimum contain equations of at least two reservoirs (the atmosphere and the surface ocean), which are solved simultaneously. A single equation is fundamentally at odds with basic theory and observations. In the following we will (i) clarify the difference between CO2 atmospheric residence time and adjustment time, (ii) present recently published information about anthropogenic carbon, (iii) present details about the processes that are missing in Harde (2017), (iv) briefly discuss shortcoming in Harde's generalization to paleo timescales, (v) and comment on deficiencies in some of the literature cited in Harde (2017).

Description: Most dissolved iron in the ocean is bound to organic molecules with strong conditional stability constants, known as ligands that are found at concentrations ranging from 0.2 to more than 10 nmol L− 1. In this work we report the first mechanistic description of ligand dynamics in two three-dimensional models of ocean biogeochemistry and circulation. The model for ligands is based on the concept that ligands are produced both from organic matter remineralization and phytoplankton processes, and that they are lost through bacterial and photochemical degradation, as well as aggregation and to some extent in the process of phytoplankton uptake of ligand-bound iron. A comparison with a compilation of in-situ measurements shows that the model is able to reproduce some large-scale features of the observations, such as a decrease in ligand concentrations along the conveyor belt circulation in the deep ocean, lower surface and subsurface values in the Southern Ocean, or higher values in the mesopelagic than in the abyssal ocean. Modeling ligands prognostically (as opposed to assuming a uniform ligand concentration) leads to a more nutrient-like profile of iron that is more in accordance with data. It however, also leads to higher surface concentrations of dissolved iron and negative excess ligand L⁎ in some ocean regions. This is probably an indication that with more realistic and higher ligand concentrations near the surface, as opposed to the traditionally chosen low uniform concentration, iron modelers will have to re-evaluate their assumption of low scavenging rates for iron. Given their sensitivity to environmental conditions, spatio-temporal variations in ligand concentrations have the potential to impact primary production via changes in iron limitation.

Description: Ikaite (CaCO3·6H2O) has only recently been discovered in sea ice, in a study that also provided first direct evidence of CaCO3 precipitation in sea ice. However, little is as yet known about the impact of physico-chemical processes on ikaite precipitation in sea ice. Our study focused on how the changes in pH, salinity, temperature and phosphate (PO4) concentration affect the precipitation of ikaite. Experiments were set up at pH from 8.5 to 10.0, salinities from 0 to 105 (in both artificial seawater (ASW) and NaCl medium), temperatures from 0 to −4 °C andPO4 concentrations from0 to 50 μmol kg−1. The results show that in ASW, calcium carbonate was precipitated as ikaite under all conditions. In the NaCl medium, the precipitates were ikaite in the presence of PO4 and vaterite in the absence of PO4. The onset time (τ) at which ikaite precipitation started, decreased nonlinearly with increasing pH. In ASW, τ increased with salinity. In the NaCl medium, τ first increased with salinity up to salinity 70 and subsequently decreased with a further increase in salinity; it was longer in ASW than in the NaCl medium under the same salinity. τ did not vary with temperature or PO4 concentration. These results indicate that ikaite is very probably the only phase of calcium carbonate formed in sea ice. PO4 is not, as previously postulated, crucial for ikaite formation in sea ice. The change in pH and salinity is the controlling factor for ikaite precipitation in sea ice. Within the ranges investigated in this study, temperature and PO4 concentration do not have a significant impact on ikaite precipitation.