Formation and Maintenance of the GEOTRACES Subsurface‐Dissolved Iron Maxima in an Ocean Biogeochemistry Model

Abstract

Recent GEOTRACES transects revealed basin‐scale patterns of dissolved iron in the global oceans, providing a unique opportunity to test numerical models and to improve our understanding of the iron cycling. Subsurface maxima of dissolved iron in the upper ocean thermocline are observed in various transects, which can play an important role in regulating marine productivity due to their proximity to the surface euphotic layer. An ocean biogeochemistry model with refined parameterizations of iron cycling is used to examine the mechanisms controlling the formation and maintenance of these subsurface maxima. The model includes the representation of three iron sources including dust deposition, continental shelves, and hydrothermal vents. Two classes of organic ligands are parameterized based on the dissolved organic matter and apparent oxygen utilization. Parameterizations of particle‐dependent scavenging and desorption are included. Although the model still struggles in fully capturing the observed dissolved iron distribution, it starts reproducing some major features, especially in the main thermocline. A suite of numerical sensitivity experiments suggests that the release of scavenged iron associated with sinking organic particles forms the subsurface‐dissolved iron maxima in high‐dust regions of the Indian and Atlantic Oceans. In low‐dust regions of the Pacific basin, the subsurface‐dissolved iron extrema are sustained by inputs from the continental shelves or hydrothermal vents. In all cases, subsurface ligands produced by the remineralization of organic particles retain the dissolved iron and play a central role in the maintenance of the subsurface maxima in our model. Thus, the parameterization of subsurface ligands has a far‐reaching impact on the representation of global iron cycling and biological productivity in ocean biogeochemistry models.

Responses of ocean biogeochemistry to atmospheric supply of lithogenic and pyrogenic iron-containing aerosols

“…Accordingly, a more detailed model including a weaker ligand and larger concentration of total ligand has led to better model-measurement agreement (Parekh et al 2004). Currently, some ocean biogeochemistry models consider variability in Fe-binding ligands (Misumi et al 2013;Völker & Tagliabue, 2015;Pham & Ito, 2018), although some still assume a constant ligand concentration of 0.6 or 1 nM (Tagliabue et al 2016). Furthermore, few ocean biogeochemistry models consider scavenging of Fe onto mineral dust, in addition to the scavenging on organic particles (Moore & Braucher, 2008;Aumont et al 2015;Ye & Völker, 2017;Pham & Ito, 2018).…”

Ligand Binding Strength Explains the Distribution of Iron in the North Atlantic Ocean

“…Biological production is modulated by the availability of nutrients (PO 3− 4 and dFe) and light through Monod function. As in Pham and Ito (2018), we consider three external supplies for dFe: atmospheric deposition, continental shelves, and hydrothermal vents. In this model, the biological Fe:P uptake ratio changes as a function of the dFe concentration, which aims to represent the luxury Fe uptake of diatoms such as in the subarctic Pacific and the Southern Ocean (Ingall et al, 2013).…”

“…There is uncertainty in the binding strength of the refractory, humic-like L 3 , as several studies reported to be weaker than 10 11 L/mol (Gledhill & Buck, 2012); thus, we vary the magnitude of K L3 between 10 11 and 10 10 L/mol in the sensitivity experiments. The two-class parameterization of Pham and Ito (2018) is essentially the same as setting K L3 = K L2 = 10 11 L/mol, so a decreased retention of dFe is anticipated if we use K L3 < 10 11 L/mol. dFe not bound to ligands (free Fe, [Fe ′ ]) can be removed from the water column by scavenging onto lithogenic and organic particles, based on a first-order bulk scavenging rate, and by precipitation (Pham & Ito, 2018).…”

“…where the reference scavenging coefficient, K inorg0 , is the same as in Pham and Ito (2018; set to 2.0 ×10 −7 s −1 ), JFe dust is the atmospheric dFe dust flux at each model grid cell in the surface ocean, and ⟨JFe dust ⟩ is the global mean dust Fe flux. Thus, K inorg can vary spatially in the surface water as a function of the atmospheric dFe flux, but its value below the surface is still set to K inorg0 .…”

Processes affecting dissolved iron across the Subtropical North Atlantic: a model study

“…The Fe tracer in the model is the sum of the two forms dF e = F eL + F e . F e is calculated as in Parekh et al (2004) and represents only a small percentage of the total dFe pool. It is assumed that Fe and ligands are bound in a 1:1 ratio.…”