Mercury speciation and isotopic fractionation processes
have been incorporated into the HAMOCC offline ocean tracer advection code.
The model is fast enough to allow a wide exploration of the sensitivity of
the Hg cycle in the oceans, and of factors controlling human exposure to
monomethyl-Hg through the consumption of fish. Vertical particle transport
of Hg appears to play a discernable role in setting present-day Hg
distributions, which we surmise by the fact that in simulations without
particle transport, the high present-day Hg deposition rate leads to an Hg
maximum at the sea surface, rather than a subsurface maximum as observed. Hg
particle transport has a relatively small impact on anthropogenic Hg uptake,
but it sequesters Hg deeper in the water column, so that excess Hg is
retained in the model ocean for a longer period of time after anthropogenic
Hg deposition is stopped. Among 10 rate constants in the model, steady-state
Hg concentrations are most sensitive to reactions that are sources or sinks
of Hg(0), the evasion of which to the atmosphere is the dominant sink term
in the surface ocean. Isotopic fractionations in the interconversion
reactions are most strongly expressed, in the isotopic signatures of
dissolved Hg, in reactions that involve the dominant dissolved species,
Hg(II), including mass independent fractionation during Hg photoreduction.
The
The element mercury (Hg) is a powerful neurotoxin (Clarkson and Magos, 2006). When transformed to methyl mercury (MeHg) it is known to amplify its toxicity by bio-accumulating up the food chain. The main human exposure to MeHg is via consumption of high trophic level seafood (Chen et al., 2016; Schartup et al., 2018). Humans have been mining and mobilizing Hg into the Earth surface environment for hundreds of years, as a by-product of coal combustion, for its use in gold mining, and in products such as electronics and light bulbs (Amos et al., 2013; Driscoll et al., 2013; Krabbenhoft and Sunderland, 2013; Lamborg et al., 2014; Obrist et al., 2018; Streets et al., 2017; Mason et al., 2012). The Hg load in the surface ocean has increased by a factor of 3–5 since the industrial revolution; this represents a massive human impact on the global Hg cycle (Streets et al., 2017).
Hg can be extremely mobile in the environment, with gaseous forms in the atmosphere, and with particle-reactive forms allowing it to travel through soils and rivers and into the oceans (Fitzgerald et al., 2007). Hg(II) has a high affinity for complexing with (or adsorbing to) sulfur-rich ligands in organic matter (Schartup et al., 2015) and this leads to Hg accumulation with organic carbon in soils (Amos et al., 2013; Smith-Downey et al., 2010; Biswas et al., 2008) and sediments (Hollweg et al., 2010). The high mobility of Hg implies that the amount of Hg in Earth surface reservoirs is transient, even in the steady-state prehuman Hg cycle (Amos et al., 2013), and that Hg can be potentially mobilized by human impacts such as the thawing of Arctic permafrost (Schuster et al., 2018; Obrist et al., 2017) or enhanced wildfire activity (Turetsky et al., 2006).
The Hg cycle is analogous to the carbon cycle, in which fossil fuel extracted from the solid Earth is released to a fast surface system consisting of soils and oceans in communication via the atmosphere. In both cases, the long-term sink for the perturbation is burial in sediments of the ocean. Because these burial fluxes are relatively slow, it will take a long time for these perturbations to subside: thousands of years for the Hg cycle (Amos et al., 2013), and hundreds of thousands of years for the carbon cycle (Archer et al., 2009). Other forms of environmental degradation that will persist for thousands of years include actinide radioactive waste, and some anthropogenic gases such as sulfur hexafluoride (Ray et al., 2017).
It is extremely challenging to predict the future of human exposure to Hg, because the Hg cycle is so complex (Blum, 2013). One challenge has been to characterize the quantitative role of Hg adsorbed onto sinking particles in the ocean (Lamborg et al., 2016), which will constrain how deeply anthropogenic Hg may have penetrated into the ocean (Lamborg et al., 2014; Munson et al., 2015; Zhang et al., 2014b). Another is to understand the factors that control the production of MeHg, which is the bio-accumulating form but which comprises only a small fraction of the Hg in the ocean (Schartup et al., 2013, 2015; Ortiz et al., 2015; Lehnherr et al., 2011; Lehnherr, 2014; Jonsson et al., 2016; Chakraborty et al., 2016; Blum et al., 2013).
Stable isotopes provide a powerful tool for determining the origins (Kwon
et al., 2014; Li et al., 2014; Sherman et al., 2013, 2015; Balogh et al., 2015; Demers et al., 2015; Donovan et al., 2013, 2014; Gehrke et al., 2011;
Sun et al., 2016; Tsui et al., 2014; Sonke,
2010; Yin et al., 2013) and transformations (Kwon et al., 2013,
2014; Rodriguez-Gonzalez et al., 2009; Chandan et al., 2015; Yang and Sturgeon,
2009; Foucher, 2013; Jiskra, 2012) of Hg in the natural environment. Hg
has seven stable isotopes, with six at high abundance (
We have incorporated a model of the chemical transformations and isotopic
fractionations of Hg in the ocean into the HAMOCC offline ocean passive
tracer advection model (Maier-Reimer and Hasselmann, 1987). The flow
field is taken from the large-scale geostrophic (LSG) dynamics model, which
is also extremely fast and efficient for 3-D ocean flow
(Maier-Reimer, 1993). The LSG physical model takes a time step of a
month by eliminating non-geostrophic parts of the circulation that would be
violated by this extremely long time step. The HAMOCC tracer advection model
takes an annual average flow field from 12 monthly time steps of the LSG
model and uses it to advect tracers through the ocean. While the tracers are
flowing, they are exchanged with the atmospheric gases (in the case of
The distribution of Hg in the ocean today is the product of a presumably steady-state natural Hg cycle, which takes thousands of years to achieve in the model due to the ocean turnover time, followed by a global human perturbation, which began in about 1850 (and which could persist for thousands of years into the future). HAMOCC is believed to still be the fastest offline 3-D ocean tracer advection code in existence and is ideal for studying the sensitivity of the ocean Hg cycle on these long timescales. This paper is also the first attempt in our knowledge to simulate the isotopic fractionation processes of Hg in the ocean, which take thousands of years to express themselves globally.
Schematic of the reaction web for Hg speciation in the model.
Starred numbers by the arrows show typical values for biologically mediated
rate constants, measured as per year (yr
The geochemical cycling of Hg in the ocean in Hg-HAMOCC is similar in conception to previous models (Fig. 1). Hg interconverts between Hg(II), Hg(0), monomethyl-Hg (MMHg), dimethyl-Hg (DMHg), and Hg adsorbed to sinking particles (Hg-P). The rates of the biological reactions are correlated to each other and to the overall rate of metabolic activity in the Semeniuk and Dastoor (2017) model and in our model, with typical values as shown in Fig. 1. The rate constant for MMHg production from Hg(II) is proportional to the rate of particulate organic carbon (POC) degradation, which is derived from the attenuation with depth of the sinking POC flux in HAMOCC (expressed as a volumetric rate of POC degradation).
Other Hg transformation reactions are provoked by light (Blum et al.,
2014; Bergquist and Blum, 2007) and only take place in the surface ocean.
The rate of photochemical reactions in Hg-HAMOCC is about a factor of 2
higher in low versus high latitudes, using the latitudinal function that
governs export production rates in HAMOCC. The photochemical reaction rates
are attenuated with water depth, using an
All of the rate constants in Hg-HAMOCC are first-order, which is to say that the chemical rates are determined by multiplying the rate constant by a single species concentration to the first power. Rates of conversion between these species are generally fast, some much faster than the 1-year time step of the tracer code. For this reason, solvers were written to find steady-state distributions of the Hg species. Because the Hg system is strongly driven at the sea surface by air–sea fluxes, a separate solver system was developed for surface grid points than the one applied to subsurface grid points.
Export of sinking Hg-P is done separately from the speciation calculations
in subsurface waters, but simultaneously with the speciation calculations in
the surface ocean. Hg advection by ocean circulation is also done in an
independent step from the chemistry and particle components. Since the Hg
speciation is imposed to be at equilibrium by the speciation solvers, there
is no need to carry speciation information through the advective system,
which only needs to carry around a single tracer for the total Hg
concentration. To treat the isotopic systematics of the Hg cycle, we added
three additional advected Hg tracers, identical to the first but with
slightly altered source fluxes or rate constants, in order to simulate
variations in the relative abundances of isotopes
For the surface ocean, the distribution of Hg among the dissolved species is
determined by a balance of Hg fluxes through the system: rain input of
Hg(II) and Hg(0), and removal by Hg(II) scavenging on sinking particles and
degassing as Hg(0) and DMHg. Concentration-dependent reaction rates in the
model are all assumed to be first-order, i.e., linear in Hg concentration.
This includes loss by gas evasion, which should be linear in Hg
concentration in the piston velocity model, and loss of bound Hg on sinking
particles, which is linear with [Hg(II)] in the adsorption model. The solver
finds values for the Hg species concentrations at which the incoming and
loss fluxes balance. The equations are as follows:
Comparison of Hg(tot) concentrations in pM from the model with
data in the top row, from Laurier et al. (2004), Bowman et al.
(2015), Bowman et al. (2016), Cossa et al. (2004), Cossa et al. (2011), Hammerschmidt
and Bowman (2012), Lamborg et al. (2012), Mason et al. (2001), and Mason et al.
(1998), at approximately the depths in the ocean given at the top. The lower
three rows are present-day (year 2010) model results using different values
of the particulate-bound Hg sinking velocity as indicated by the labels on
the left, with 500 m yr
The Hg cycle in the deep ocean differs from that of the surface in that
fluxes of Hg into and out of the system (by desorption of Hg(II) from
particles) are slow relative to the rates of interconversion between the Hg
species. Because all of the rate constants are first-order, the relative
proportions of the species are independent of the total Hg concentration.
The solver finds steady-state values of all species relative to that of
Hg(II), then scales everything to fit the total Hg concentration as produced
by the advection routine. The equations are as follows:
Depth profiles (in meters) of the total Hg concentration in the
model: global mean and from locations shown in Fig. 4, showing preanthropogenic
(1850), present-day (2010), and the difference between the two, for
different values of the particulate-bound Hg flux in meters per year (m yr
Hg has a strong chemical affinity for organic matter, in particular for organic sulfur ligands. This chemistry leads Hg to adsorb onto organic matter in the ocean, leading to a vertical sinking flux of adsorbed Hg on particles (Lamborg et al., 2016). Characterizing this flux is complicated by the fact that sinking particles compete for Hg with suspended and dissolved organic carbon (Han et al., 2006; Fitzgerald et al., 2007).
The biological pump in HAMOCC is represented as an instantaneous vertical
redistribution of nutrients and other associated biological elements,
without ever resolving them into particles or tracking their sinking. We
constructed a hypothetical POC profile from this functioning of HAMOCC by
choosing a POC sinking velocity that would transform the export production
from the euphotic zone in HAMOCC into surface POC concentrations that are
close to the observed mean concentration of about 5
A map of the bound fraction of Hg(II) (relative to bound
Because POC in the real ocean varies in size from dissolved to fast-sinking,
the imposition of a single velocity in the model formulation, to be applied
to the entire adsorbed Hg pool, is an oversimplification of reality, and the
velocity required for the best fit is not a simple thing that can be
measured directly in the real ocean. The sensitivity of the model to the
sinking velocity is shown in Fig. 4. With an increase in sinking velocity,
the “biological pump” for Hg becomes stronger, increasing the
concentration in the deep ocean. The scavenging lifetime of Hg decreases as
the sinking flux increases with increasing sinking velocity. When the Hg
sinking velocity is set to 500 m yr
Depth profiles (in meters) of the bound fraction of Hg(II) from
the same locations as in Fig. 3, as a function of the
A second degree of freedom in the system of sinking Hg on particles is the
adsorption constant
Global model fluxes as a function of Hg(II) sinking velocity
imposed in the model. Colors represent preanthropogenic and present-day
results from our model. For comparison broken lines are model results from
Zhang et al. (2014a) and Semeniuk and Dastoor (2017),
sediment trap data from Mason et al. (2012),
17
Mercury isotope fractionations associated with any of the processes in the
Hg cycle are treated in the model as kinetic effects: slight perturbations
in the rates of chemical transformations between the isotopes (rather than
fractionation of the equilibrium state). This allows Hg-HAMOCC to impose
fractionation effects onto the kinetic expressions in the solvers for
surface and subsurface Hg speciation. The altered kinetic rate constants are
applied to alternative total Hg concentration fields representing the
isotopes
The way that the model treats the isotopes differs from reality, for a numerical convenience, following a technique developed by Ernst Maier-Reimer in HAMOCC many years ago for carbon isotope ratios (Maier-Reimer, 1984). In the real world, the total Hg concentration is comprised of multiple isotopes. In the model, the concentrations of Hg in the ocean are taken as that of a base isotope. Then the entire Hg cycle in the model is duplicated, and the kinetic constants slightly altered, to represent the behavior of a different Hg isotope. Each isotopic field corresponds to how the total Hg field would behave if it were entirely comprised of its particular Hg isotope, subject to slightly altered sources and kinetic rate constants for that isotope.
Profiles of the turnover time of Hg(II) with respect to transiting
through the system on sinking particles, as a function of the sinking
velocity in the legend, in meters per year (m yr
The deviations of the fields for the other isotopes are represented as ratios relative to the base field, and presented in per mill (‰), where the ratio of the “standard” is 1 rather than the particular ratio of the isotopic reference standard for natural samples. The relative differences, represented as ratios in per mill, are the same between variations in the isotopes in reality and between the altered fields in the model, even though the concentrations are different between the two cases. The advantage of this scheme is that the fields representing the different isotopes are subjected to similar computational rounding errors, because their values are similar. Also, it is simpler to simulate the behavior of total Hg in a single field, rather than as a more complicated sum of isotopic concentrations as in reality.
Because the chemical speciation of Hg is solved for each time step, there is no need to advect the concentrations of chemical species such as MMHg. The advection scheme in the model carries the total concentrations representing each isotope. Each isotope field is divided into the different Hg species, assuming steady state and using the web of kinetic rate constants appropriate to that isotope. The slightly altered speciation of one isotope relative to another, and the slightly differing sources and sinks for that isotope, lead to slight differences between the abundances of each isotope overall in the Hg pool.
Mass-dependent fractionation processes are imposed on all isotopic systems,
with the rates depending on how much heavier an isotope is than mass 198.
Mass-independent fractionations in the ocean are applied only to the
Human activity has resulted in significantly increased Hg emission to the global biosphere since about 1850 (Streets et al., 2011, 2017; Amos et al., 2013; Horowitz et al., 2014), which has lead to an increase in Hg deposition to the ocean. Because of the tendency for Hg to recycle in the environment, the relationship between emissions and deposition is not simple and immediate, but rather reflects the entire cumulative emission and re-emission of Hg. Guided by a reconstructed history of atmospheric Hg through time (Streets et al., 2017), we subject our model to a 4-fold increase in Hg deposition, following an initial spin-up equilibration period of 10 000 years. The beginning of the anthropogenic period corresponds to approximately the year 1850. We show natural steady-state results from model year 1850, which are useful for understanding how the ocean Hg cycle works, and contemporary results from model year 2010, for comparison with field measurements. Anthropogenically enhanced deposition is continued at a constant rate until the year 2100, after which we follow two scenarios: an abrupt and unrealistic return to natural Hg deposition fluxes, useful to determine the time constant of the oceanic recovery, and a “hangover” scenario in which an abrupt cessation of human Hg emissions triggers a gradual slowdown of enhanced deposition, over an ocean overturning timescale of 1000 years.
The steady-state assumption in the Hg solvers limits the ability of Hg-HAMOCC to explore detailed shallow-water interactions of turbulence, ventilation, and photochemistry, and the physics of the tracer advection code preclude exploration of processes on short timescales, such as the seasonal cycle near the surface. The model allows us to explore the interaction of the Hg chemistry and particle adsorption with the ocean circulation on long timescales.
A peculiarity of the surface ocean solver is that fluxes of Hg across the
sea surface are always locally balanced, by construction, neglecting the
impact of any upwelling Hg driving sea surface Hg concentrations and evasion
rates to higher values. Similarly to the treatment of
There are two competing mechanisms for Hg invasion into the deep ocean: advection by the overturning circulation and the flux of Hg adsorbed on sinking particles. We use our model to explore the interaction of these pathways. There are two end-member cases to consider; one with particles dominating the distribution and transport of Hg, and the other with circulation dominating. The particle-flux dominating end-member conditions can be achieved in Hg-HAMOCC by disabling the advection of the Hg tracers (Fig. 8, orange line). In the steady state, in order to achieve Hg concentrations that are not changing through time, the vertical flux of Hg through the water column must be the same at all depth levels. The flux of sinking POC decreases with depth in the ocean due to degradation. The abundant POC sinking flux in the surface ocean carries the same Hg sinking flux as the rarefied POC sinking flux in the deep sea.
This means that in the steady state, the POC in the deep sea has to carry
more Hg than it would in the surface ocean. The adsorbed Hg is linearly
related to the dissolved Hg by the adsorption Eq. (1). Rearranging Eq. (1)
gives the following:
If we take the sinking Hg-P flux to be proportional to [Hg-P] (assuming a uniform sinking velocity), then a decrease in the flux of POC (proportional to [POC] for the same reason) requires a higher dissolved [Hg(II)]. The result is that, in the steady state, Hg concentrations rise with depth in the ocean, to compensate for the decrease in sinking POC flux. A smaller POC sinking flux will have to carry a higher Hg concentration in order to sustain the required depth-uniform Hg flux, and the higher adsorbed Hg concentration requires a higher Hg concentration in the water column.
The other end-member case comes much closer to the observed distribution of Hg in the deep ocean. When circulation dominates, and particle transport of Hg is disabled, the Hg concentrations maintained in the surface ocean (by balancing evasion against deposition) are imposed on the deep ocean, resulting in a nearly uniform distribution of Hg throughout the ocean (Fig. 8, blue line). There are some regional variations in Hg in this scenario, but they are not systematic, as compared to the clear Pacific–Atlantic differences exhibited by nutrient-type elements (concentrated in the Pacific) versus those exhibited by strongly scavenged elements like Al (concentrated in the Atlantic, where deposition is more intense).
Profiles of mean Hg concentration in (preanthropogenic) steady
state, as a function of depth in the ocean, in equilibrium, for the
end-member cases of no particles (blue lines), and no advection (orange
lines).
The balance between advection versus sinking particles affects the uptake of anthropogenic Hg by the ocean. Profiles of total Hg changes from the preanthropogenic period to the present day, after 130 years of enhanced Hg deposition (to 2010), are shown in Fig. 3. If particles are neglected or sink so slowly as to be negligible in the Hg cycle, there is a sharp surface spike in Hg concentrations in the model simulation of the present day (2010), due to increased deposition. The increasing importance of particle transport tends to moderate a surface ocean spike, while transferring much of the anthropogenic Hg load to a subsurface maximum corresponding to the location of POC degradation in the thermocline. Particulate Hg transport to depth is required in order to simulate a subsurface maximum in Hg concentration, as observed in the present-day real ocean. In the steady state, with no anthropogenic enhanced deposition, a somewhat slower Hg sinking flux would still generate a subsurface maximum, but it is harder to have a subsurface maximum at the end of a period of enhanced Hg deposition, such as today.
Time series of the ocean load of Hg (Mmol), in response to 250 years of enhanced Hg(II) deposition (1850–2100), followed by abrupt return to natural Hg deposition rates, or 1000-year wind-down in anthropogenic deposition due to recycling from the ocean.
Figure 9 shows the total ocean inventory of anthropogenic Hg throughout the anthropogenic deposition period (ending in the year 2100) and beyond, as a function of the Hg particle sinking velocity. Particle transport has only a minor impact on the global rate of Hg uptake during the Anthropocene stage, but strong particle transport has the effect of sequestering the anthropogenic Hg deeper in the ocean (Fig. 3), where it is retained somewhat longer than in models with less particle transport. The model, when forced with an instantaneous end to anthropogenic emissions, predicts that the ocean will continue degassing Hg for 1000 years. When this prediction is turned around, to impose a condition that the Hg deposition rate declines over 1000 years after the year 2100, the duration of the anthropogenic Hg load on the oceans increases to several thousand years.
For each of eight kinetic rate constant parameters in the Hg system, we ran simulations to a natural steady state with factor-of-2 increases and decreases in each parameter in turn, as shown in Fig. 10. In general, increasing the rate constant for a given reaction will increase the concentration of the product and decrease that of the reactant. The other species' concentrations will also change in the new steady-state balance. The concentration of Hg overall depends on the rate of Hg removal from the system, primarily by gas evasion of the minor species Hg(0), with secondary sinks by DMHg evasion and Hg(II) adsorption onto sinking particles (Table 1). Increasing the rate of MMHg production from Hg(II), for example, decreases [Hg(II)] and increases [MMHg]. The concentration of DMHg increases due to its close coupling with MMHg (see Fig. 1).
Model sensitivity to kinetic rate constants in the Hg system. For each kinetic rate constant as indicated in the centered titles, global mean concentrations of each species are given in the four plots in that row, as indicated by the labels at the top of each column. Black lines represent the base case, and red and blue represent factors of 2 higher and lower for that kinetic rate constant, respectively.
Changes in the rate constants that produce or consume Hg(0) tend to result in larger changes in Hg concentrations than the rate constants for reactions that involve DMHg, because Hg(0) is responsible for a larger fraction of the gas evasion flux. The exception is reductive degradation of MMHg to Hg(0), which occurs primarily in the surface ocean, changing the MMHg concentration there without changing concentrations appreciably in the deep ocean. The highest model sensitivity in the suite of runs is to the rate of evasion of Hg(0), which drives large changes in the total Hg concentration of the entire ocean, in the steady state.
In general, the rates of chemical transformation of Hg are much faster than that of the ocean overturning circulation, so the distribution of Hg species at any location reflects a local balance between sources and sinks of each form of Hg. However, reactions at the sea surface that provide a pathway for Hg evasion into the atmosphere have the potential to alter the Hg concentrations throughout the ocean in the steady state.
Fluxes (in Mmol yr
Schematic of the expression of isotopic fractionations on the
global mean sea surface isotopic signatures of the Hg species in the model,
for preanthropogenic steady state. Fractionation epsilon values are shown in
red, expressed as per mill differences in the
Isotopic fractionations in the Hg cycle can be “expressed” in the isotopic signatures of Hg species, or not, depending on how the fractionating process fits into the network of reactions in the cycle. Figure 11 shows the isotopic compositions of Hg species resulting from a variety of fractionation mechanisms, in schematic diagrams of the ocean Hg cycle. Red numbers indicate isotopic fractionations and black numbers show global mean oceanic isotopic compositions. The results represent preanthropogenic steady state. The model is run to equilibrium for each of seven fractionation mechanisms in isolation, and finally for all mechanisms combined. For ease of comparison with oceanic measurements, all scenarios are subject to fractionation in Hg deposition, as indicated by the red numbers next to these fluxes. Figure 12 shows depth profiles of isotopic composition. Figure 13 shows maps of isotopic compositions, at the sea surface and at depth.
A guiding principle in understanding these results is that in the steady state the isotopic composition of the sinking fluxes have to balance the isotopic compositions of the inputs of Hg(II) from rain and Hg(0) from atmosphere–sea surface exchange. A fractionation mechanism that alters the isotopic signature of one of the sink fluxes will require the steady-state signatures of the other sink fluxes to change in compensation. Then the values of the other species (MMHg and Hg(II)) are pulled in various ways by their connections with the two potential gases Hg(0) and DMHg.
Profiles of
The schematic diagram in Fig. 11a shows the fractionation associated with
evasion of Hg(0) to the atmosphere. The lighter isotope reacts faster (as is
typical), leaving a dissolved Hg(0) pool that has residually higher
The expression (or not) of a fractionation in a specific reaction pathway in
the Hg cycle depends on the web of reactions between the species and the
mass balance constraints. For example, fractionation during the reduction
step from Hg(II) to Hg(0) (Fig. 11c and d) pulls the
Fractionation in the photochemical MMHg
Fractionation in MMHg production (Fig. 11g) results in a decrease in
Profiles of
Profiles of the
The
Maps of steady-state distribution of
The
We have embedded a model of Hg chemistry and dynamics into the HAMOCC offline ocean tracer advection model, including treatment of isotopic fractionation of Hg in the ocean Hg cycle. The efficiency of the model makes it possible to do numerous sensitivity experiments for testing hypotheses and developing intuition about this complex system: 55 simulations of over 10 kyr each are presented in this paper.
The model demonstrates that the Hg cycle in the ocean is closer to an advective end-member than to a system in which transport on sinking particles dominates. The interplay of advection by fluid flow and sinking of Hg adsorbed on sinking particles is illustrated by end-member cases in which one or the other dominates. In an advection-dominated case in which particle transport is disabled, the Hg concentration in steady state is relatively uniform with depth, displaying the same pattern as for salinity. In a particle-dominated scenario in which fluid advection of Hg is disabled, the concentration of Hg in steady state increases with depth, in proportion to the decrease in the POC sinking flux with depth (due to particle decomposition). This is because in the steady state in which Hg concentrations are not changing with time, the sinking flux of Hg through the ocean must be the same (on a horizontal average) at all depth levels.
A series of sensitivity runs with different Hg-P sinking velocities shows that the observed present-day subsurface maximum in Hg(II) is a product of Hg sinking on particles and the anthropogenic increase in Hg deposition to the surface ocean. Given the 4-fold enhanced Hg deposition flux since about 1850 (Streets et al., 2017), if there were no Hg sinking and subsurface release from particles, the highest Hg concentrations would be at the sea surface today. Anthropogenic Hg sinking on particles does not have a strong impact on the net uptake rate of anthropogenic Hg by the ocean, but if the enhanced rates of Hg deposition were suddenly to return to natural levels, a model with strong Hg sinking takes longer to shed its anthropogenic Hg burden. Since oceanic Hg evasion will be recycled and re-deposited, the ocean system seems poised to buffer the environmental Hg concentration for thousands of years.
We show the sensitivity of the steady-state (preanthropogenic, 1850) Hg species concentrations to eight kinetic rate constants in the aqueous Hg cycle. Allowing a reaction to proceed more quickly than a base case tends to result in more of the product and less of the reactant, but the magnitude of the change and the impact on the rest of the Hg species and the total Hg concentration vary widely between the various reactions. Changes to the budget of Hg(0), the evasion of which is the dominant loss mechanism for Hg in surface waters, have a strong impact on the rest of the Hg concentrations. Changes to reactions involved in the MMHg budget have a stronger impact on the Hg cycle than changes to DMHg sources or sinks, because MMHg is kinetically tied more closely to Hg(II).
Isotopic variations in Hg have multiple “dimensions” of fractionation,
with mass-dependent fractionation produced by most processes, and several
forms of mass-independent fractionation produced by photochemical reactions.
The Hg cycle in the ocean is complex enough that a model is required to
predict the “expression” of isotopic fractionations in processes, on the
isotopic signatures of Hg species in the ocean, and on the distribution of
variations in those signatures. There is wide variation in the expression of
isotope fractionation effects in the isotopic composition of Hg standing
stocks. In the model, surface–deep contrasts in
Fortran source code is available in a
repository at
The supplement related to this article is available online at:
DEA did the coding and plotting, both authors designed the study, analyzed the results, and wrote the paper.
The authors declare that they have no conflict of interest.
This article is part of the special issue “Progress in quantifying ocean biogeochemistry – in honour of Ernst Maier-Reimer”. It is not associated with a conference.
This work stands on the shoulders of Ernst Maier-Reimer who created the HAMOCC model. It also benefitted immensely from the constructive criticism of Jeroen Sonke and another anonymous reviewer. Edited by: Christoph Heinze Reviewed by: Jeroen Sonke and one anonymous referee