![Map of the North Sea, with the International Council for the Exploration of the Sea (ICES) statistical rectangles. The orange rectangle highlights the area considered here. Adapted from ICES, 2018 [59].](https://www.researchgate.net/publication/378933109/figure/fig1/AS:11431281263087743@1721905761496/Map-of-the-North-Sea-with-the-International-Council-for-the-Exploration-of-the-Sea_Q320.jpg)
Access to this full-text is provided by The Royal Society.
Content available from Royal Society Open Science
This content is subject to copyright.
An approach to assessing
subsea pipeline-associated
mercury release into the
North Sea and its potential
environmental and human
health impact
Rebecca von Hellfeld1,2 and Astley Hastings1,2
1School of Biological Sciences, University of Aberdeen, 23 St Machar Drive, Aberdeen AB24
3UL, UK
2National Decommissioning Centre, Main Street, Newburgh AB41 6AA, UK
RvH,0000-0003-4283-7813
Mercury is a naturally occurring heavy metal that has
also been associated with anthropogenic sources such as
cement production or hydrocarbon extraction. Mercury is
a contaminant of concern as it can have a signicant
negative impact on organismal health when ingested. In
aquatic environments, it bioaccumulates up the foodweb,
where it then has the potential to impact human health.
With the oshore hydrocarbon platforms in the North
Sea nearing decommissioning, they must be assessed as a
potential source for the environmental release of mercury.
International treaties govern the handling of materials
placed in the ocean. Studies have assessed the ecologic
and economic benets of (partial) in situ abandonment of
the infrastructure as articial reefs. This can be applied
to pipelines after substantial cleaning to remove mercury
accumulation from the inner surface. This work outlines
the application of an approach to modelling marine
mercury bioaccumulation for decommissioning scenarios in
the North Sea. Here, in situ decommissioning of cleaned
pipelines was unlikely to have a negative impact on
the North Sea food web or human health. However,
signicant knowledge gaps have been determined, which
must be addressed before all negative impacts on
ecosystems and organismal health can be excluded.
© 2024 The Authors Published by the Royal Society under the terms of the Creative
Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits
unrestricted use, provided the original author and source are credited.
Research
Cite this article: von Hellfeld R, Hastings A. 2024
An approach to assessing subsea pipeline-
associated mercury release into the North Sea and
its potential environmental and human health
impact. R. Soc. Open Sci. 11: 230943.
https://doi.org/10.1098/rsos.230943
Received: 1 July 2023
Accepted: 5 February 2024
Subject Category:
Earth and environmental science
Subject Areas:
ecosystem, environmental science, oceanography
Keywords:
contaminant modelling, ecopath with ecosim,
food standards for mercury, estimated weekly
intake, hazard quotient
Author for correspondence:
Rebecca von Hellfeld
e-mail: rebecca.vonhellfeld@abdn.ac.uk
Electronic supplementary material is available
online at https://doi.org/10.6084/
m9.gshare.c.7095868.
1. Introduction
The North Sea is a very biodiverse ecosystem [1], with strong primary producers and a stable benthic
community [2]. Furthermore, the North Sea is of global importance for seabirds and marine mammals
[1]. In addition to its high biodiversity, the North Sea is one of the world’s most important shing
grounds, with herring and mackerel being among the most commercially relevant species [3]. The
North Sea also holds a total of 590 oil and gas platforms [4], of which >470 platforms are expected to be
decommissioned over the coming decades in the UK sector alone. This includes around 5000 extraction
wells and ~45 000 km subsea pipelines [5]. This can be completed through a variety of methods ranging
from complete removal of all installed infrastructure to in situ abandonment of the cleaned pipeline
and jacket for articial reef creation (known as ‘rigs-to-reefs’) [6]. Such options have been implemented
in the Gulf of Mexico [7] and are being considered for the North Sea [8]. Here, the proposed decommis-
sioning plan should aim to minimize the impact on the marine environment, as well as limit human
risks in line with current regulations. These include the ‘United Nations Convention on the Law of the
Sea’ (UNCLOS) [9], the ‘Convention on the Prevention of Marine Pollution by dumping of wasted and
other materials 1972 (London Convention)’ and the London Protocol [10], as well as the Oslo and Paris
Conventions (OSPAR) decision 98/3 on the disposal of disused oshore installations [11]. In addition,
legislations such as the Marine Strategy Framework 46 Directive 2008/56/EU [12] and 2011/92/EU [13]
must be considered, which state that each ‘project’ must assess all direct and indirect impacts on biota,
the environment, material assets and cultural heritage. This provides guidance for considerations to
be taken in the assessments conducted in decommissioning plans, determining scenarios with minimal
impact of the planned operations on the ecosystem and surrounding environment, as well as limiting
human risk and economic costs [14].
It has been argued that the in situ abandonment of oshore infrastructure could lead to an increase
in biodiversity or biomass of the associated sh assemblage through the formation of articial reefs [8].
However, lile focus is currently placed on the eects of residual contaminants potentially associated
with the infrastructure, even after cleaning [15–17]. One contaminant of concern is mercury, a naturally
occurring heavy metal that is extracted alongside the hydrocarbon from oshore reservoirs [18,19].
Mercury is known to associate with the internal pipeline surfaces owing to rapid changes in tempera-
ture and pressure [20,21]. Mercury speciates based on surrounding parameters and the availability of
other compounds [22]. In most aquatic ecosystems, 95% of mercury is sediment-associated, bound to
organic maer and mineral phases [23]. Common insoluble mercury species found in marine ecosys-
tems include cinnabar (HgS), calomel (Hg2Cl2), mercuric oxide (HgO) and elemental mercury (Hg0).
Soluble mercury chloride complexes (HgCl2, HgCl3−, HgCl42+) and dissolved or particulate organic
maer complexes (DOM-Hg and POM-Hg, respectively) are also present [24]. Although it is known
that parameters such as temperature, pH, salinity, pore water sulphide, organic maer and sediment
redox potential inuence speciation [18], the complexity of the marine environment increases the
diculty in accurately determining the fate of mercury in the ocean and its accumulation in marine
biota [19]. Mercury also speciates in oil and gas production systems based on their physical and
chemical properties. The water-soluble Hg0 forms more stable complexes with chloride or sulphur
ions [25]. Studies showed that Hg0 adsorbs to the inner steel surface of pipelines [26–28]. Mercury
also adsorbs on corrosion products formed within the pipelines. Here, the speciation process is based
on environmental parameters, as well as the corrosion scale make-up [18]. Although sulphur-bound
species are initially unavailable for biota uptake, at low sulphide concentrations, such as those found in
oxic seawater, the dissolved fraction of mercury increases [29]. For a detailed analysis of the parameters
inuencing pipeline-bound mercury speciation, refer to [18]. While pipeline-cleaning methods such as
pigging and chemical cleaning exist [30,31], studies have estimated that even cleaned pipelines may
still contain up to 98 mg total mercury per kg pipeline (herein mg kg-1) [32]. The same study further
determined a sediment concentration of 260 µg total mercury per kg sediment (herein µg kg-1) at
the pipeline. This concentration decreased to 10 µg kg-1 at 125 m from the pipeline. In comparison,
the concentration given by OSPAR for pristine or remote sites is 50 µg kg-1, while the accepted
concentration assessment is 70 µg kg-1 for marine sediments [33]. However, considering that published
data for cleaning previously active subsea pipelines is scarce, and we currently lack the technology for
measuring in situ mercury contamination in pipelines, it cannot be excluded that parts of to-be-aban-
doned pipelines may hold higher mercury concentrations than laboratory studies indicated [18].
For organismal health, the most important speciation process is the methylation of mercury into
organic methylmercury, a known neurotoxicant [34]. While freshwater sediments are assumed to be
2
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 11: 230943
the main source of mercury methylation [35], processes driving marine methylation are less well
understood. Studies suggest that the formation here is driven by the mixed layer [36]. On average, 10%
of the total mercury in the water column is methylmercury. This increases to 15% in phytoplankton,
30% in zooplankton and 95% in shes and higher trophic-level organisms, owing to its bioaccumula-
tion and biomagnication potential [19]. Once ingested, methylmercury has a long retention time, thus
leading to its bioaccumulation within an organism, as well as its biomagnication in food webs [37].
Approximately 90% of ingested methylmercury is thought to be absorbed through the gastrointestinal
tract, as its strong anity for L-cysteine allows it to cross cell membranes and enter the blood stream
[38,39]. The L-type neutral amino acid carrier transport 1 system then enables its transport into the
central nervous system [40]. Methylmercury also covalently bonds with sulydryl group [41], which
leads to enzyme inhibition and may be the cause of its neurotoxicity [42]. To account for this toxicity, a
tolerable weekly intake (TWI) for methylmercury has been developed [43]. This governs the maximum
permissible body burden from dietary intake that would not likely induce negative health eects in
humans. The current TWI of 1.3 µg methylmercury per kg body weight [44] (herein referred to as
µg kg bw-1) was based on epidemiological studies and is considered protective of the developmental
neurotoxic eects of methylmercury ingestion in foetuses and children and should be followed by
women of childbearing age or those who are pregnant. However, special focus should be placed on
considering changes in dietary mercury exposure for future generations.
1.2. Aims and objectives
The research aims to apply a previously outlined method [45] for mercury food web bioaccumulation
modelling to the North Sea considering dierent exposure scenarios. An existing eEcopath with an
eEcosim model of the North Sea was used [46], with varied mercury inux into the model environ-
ment, based on dierent decommissioning scenarios for the North Sea. The modelled data were
validated against current mercury concentrations in dierent organisms. This approach allowed for the
determination of potential data gaps for mercury accumulation in the North Sea, as well as knowledge
gaps that require addressing to make this method suitable for environmental impact assessments in
the future. Furthermore, the economic implications were calculated in terms of revenue lost owing to
increased mercury accumulation beyond the current food standards (FS) for mercury. Moreover, the
estimated weekly intake of methylmercury (EWIMeHg) was calculated for men and women, in general,
as well as children and pregnant women, comparing the obtained EWIMeHg to the current TWI. In
addition, the hazard quotient (HQ) for these groups was derived to assess the overall risk posed to
each.
2. Material and methods
To beer understand bioaccumulation of mercury in the North Sea marine food web, a previously
outlined in silico food web modelling approach [45] for Ecopath with Ecosim (EwE, V.6 [47]) and its
contaminant tracking tool ‘Ecotracer’ [48] was used. The model was calibrated using literature-derived
data for environmental mercury concentrations in the North Sea, and biota uptake rates. The results
obtained in the present study were assessed against literature-derived mercury concentrations in biota
samples from the North Sea. A previously developed food web model of the North Sea [46] was used
to test the method, and dierent environmentally relevant exposure scenarios (ES) were determined.
Data conversions were conducted using Microsoft Excel®, and the data visualization was done using
the Tidyverse package for R or SigmaPlot (V.14.0, Systat Software Inc.) [49,50].
2.1. Food web modelling with Ecopath with Ecosim and Ecotracer
A detailed explanation of the underlying equations and functions of EwE can be found in Christensen
et al. [51]. Briey, EwE is a food web modelling programme based on determining the static, temporal
or spatiotemporal mass balance of a modelled system. The underlying calculations and input values
regarding parameters such as mortality, migration and predation, form the basis for the contaminant
tracking tool ‘Ecotracer’. Ecotracer simulates the transport of a chosen contaminant through the food
web, solving the contaminant dynamic equation simultaneously with the mass balance [52]. It allows
for a varied contaminant inux over time and considers dierent decomposition/outow pathways.
3
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 11: 230943
A detailed description of the tool equations and assumptions was recently published [48]. Ecotracer
has been used to track mercury and methylmercury in marine food webs in previous studies. This
includes a theoretical study highlighting the suitability of the programme suite for environmental risk
assessment for oshore decommissioning [45] and the impacts of climate change on the health risks to
the Faroe Island population owing to their reliance on marine mammal meat [53].
2.2. Model calibration
In the present study, an existing and validated EwE model of the North Sea was used [46], covering the
area outlined in gure 1. The model contains 29 functional groups, including the detritus, and a full
species list and the EwE input parameters can be found in the electronic supplementary material, S1.
To calibrate the model, an initial environmental concentration was derived from published data, and
separate model runs were initiated to incorporate the environmental mercury into the dierent species.
This was done following the method outlined in a recent publication on a hypothetical foodweb
[45]. Direct mercury uptake was only derived for photosynthetic plankton groups, copepods and
euphausiids, as the literature suggests most uptake for other marine organisms is diet-based [53–56];
see the electronic supplementary material, S2, for details. For the remaining functional groups, the
direct uptake of mercury from seawater is considered negligible compared with the dietary uptake [57]
and thus not computed. For this study, exposure to total mercury was modelled, using the uptake and
retention time parameters to simulate methylmercury biota uptake. To this end, an excretion rate of
10% was modelled. This accounts for the uptake of non-accumulating mercury species, as well as the
potential for some species to demethylate small amounts of mercury [58].
The initial environmental total mercury concentration used here is 0.57 µg l-1, based on the data
from the International Council for the Exploration of the Sea (ICES) ‘Contaminants and biological
eects of contaminants in water’ dataset [60]. This value is derived from a dataset encompassing
almost 300 000 data points from 1979 to 2021 (see hps://dome.ices.dk/views/ContaminantsBiota.aspx
for a detailed review of quantication methods and study characteristics). This value is higher
than other publications provided, as a large uncertainty pertains to values obtained pre-1958 owing
to changes in measurement method. However, for the present work, a conservative background
concentration was found suitable, as it aims to test a method and determine data gaps. Moreover, the
model validation outlined below highlights that, even with this conservative environmental concentra-
tion, the species modelled accumulate less mercury than was measured in comparable species in the
North Sea. A representation of the environmental mercury concentrations in the North Sea can be
found in the electronic supplementary material, S3.
2.3. Model validation at equilibrium
Before determining the changes in bioaccumulation of mercury, the environmental mercury concentra-
tion and direct uptake rates for the selected groups were input into Ecotracer and the system was
allowed to reach equilibrium. The resulting mercury burden for each species was converted into
muscle tissue concentration and compared to mercury concentrations measured in biota samples.
Based on literature ndings, approximately 50% of total mercury accumulates in muscle tissue for
most sh species [61], 15% for shrimp and 100% for benthos species [62–64], plankton [65,66] and
detritus. Such values can be compared with the FS for mercury in foodstus, determining that 0.5 and
1 mg kg-1 muscle tissue are guideline values for low and high trophic level species, respectively [67].
These values indicate whether landed sh is eligible for sale commercially.
The initial mercury concentrations for each functional group derived from EwE were compared
with literature-derived concentrations of North Sea species with the same trophic level. Only publica-
tions were considered where the species’ trophic level was discernible, muscle samples were analysed,
it was evident whether the concentration was provided as wet or dry weight fraction, only fresh
samples with an uninterrupted freezing were assessed and the work was published in English. To
validate the modelled values, the model/observed (M/O) ratio was derived as follows [68]:
(2.1)M/O ratio = M
O
where O is the modelled and O is the mean observed mercury concentration. The closer to 1 the
derived value is, the more closely the modelled value represents the observed value. The normalized
4
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 11: 230943
mean bias (NMB), where n is the number of empirical studies generating the observed mean mercury
concentration value [68]:
(2.2)NMB = ∑1
nM−O
∑1
nO.
If the NMB ≥ 0, the model overpredicts the observation by a factor equivalent to the NMB + 1. For
example, a derived factor of 1.2 shows an overestimation by the model of a factor 2.2, while a value
of −0.2 highlights an underprediction by a factor of 1.2, for example [69]. The results of the validation
can be found in the electronic supplementary material, S4. The implemented environmental mercury
concentration may exceed the current concentration measured in North Sea marine samples, but when
calibrating the model, current concentrations did not lead to comparable concentrations in modelled
organisms with biota samples analysed. This indicates the suitability of the selected value to the food
Figure 1. Map of the North Sea, with the International Council for the Exploration of the Sea (ICES) statistical rectangles. The orange
rectangle highlights the area considered here. Adapted from ICES, 2018 [59].
5
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 11: 230943
web model. Such discrepancies can be based on the limitations inherent in modelling studies, such as
those discussed in this work. This publication intends to highlight the method and its current data and
knowledge gaps, this deviation was accepted as suitable to the task.
2.4. Exposure scenarios
Once the model equilibrium was validated, four scenarios were derived, representing current
environmental mercury concentrations (ES A), in situ decommissioning of some pre-cleaned pipelines
based on current decommissioning proposals (ES B), and in situ decommissioning of all pipelines in the
North Sea after cleaning (ES C and D); see the electronic supplementary material, S5 for actual values
and details.
2.4.1. Exposure scenario A: current environmental concentrations
ES A represents the current environmental total mercury concentration of 0.57 µg l-1 [60], as outlined
above. To estimate the annual anthropogenic inux of total mercury into the North Sea, the ‘inputs
of mercury, cadmium and lead via water and air to the Greater North Sea between 1990 and 2014’
dataset [70], provided by the Centre for Environment, Fisheries and Aquaculture Science (CEFAS)
was used. The average inux from 2005 to 2014 of the CEFAS-provided dataset was used (0.01 t yr-1),
providing a more conservative input value than an average of the entire dataset would provide [70].
For conversions into the model environment, a mean depth of 95 m and an area of 570 000 km2 was
assumed, as these are the average depth and the overall area of the North Sea.
2.4.2. Exposure scenario B: 317 km in situ decommissioned cleaned pipeline
ES B represents a scenario where all 317.12 km of pipeline currently abandoned in the North Sea and
without a clear plan for removal [71] are accepted for in situ decommissioning after cleaning. From
the publicly accessible data, the average pipeline is a schedule 60 with a 10″ diameter (outer diameter:
27.31 cm, wall thickness: 12.7 mm). The recent decommissioning report states that a cleaned pipeline
is estimated to lead to a concentration of 260 µg kg-1 in the sediment directly surrounding the pipeline
[32]. To model the potential release of a decommissioned pipeline, it was assumed that a pipeline
released 260 µg kg-1, which equates to a total concentration of 0.07 t mercury in 317.12 km pipeline.
The environmental mercury concentration and annual inux, as well as initial species concentration as
stated in ES A were applied in ES B. No dierence in the nal accumulation of contaminants in the
model was observed when comparing single-release events versus setwise release (data not shown)
[45]. Thus, ES B assumes a single-release event after 50 years of 0.07 t yr-1, before returning to the
CEFAS database-derived annual anthropogenic inux.
2.4.3. Exposure scenario C: 45 000 km in situ decommissioned cleaned pipeline
ES C represents a release scenario where all 45 000 km pipelines currently in the North Sea are
decommissioned in situ post cleaning. The assumptions and values of ES A apply, and, similar to ES
B, at 50 years, an additional inux of 9.9 t mercury was added to the system, before returning to the
previously used 0.000004 t yr-1, this would equate to a total concentration of 9.9 t mercury in 45 000 km
cleaned pipeline.
2.4.4. Exposure scenario D: 45 000 km in situ decommissioned pipelines cleaned to smelting regulations
ES D represents a release scenario where all 45 000 km pipeline currently in the North Sea are decommis-
sioned in situ post cleaning. The assumptions and values of ES A apply, and the same calculations as for
ES B and C apply. However, here, the pipelines are cleaned to the maximum threshold given by many
smelting companies, which is 2 mg kg-1 steel. For 45 000 km pipeline, this would lead to an additional
inux of 72.93 t mercury after 50 years before returning to the previously used 0.000004 t yr-1.
2.5. Model assumptions
ES B and C represent a concentration series as all underlying assumptions remain the same as outlined
in ES A with the addition of the single release of additional mercury. While the above described
6
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 11: 230943
method is the most suitable option for the present study, incremental release modelling may be more
suitable for shorter timescale studies or instances where spatial modelling is considered. Further
limitations include that it was assumed that all mercury entering the marine environment is available
for uptake, and actual body burden data of North Sea species were used to compute uptake rates. This
represents a very conservative (i.e. ‘worst case scenario’) example that may not be applicable to every
ecosystem. Moreover, all applied simplications follow the method initially outlined in von Hellfeld
et al. [45]. To determine the partitioning of a contaminant between the water and sediment/particu-
late compartment, a partitioning coecient (Kd value) can be employed [72]. However, these values
must be derived for each environment and are highly dependent on the mercury species, as well as
parameters such as organic maer content. To this end, the assumptions in the previous work [45]
were accepted for the sake of testing the method for the North Sea. It was further assumed that no
sediment mixing or dilution occurs post pipeline release, limiting the dispersion potential for mercury
in the aqueous phase. A sediment compartment was not modelled. A closed model environment was
assumed, with no migration of biota or import/export of aqueous mercury, owing to a lack of data.
Although the model validation highlighted the t of the model to current environmental mercury
concentrations and accumulation, the change in marine organism muscle concentration in ES B and
C may not be representative of future scenarios. The results presented here merely outline the nal
mercury concentration based on the outlined assumptions and estimates. Considering that mercury
biomagnication will be highly sensitive to local environments and ecosystems, as well as to changes
in local parameters and characteristics, these results should not be interpreted as being predictive of
impacts from real mercury releases.
2.6. Estimated weekly intake of methylmercury and hazard quotient determination
To derive the EWIMeHg for humans from consuming seafood with total mercury concentrations
equivalent to those derived from the ES, previously described methods were followed [45]. Here, only
sh landed in ICES ecoregion codes IIIa, IV (a–c) and VIId and VIIe were considered as ‘North Sea’
[73]. This approach determined that 79% of sh and 77% of shellsh in the United Kingdom originated
from non-North Sea landings [74]. Moreover, all seafood imported into the UK was also considered for
the determination of the EWIMeHg. As not enough data for background mercury concentrations in sh
landed abroad or imported could be obtained, the concentration determined in ES A was applied to the
imported proportion. For seafood originating from the North Sea, muscle concentrations determined
through the model scenarios were applied.
To determine the EWIMeHg, the bioaccessibility of methylmercury during consumption (the fraction
of tissue-bound methylmercury that is released from the food matrix during consumption and in a
soluble form in the gastrointestinal tract, see the electronic supplementary material, S6) and uptake
rates of methylmercury by human stomach epithelial cells (79% [75]) were calculated by adapting the
following equation [45]:
(2.3)EWIMeHg =AC × F × WI × AB
W× Frac .
Using the mercury muscle tissue concentration (AC, µg kg-1), the fraction represents bioaccessible
methylmercury (F, %), the weekly sh or seafood intake (WI, kg), the methylmercury stomach
epithelial cell absorption rate (AB, %), the body weight (W, kg) and the fraction of consumed seafood
(Frac). The weekly seafood consumption was determined based on the National Health Service (NHS)
recommendations for seafood consumption [76], and the average weekly consumed seafood was given
by the National Diet and Nutrition Survey (NDNS) [77] (electronic supplementary material, S7).
To determine the impact of increased mercury concentration in sh, the weekly intake for each food
group was determined by dividing the respective weekly intake value by the number of species in the
group:
(2.4)foodgroup = ∑
F × W
ngroup
ngroup .
The results for all three seafood groups were then combined to provide the EWIMeHg for total weekly
seafood consumption.
The HQ for methylmercury (HQMeHg) was calculated to determine the ratio between the potential
exposure to a substance, and the level of exposure at which no adverse eect is anticipated [78]. An HQ
7
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 11: 230943
≤ 1 indicates that adverse eects are not likely to occur, while an HQ > 1 states whether, and by how
much, an exposure dose (ED) exceeds the reference concentration (RfC) for dietary exposure. The HQ
is calculated as outlined below [79]:
(2.5)HQMeHg =ED
RfC .
For this study, the RfCMeHg of 0.0001 mg kg-1 d-1 (or 0.0007 mg kg-1 week-1) as dened by the United
States Environmental Protection Agency (EPA) was applied [80].
3. Results
3.1. No exceedance of the food standards for mercury in seafood in any modelled scenario
The muscle tissue concentration in all species increased in line with an increase in inux of mercury
into the model environment (gure 2). In ES A, representing an unchanged inux of mercury into
the marine environment, none of the commercially relevant species were observed to exceed either of
the FS for mercury. None of the observed species accumulated mercury in amounts that would lead
to an exceedance of the FS for mercury in ES B to D. However, in ES D, mackerel and saithe were
modelled to contain muscle tissue mercury concentrations above 0.2 mg kg-1. None of the modelled
concentrations exceed the FS for mercury of 0.5 and 1 mg kg-1 for high and low trophic level organisms,
respectively.
3.2. No exceedance of the estimated weekly intake of methylmercury and limited increase of the
hazard quotient for all modelled scenarios
In none of the modelled scenarios did the EWIMeHg of total seafood consumption per week by any
consumer group under either of the consumption rates exceed the current TWI of 1.3 µg kg bw-1 [44]
(gure 3). Following the NHS dietary recommendations [76], all groups were modelled to take in more
methylmercury per week, than when following the actual current seafood consumption according to
the NDNS [77]. A similar trend was observed when determining the HQ for all consumer groups
(gure 4).
4. Discussion
Overall, the data presented in this study have shown that no direct risk from in situ decommissioning
of oil and gas pipelines in the North Sea is to be expected, under the outlined assumptions. Although
increased concentrations of mercury in muscle tissue were modelled, the burden was not found to
surpass the FS for mercury in seafood. Current North Sea biota mercury concentrations cited in the
ICES database [81], as well as in the Marine Scotland database [82] and peer-reviewed publications
[83,84], for example, highlighted that various species of interest exceed the FS for mercury. Although
the presented data for all scenarios determined that modelled sh species did not exceed the FS for
mercury (gure 2; electronic supplementary material, S9), the model validation process highlighted
that the M/O and NMB values derived for the model were conservative to highly conservative for
all but one species modelled (electronic supplementary material, S4). This suggests that although
approaching biota concentrations, the mercury accumulation in many modelled species was still
underestimated. Lastly, although increases in the EWIMeHg and the HQ were computed, none of the
respective threshold values were surpassed.
The modelled approach of the present work does not account for sub-lethal impacts on marine
organisms, which may occur at modelled concentrations below threshold values such as the FS for
mercury in seafood. To this end, the following discussion will provide a more detailed insight into the
current understanding of the ecological and organismal health impacts of mercury overall. Although
no immediate risk to human health was modelled, a deeper assessment of the potential implications
for future communities is required to beer understand the risk factors and implications. Considering
the model limitations and considerations that will be discussed, risk and impact assessments should
nonetheless be carried out for each new decommissioning plan.
8
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 11: 230943
4.1. Ecological and organismal impacts of sub-lethal mercury concentrations
Current approaches to food web contaminant accumulation rarely assess the deleterious eect such
concentrations may have on the organisms at or below the modelled concentration. It is known that
long-term exposure to sub-lethal concentrations of mercury can impact organismal health over time.
Such impacts include developmental alterations and impaired larval predator avoidance and prey-cap-
ture abilities in mummichog embryos (Fundulus heteroclitus) exposed to <0.1 mg l-1 methylmercury [85–
87]. The direct link between laboratory studies and environmental exposure is lacking. Publications
indicating the mercury content in wild-caught sh are available [88,89], but in such cases, behavioural
or developmental alterations are not recorded. Only a few publications examine whether environmen-
tal mercury exposure aects the development or behaviour of wild organisms [90,91]. Similarly, limited
data are available on mercury and methylmercury toxicity of marine species, a shortcoming recently
highlighted by Gissi et al. [19]. This highlights that although we are capable of modelling mercury
in marine ecosystems and understanding the lethality in wild animals, we still lack the translation of
what impact exposure to sub-lethal concentrations has outside of laboratory conditions. This is a vital
gap that requires addressing, as current threshold values such as the FS for mercury are tailored to
human health, rather than animal wellbeing.
Figure 2. Final muscle tissue total mercury concentration (mg kg-1) in commercially relevant modelled North Sea species (a–q)
(shown as circles) for the four exposure scenarios (A–D) as highlighted in the methodology. Total muscle concentrations are listed in
the electronic supplementary material, S9.
9
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 11: 230943
4.2. Human health assessment: the global context
In the present study, no model scenario led to an EWIMeHg exceeding the TWI of 1.3 µg kg bw-1
(gure 3). The TWI estimates the amount per unit body weight of a contaminant that can be ingested
over time without risk of adverse eects. For methylmercury, the TWI takes into consideration that
extended exposure will increase the health risks through its accumulating potential [44]. Contrary
to the data presented here, a recent study highlighted that ve of the assessed European countries
(Belgium, Ireland, Italy, Portugal and Spain) had higher subgroups of the population that were
potentially at risk of EWIMeHg > 1.3 µg kg bw-1 TWI (table 1). With the overall conservative M/O values
for the present study, an underestimation of the EWIMeHg is plausible. Although overall comparable,
future studies would benet from a unied methodology for the determination of methylmercury
intake in dierent groups of the population. Additionally, testing of both raw and processed samples
would provide more insight into the potential EWIMeHg for dierent processing methods and would
allow a beer understanding of a consumer group’s risk based on personal and regional preferences.
Such an approach was successfully employed by Bradey et al. [92]. They also highlighted that the thus
far accepted fraction of >95% of total mercury being methylmercury does not hold true for all aquatic
species. More in-depth sample analysis is necessary to avoid an overestimation of risk from lower
trophic level species (i.e. accumulating less mercury) in the future.
Like the EWIMeHg, no modelled scenario was determined to lead to HQ >1 through seafood
consumption (gure 4). The HQ is the ratio between the dietary exposure dose and the reference dose,
which is based on a compound-specic benchmark concentration and an uncertainty factor to account
for bioaccumulation and other toxic eects [78]. Although the derived HQ value does not provide a
quantitative assessment of the potential risk, the larger the value, the higher the dierence between
Figure 3. EWIMeHg(µg kg bw -1) for the total weekly seafood intake derived for NHS recommendation of seafood consumption (teal)
and the UK average weekly consumed seafood as given by the NDNS (coral) in the respective consumer group (gure labels); see the
electronic supplementary material, S10 for more information.
10
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 11: 230943
the exposure concentration and the reference concentration and thus the greater the level of concern
[79]. Various studies have assessed the HQ for methylmercury from seafood consumption (table 2). An
exceedance of a HQ > 1 by certain consumer groups was not always concurrent with an exceedance
of the accepted TWI [99]. Similarly, an exceedance of the FS for mercury by certain sh species was
not always indicative of an increased HQ by consumer groups. Such ndings highlight the need for a
beer understanding of risk factors to dierent consumer groups, as well as taking dietary preferences
and cultural factors into account.
4.3. The foodweb model: limitations and considerations
The modelled data presented here show that EwE can successfully be used to model mercury
accumulation in a North Sea food web. Only one species accumulated more mercury than current biota
samples indicate present in such species (electronic supplementary material, S4). This supports the
future use of easy-to-use programmes such as EwE for comprehensive impacts and risk assessments
for ecosystems. As modelling remains a simplication of real-world scenarios, limitations must be
considered. For the present study, these limitations include: (i) the bioaccumulative nature of mercury
in marine animals not being represented in EwE, (ii) the lack of knowledge of additional sources of
mercury, (iii) the insucient data on uptake rates and contaminant burden for model calibration, and
(iv) the speciation of mercury in marine ecosystems.
Figure 4. HQ for dierent consumer groups (as listed on the x-axis) and dietary intake (top labels for the UK average weekly seafood
consumption according to the NDNS, and the NHS weekly recommendations) for all four exposure scenarios (vertical labels on the
right). The red line indicates the HQ = 1 threshold; see the electronic supplementary material, S11 for more information.
11
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 11: 230943
When calibrating the model, an inated environmental mercury concentration was used to achieve
a biota accumulation that is representative of published data for dierent North Sea species. This may
be addressed by including factors like bioaccumulation and biomagnication in the future EwE-Eco-
tracer version. To account for the increase in mercury with each trophic level, a highly conservative
environmental mercury concentration was employed here. Such approaches can increase the data
uncertainty at higher concentrations and should thus be addressed prior to the implementation on a
larger scale. Such improvements could include the addition of factors to account for the biomagnica-
tion within the food web or dierent species. Although methods for the determination of bioaccumula-
tion factors for contaminants exist, and data are available for food webs, these are often standalone
[104] rather than an integrated variable in food web modelling.
The environmental mercury concentration in the North Sea and the annual inux were derived
from publicly accessible databases [60,70]. Additional inux sources may not have been monitored,
leading to an underestimation of the North Sea mercury inventory in the present study. Such sources
may include produced water (water from the formation, produced along with the oil and gas, which
can sometimes also contain the injection uid and condensation water [105]) from oshore installations
or cuings piles (accumulation of drill cuings that form during hydrocarbon drilling operations and
consist of drilling uid, subsurface rock debris and hydrocarbons [106]). Ongoing studies indicate
Table 1. Summary of EWIMeHg from seafood from dierent European and international studies in comparison to the present study.
Study, Region EWIMeHg (µg kg bw-1) Additional information Reference
Present study, UK > 0.5 Present study
FP7-ECsafeSEAFOOD-project,
EU
>TWI From hake, cod, seabream, sea bass and octopus [93]
‘CALIPSO study’, France 1.3–1.6 Tuna, cod, ling, sole and hake were the main
contributors to the methylmercury intake [94]
[44]
‘Seafood frequency
consumption
questionnaire’,
Mediterranean Sea
0.12–6.11 For women of reproductive age [95]
N/A, Spain 0.98–2.60 For vulnerable population groups (children,
pregnant women and women of childbearing
age)
[96]
N/A, Mallorquín swamp 5.3, 3.7 and 4.4 For children, women of childbearing age and the
rest of the population, respectively
[97]
BfM MEAL, Germany >TWI From the consumption of tuna, herring, pollock
and trout (smoked, canned in oil or brine, fried,
pickled or in sauce)
[98]
Table 2. Summary of EWIMeHg from seafood from dierent European and international studies.
Region Finding Reference
Zhoushan, China The HQ for children and adults exceeded 1 for total seafood and sh. This was
found even when the EWIMeHg did not exceed the TWI
[99]
Baja California Sur,
Mexico
Not all species with muscle concentrations exceeding the FS for mercury also
lead to a HQ > 1
[100]
Aracaju, Brazil Six out of 13 analysed species exceeded the 0.5 mg kg-1 and two species
exceeding the 1 mg kg-1 FS for mercury. Only two species induced a HQ > 1
[101]
Malaysia HQ Between 1.5 and 2.7 for urban and coastal rural women of childbearing
age, respectively
[102]
Jeddah, Saudi Arabia Only one of the 13 samples sh species exceeded the 0.5 mg kg-1 FS for
mercury. Seven species had a HQ > 1
[103]
12
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 11: 230943
that the bioavailable fraction of mercury from these piles is thought to be minimal. By contrast, more
recent studies measuring mercury in North Sea seawater samples found concentrations around 0.5 ng
l-1 [107], although others stated concentrations between 0.5 and 200 ng l -1 [108]. The additional inux
of mercury from oshore installations, while possible, is more likely to lead to concentration changes
in the close vicinity of the platforms rather than on an ocean basin scale. Such limitations may, in
part, be the cause for the need to use inated environmental concentrations for the model calibration
presented here. A further source of mercury to account for is decommissioning pipelines. Estimates
provided for post-cleaning mercury concentrations assume that there is a uniform distribution of
mercury along the length of the pipeline [32] and that the cleaning ecacy of 97% can be obtained
throughout [109]. Past studies have highlighted that mercury accumulation is likely to increase the
low points of pipelines, as well as areas of stagnation and changes in the ow regime, such as corners
[110]. It was determined that post-cleaning, 7.7% of the remaining mercury were stable mercury salts
and 92.3% were Hg0 [109]. Although leaching tests run with these samples indicated no increase in
seawater mercury concentration over 112 days [111], it is not known how steel corrosion and marine
microbes may inuence such behaviour. Additionally, various parameters are known to inuence the
speciation of mercury within the pipeline, as highlighted in Kho et al. [18], which may further inuence
the risk posed to the local food web. This work has highlighted the need to improve our understanding
of the actual mercury inventory in the North Sea and other ocean basins.
The previously discussed lack of data also meant that concentrations in some of the species
modelled here had to be approximated from species of the same trophic level, but not necessarily
with the same behaviour. This was the case for the plankton communities, as no data could be found
for North Sea species. Thus, the uptake rates calculated here, and in turn the bioaccumulation, of
mercury into the North Sea model was based on phytoplankton communities in the southern Baltic
Sea [66]. This may account for a large fraction of the discrepancy between the modelled and observed
mercury concentration in higher trophic-level species.
Lastly, a beer understanding of the mobility and bioavailability of dierent mercury species,
marine mercury methylation [19] and the mercury inventory in subsea infrastructure [15] is needed.
The chemical speciation may be modelled by, for example, PHREEQc [112], a programme for geo-
chemical calculations for one-dimensional transport, speciation and batch reactions. However, here
too, the output is limited by our knowledge of speciation behaviour and inuencing characters, as
well as the reliability of input data. In addition to chemical speciation modelling, biological specia-
tion (or methylation) needs to be taken into consideration. Past studies worked with concentration
values for methylmercury only, examining the potential impact of changing climates on the uptake
of methylmercury [53]. A study aiming to determine the methylation rate and potential of marine
ecosystems highlighted that while certain parameters can be considered globally important, this
biological speciation was a highly localized event and could not be predicted with current tools [113].
In addition, the inclusion of additional environmental compartments to represent the sediment and
microbial community would improve the modelling of mercury speciation and availability in EwE.
4.4. Further considerations
Government statistics indicate that 80% of the seafood consumed in the UK consists of cod, had-
dock, salmon, tuna and various prawn species [114]. Such preferences were not considered in the
present work, but future research would benet from a more detailed assessment of society’s seafood
consumption habits. This is especially relevant when considering contaminants such as methylmer-
cury, with the potential to bioaccumulate and biomagnify, as all preferred sh species fall within
trophic levels ≥4 and thus pose an increased risk of exceeding the FS for mercury, as well as the
EWIMeHg [115]. It should also be noted that most of the salmon consumed in the UK is farmed,
thus being fed with shmeal containing marine caught species of a lower commercial value such as
shrimp or sand eel. A study conducted on Canadian farmed versus wild-caught salmon determined
a negligible dierence in mercury concentration in both salmon types but did determine a higher
bioaccumulation factor for farmed individuals [116]. Although observing a similar trend in a larger
database study for mercury in seafood, it was noted that farmed sh were understudied and that large
discrepancies existed between the database estimated mean mercury concentration and the US Food
and Drug Administration mercury monitoring programme estimates for most seafood examined in the
study [117]. Future studies could also assess the impact of shmeal diets on mercury accumulation in
farmed sh in more detail to determine the potential health benets or risks. These insights, and the
bioaccumulative and biomagnifying nature of methylmercury, mean that a potential consequence of
13
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 11: 230943
a future increase in seafood contamination is unavoidable and may be a reassessment of the dietary
recommendations for seafood consumption. While reducing or eliminating seafood consumption may
initially address the risk of increased methylmercury exposure, this may lead to consequences for both
human health and the environment [118]. This highlights that there are many factors highlighting the
importance of monitoring contaminant levels in marine food webs.
5. Conclusion
With decommissioning being in its infancy, all current proposals and conducted projects are executed
in accordance with best practices and current scientic knowledge [119]. The present study, along with
previous works [15,16,19], has highlighted gaps in the understanding of how contaminants associated
with oshore infrastructures behave in the marine environment over time. With mercury having been
associated with oshore oil and gas installations, the resulting potential environmental release in the
case of in situ decommissioning could be a serious threat to the ecosystem and human health [15,16,19].
We determined data gaps in the mercury inventory for the North Sea, which may also aect other
oceans. A global eort to address such gaps, for all contaminants of concern to the marine environ-
ment, would signicantly improve our understanding of the potential risks of oshore activities on the
marine food web and, by extension, human health and the relevant industries. Additionally, research
should focus on determining the bioavailable fraction of methylmercury in seafood and the stomach
uptake, thus allowing for more accurate estimates and future risk predictions.
This study aimed to apply a previously developed method for mercury bioaccumulation modelling
to the North Sea marine ecosystem in order to assess potential future economic and human health risks
of mercury released into the North Sea. All ndings indicate that although an increase in accumula-
ted mercury could be assumed, no further human health implications are expected under present
assumptions. Uncertainties remain in the current data on both environmental mercury inventory and
inux, as well as the mercury behaviour in the marine environment and its uptake by and speciation
within marine organisms. Bridging these gaps would provide a great benet for both the industry
and the scientic community. It would allow for a more robust risk assessment to be conducted for
oshore decommissioning practices, ensuring that decisions are made on the most holistic assessment
of each asset. Moreover, it would provide a basis for the assessment of other contaminants, as well
as the cumulative impact of various contaminants, greatly improving our ability to conserve the
world oceans and oer more targeted mitigation solutions for marine pollution. The development of
a global database for contaminant concentrations in marine species would greatly improve the ability
to determine potential economic implications for the future of the sheries industry. Moreover, the
human risk determination method presented here would benet from being tested on larger datasets
with more detailed insights into consumer habits, as well as mercury organ/tissue distribution and
speciation in marine organisms, species-specic assimilation eciency and the uptake availability and
ecacy of methylmercury by humans from dierent sh species and food preparation methods [92].
Ethics. This work did not require ethical approval from a human subject or animal welfare commiee.
Data accessibility. All used data has been included in supplementary materials and is freely accessible with the
publication [120].
Declaration of AI use. We have not used AI-assisted technologies in creating this article.
Authors’ contributions. R.v.H.: conceptualization, data curation, formal analysis, funding acquisition, investigation,
methodology, project administration, resources, software, supervision, validation, visualization, writing—original
draft, writing—review and editing; A.H.: funding acquisition, supervision, validation, writing—review and editing.
Both authors gave nal approval for publication and agreed to be held accountable for the work performed
therein.
Conict of interest declaration. We declare we have no competing interests.
Funding. R.v.H. and A.H. are funded by the Net Zero Technology Centre and the University of Aberdeen, through
their partnership with the UK National Decommissioning Centre. R.v.H. received funding from the University of
Aberdeen under the interdisciplinary project funding in 2022.The research by A.H. was also undertaken as part of
the UK Energy Research Centre research programme (UKERC-4, EP/S029575/1).
Acknowledgements. This publication contains data obtained from dierent research councils. The authors
acknowledge the the contributions of the following councils and respective works: The International Council for the
Exploration of the Sea (ICES) 'Contaminants dataset' [80]; The Centre for Environment, Fisheries and Aquaculture
Science (CEFAS) 'Inputs of mercury, cadmium and lead via water and air to the Greater North Sea between 1990 and
2014' [70] and 'Time trends and status for cadmium, mercury and lead in sh and shellsh' [82]; and The North Sea
14
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 11: 230943
Transition Authority (NSTA) 'SDC Oshore Infrastructure dataset' [71]. Through this, the authors acknowledge all
institutions that have contributed to the creation of the datasets used in the present publication.
References
1. OSPAR. 2000 Quality status report 2000 region II - greater North Sea. London, UK: OSPAR Commission.
2. Knijn RJ, Boon TW, Heessen HJL, Hislop JRG.1993 Atlas of Nor th Sea shes. based on bottom-Trawl survey data for the years 1985-1987. Report
No.: ICES Cooperative Research Report 194.
3. Walday M, Kroglung T. 2008 Europes Biodiversity - Biogeographical regions and seas. the North sea - bottom Trawling and oil/gas exploitation
European Environment Agency. See https://www.eea.europa.eu/publications/report_2002_0524_154909/regional-seas-around-europe/
NorthSea.pdf/view
4. Martins MCI, Carter MI, Rouse S, Russell DJ. 2023 Oshore energy structures in the North Sea: past, present and future. Mar. Policy 152, 105629.
(doi:10.1016/j.marpol.2023.105629)
5. Mackenzie W. 2020 UK North sea decommissioning: the £17 billion challenge. See https://www.woodmac.com/news/opinion/uk-north-sea-
decommissioning-the-17-billion-challenge/ (accessed 10 November 2022).
6. Bull AS, Love MS. 2019 Worldwide oil and gas platform decommissioning: a review of practices and reeng options. Ocean Coast. Manag. 168,
274–306. (doi:10.1016/j.ocecoaman.2018.10.024)
7. Schroeder DM, Love MS. 2004 Ecological and political issues surrounding decommissioning of oshore oil facilities in the Southern California
Bight. Ocean Coast. Manag. 47, 21–48. (doi:10.1016/j.ocecoaman.2004.03.002)
8. Fowler AM, et al. 2018 Environmental benets of leaving oshore infrastructure in the ocean. Front Ecol. Environ. 16, 571–578. (doi:10.1002/
fee.1827)
9. United Nations. 1982 United Nations convention on the law of the sea. Treaty Series. See https://www.un.org/depts/los/convention_
agreements/texts/unclos/unclos_e.pdf
10. IMO. 1996 London Convention: protocol of the convention on the prevention of marine pollution by dumping of wastes and other matter. See
https://www.epa.gov/sites/production/les/2015-10/documents/lpamended2006.pdf
11. OSPAR. 1998 OSPAR decision 98/3 in the disposal of disused oshore installations. See https://cil.nus.edu.sg/wp-content/uploads/formidable/
18/1998-OSPAR-Decision-98-3.pdf
12. EU. 2008 Directive 2008/56/EC of the European Parliament and of the council on 17 June 2008 establishing a framework for community action
in the eld of marine environmental policy (Marine Strategy Framework Directive). See https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?
uri=CELEX:32008L0056&from=EN
13. EC. 2012 Directive 2011/92/EU of the European parliament and the council of 13 December 2011 on the assessment of the eects of certain
publica and private projects on the environment. See https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32011L0092&from=EN
14. Shaw ED. 2018 Skipping ahead to the good part: the role of civic technology in achieving the promise of E-government. J. eDEM. 10, 74–96.
(doi:10.29379/jedem.v10i1.455)
15. Kho F, Koppel DJ, von Hellfeld R, Hastings A, Gissi F, Cresswell T, Higgins S. 2022 Current understanding of the ecological risk of mercury from
subsea oil and gas infrastructure to marine ecosystems. J. Hazard Mater. 438, 129348. (doi:10.1016/j.jhazmat.2022.129348)
16. Koppel DJ, Kho F, Hastings A, Crouch D, MacIntosh A, Cresswell T, Higgins S. 2022 Current understanding and research needs for ecological risk
assessments of naturally occurring radioactive materials (NORM) in subsea oil and gas pipelines. J. Environ. Radioact. 241, 106774. (doi:10.
1016/j.jenvrad.2021.106774)
17. MacIntosh A, Daorn K, Penrose B, Chariton A, Cresswell T. 2022 Ecotoxicological eects of decommissioning oshore petroleum infrastructure:
a systematic review. Crit. Rev. Environ. Sci. Technol. 52, 3283–3321. (doi:10.1080/10643389.2021.1917949)
18. Kho F, Koppel DJ, von Hellfeld R, Hastings A, Gissi F, Cresswell T, Higgins S. 2022 Current understanding of the ecological risk of mercury from
subsea oil and gas infrastructure to marine ecosystems. J. Hazard Mater. 438, 129348. (doi:10.1016/j.jhazmat.2022.129348)
19. Gissi F, Koppel D, Boyd A, Kho F, von Hellfeld R, Higgins S, Apte S, Cresswell T. 2022 A review of the potential risks associated with mercury in
subsea oil and gas pipelines in Australia. Environ. Chem. 19, 210–227. (doi:10.1071/EN22048)
20. Wilhelm SM, Bloom N. 2000 Mercury in petroleum. Fuel Process Technol. 63, 2000–2001. (doi:10.1016/S0378-3820(99)00068-5)
21. Ryzhov VV, Mashyanov NR, Ozerova NA, Pogarev SE. 2003 Regular variations of the mercury concentration in natural gas. Sci. Total Environ. 304,
145–152. (doi:10.1016/S0048-9697(02)00564-8)
22. Mason RP, Fitzgerald WF. 1991 Mercury speciation in open ocean waters. Water Air Soil Pollut. 56, 779–789. (doi:10.1007/BF00342316)
23. Craig PJ (ed). 2003 Organometallic compounds in the environment. New York, NY: Wiley. (doi:10.1002/0470867868)
24. Gworek B, Bemowska-Kałabun O, Kijeńska M, Wrzosek-Jakubowska J. 2016 Mercury in marine and oceanic waters-a review. Water Air Soil
Pollut. 227, 371–390. (doi:10.1007/s11270-016-3060-3)
25. Wilhelm SM, Nelson M. 2010 Interaction of elemental mercury with steel surfaces. The Journal of Corrosion Science and Engineering 13.
26. Chaiyasit N, Kositanont C, Yeh S, Gallup D, Young L. 2009 Decontamination of mercury contaminated steel of API 5L-X52 using iodine and iodide
lexiviant. MAS. 4. (doi:10.5539/mas.v4n1p12)
15
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 11: 230943
27. Roseborough D, Aune RE, Seetharaman S, Göthelid M. 2006 The surface behavior of mercury on iron systems. Metall. Mater. Trans. B 37, 1049–
1056. (doi:10.1007/BF02735027)
28. Linderoth S, Morup S. 1992 Stability and magnetic properties of an iron-mercury alloy. J. Phys. Condens. Matter. 4, 8627–8634. (doi:10.1088/
0953-8984/4/44/024)
29. Morse JW, Luther GW. 1999 Chemical inuences on trace metal-sulde interactions in anoxic sediments. Geochim. Cosmochim. Acta. 63, 3373–
3378. (doi:10.1016/S0016-7037(99)00258-6)
30. Chanvanichskul C, Punpruk S, Silakorn P, Thammawong C, Pornnimitthum S, Kumseranee S. 2017 In situ mercury decontamination for pipeline
decommissioning in the Gulf of Thailand. In Abu Dhabi International Petroleum Exhibition & Conference. https://onepetro.org/SPEADIP/
proceedings/17ADIP/3-17ADIP/Abu Dhabi, UAE/193635
31. Baker S, Andrew M, Kirby M, Bower M, Walls D, Hunter L, Stewart A. 2021 Mercury contamination of process and pipeline infrastructure - a
novel, all- encompassing solution for the evaluation and decontamination of mercury from pipelines and topside process equipment to allow
safe disposal. In SPE Symposium. https://onepetro.org/SPESM02/proceedings/22SM02/1-22SM02/D011S001R001/472794 (accessed 30
November 2021).
32. Advisian. 2022 Grin gas export pipeline decommissioning environment plan. See https://docs.nopsema.gov.au/A834478
33. OSPAR. 2013 Levels and trends in marine contaminants and their biological eects—CEMP assessment report 2013. Monitoring and assessment
Series. See http://dome.ices.dk/osparmime2018/main.html and https://ocean.ices.dk/oat/ (accessed 2014).
34. Benoit JM, Gilmour CC, Mason RP, Heyes A. 1999 Sulde controls on mercury speciation and bioavailability to methylating bacteria in sediment
pore waters. Environ. Sci. Technol. 33, 951–957. (doi:10.1021/es9808200)
35. Regnell O, Watras CJ. 2019 Microbial mercury methylation in aquatic environments: a critical review of published eld and laboratory studies.
Environ. Sci. Technol. 53, 4–19. (doi:10.1021/acs.est.8b02709)
36. Hammerschmidt CR , Bowman KL. 2012 Vertical methylmercury distribution in the subtropical North Pacic Ocean. Mar. Chem. 132–133, 77–
82. (doi:10.1016/j.marchem.2012.02.005)
37. Hammerschmidt CR , Fitzgerald WF. 2006 Bioaccumulation and trophic transfer of methylmercury in Long Island Sound. Arch. Environ. Contam.
Toxicol. 51, 416–424. (doi:10.1007/s00244-005-0265-7)
38. Aschner M, Clarkson TW. 1987 Mercury 203 distribution in pregnant and nonpregnant rats following systemic infusions with thiol-containing
amino acids. Teratology 36, 321–328. (doi:10.1002/tera.1420360308)
39. Clarkson TW. 2002 The three modern faces of mercury. Environ. Health Perspect. 110 (Suppl. 1), 11–23. (doi:10.1289/ehp.02110s111)
40. Kerper LE, Ballatori N, Clarkson TW. 1992 Methylmercury transport across the blood-brain barrier by an amino acid carrier. A m. J. Physiol-Regul.
Integr. Compar. Physiol. 262, R761–R765. (doi:10.1152/ajpregu.1992.262.5.R761)
41. Hastings FL, Lucier GW, Kleini R. 1975 Methylmercury: recent advances in the understanding of its neurotoxicity. Environ. Health Perspect. 12,
127–130. (doi:10.1289/ehp.7512127)
42. Aschner M, Sy versen T. 2005 Methylmercury: recent advances in the understanding of its neurotoxicity. Ther. Drug Monit. 27, 278–283. (doi:10.
1097/01.ftd.0000160275.85450.32)
43. Farina M, Rocha JBT, Aschner M. 2011 Mechanisms of methylmercury-induced neurotoxicity: evidence from experimental studies. Life Sci. 89,
555–563. (doi:10.1016/j.lfs.2011.05.019)
44. EFSA. 2012 Scientic opinion on the risk for public health related to the presence of mercury and methylmercury in food. EFSA J. 10, 241. (doi:.
org/10.2903/j.efsa.2012.2985)
45. von Hellfeld R, Gade C, Koppel DJ, Walters WJ, Kho F, Hastings A. 2023 An approach to assess potential environmental mercury release, food
web bioaccumulation, and human dietary methylmercury uptake from decommissioning oshore oil and gas infrastructure. J. Hazard. Mater.
452, 131298. (doi:10.1016/j.jhazmat.2023.131298)
46. Christensen V. 1995 A model of trophic interactions in the North Sea in 1981, the year of the stomach. Dana. 11, 1–19.
47. Christensen V, Walters CJ, Pauly D. 2005 Ecopath with ecosim: a user’s manual. See www.ecopath.org
48. Booth S, Steenbeek J, Charmasson S. 2020 Ecotracer: a user’s guide to tracking contaminants using the Ecopath with Ecosim (EwE) approach.
Figshare Media (doi:10.6084/m9.gshare.12821333.v2)
49. Wickham H, etal. 2019 Welcome to the Tidyverse. J. Open Source Soft. 4, 1686. (doi:10.21105/joss.01686)
50. R Core Team. 2019 R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing.
51. Christensen V, Walters CJ. 2004 Ecopath with Ecosim: methods, capabilities and limitations. Ecol. Modell. 172, 109–139. (doi:10.1016/j.
ecolmodel.2003.09.003)
52. Walters WJ, Christensen V. 2018 Ecotracer: analyzing concentration of contaminants and radioisotopes in an aquatic spatial-dynamic food web
model. J. Environ. Radioactiv. 181, 118–127. (doi:10.1016/j.jenvrad.2017.11.008)
53. Booth S, Zeller D. 2005 Mercury, food webs, and marine mammals: implications of diet and climate change for human health. Environ. Health
Perspect. 113, 521–526. (doi:10.1289/ehp.7603)
54. Canli M, Furness RW. 1995 Mercury and cadmium uptake from seawater and from food by the Norway lobster Nephrops norvegicus. Environ.
Toxicol. Chem. 14, 819–828. (doi:10.1002/etc.5620140512)
55. Mason RP, Reinfelder JR, Morel FMM. 1996 Uptake, toxicity, and trophic transfer of mercury in a coastal diatom. Environ. Sci. Technol. 30, 1835–
1845. (doi:10.1021/es950373d)
56. Downs SG, MacLeod CL, Lester JN. 1998 Mercury in precipitation and its relation to bioaccumulation in sh: a literature review. Water Air Soil
Pollut. 108, 149–187. (doi:10.1023/A:1005023916816)
16
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 11: 230943
57. Booth S, Zeller D. 2005 Mercury, food webs, and marine mammals: implications of diet and climate change for human health. Environ. Health
Perspect. 113, 521–526. (doi:10.1289/ehp.7603)
58. Wang X, Wang WX. 2017 Selenium induces the demethylation of mercury in marine sh. Environ. Pollut. 231, 1543–1551. (doi:10.1016/j.
envpol.2017.09.014)
59. ICES. 2018 DATRAS Specication Document—Areas in DATRAS Products. See https://www.ices.dk/data/Documents/DATRAS/Survey_Maps_
Datras.pdf
60. ICES. 2021 Contaminants and biological eects. Copenhagen, Denmark: ICES.
61. Wang R, Wong MH, Wang WX. 2010 Mercury exposure in the freshwater tilapia Oreochromis niloticus. Environ. Pollut. 158, 2694–2701. (doi:10.
1016/j.envpol.2010.04.019)
62. Taylor DL, Calabrese NM. 2018 Mercury content of blue crabs (Callinectes sapidus) from southern New England coastal habitats: contamination
in an emergent shery and risks to human consumers. Mar. Pollut. Bull. 126, 166–178. (doi:10.1016/j.marpolbul.2017.10.089)
63. Roveta C, Pica D, Calcinai B, Girolametti F, Truzzi C, Illuminati S, Annibaldi A, Puce S. Hg levels in marine porifera of montecristo and Giglio
Islands (Tuscan Archipelago, Italy). App. Sci. 10, 4342. (doi:10.3390/app10124342)
64. Rivera-Hernández JR, Fernández B, Santos-Echeandia J, Garrido S, Morante M, Santos P , Albentosa M. 2019 Biodynamics of mercury in mussel
tissues as a function of exposure pathway: natural vs microplastic routes. Sci. Total Environ. 674, 412–423. (doi:10.1016/j.scitotenv.2019.04.
175)
65. Bełdowska M, Kobos J. 2016 Mercury concentration in phytoplankton in response to warming of an autumn - winter season. Environ. Pollut.
215, 38–47. (doi:10.1016/j.envpol.2016.05.002)
66. Bełdowska M, Kobos J. 2018 The variability of Hg concentration and composition of marine phytoplankton. Environ. Sci. Pollut. Res. Int. 25,
30366–30374. (doi:10.1007/s11356-018-2948-4)
67. EC. 2006 Commission regulation (EC) no. 1881/2006 of 19 December 2006 setting maximum levels for certain contaminants in foodstus. See
https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32006R1881&from=EN
68. Li ML, Gillies E J, Briner R, Hoover CA, Sora KJ, Loseto LL, Walters WJ, Cheung WWL, Giang A. 2022 Investigating the dynamics of methylmercury
bioaccumulation in the Beaufort Sea shelf food web: a modeling perspective. Environ. Sci. Process Impacts 24, 1010–1025. (doi:10.1039/
d2em00108j)
69. Yu S, Eder B, Dennis R, Chu SH , Schwartz SE. 2006 New unbiased symmetric metrics for evaluation of air quality models. Atmos. Sci. Lett. 7, 26–
34. (doi:10.1002/asl.125)
70. Moxon R. 2019 Inputs of mercury, cadmium and lead via water and air to the Greater North Sea between 1990 and 2014. UK: CEFAS.
71. NSTA. 2022 SDC oshore infrastructure pilot (WGS84). See https://experience.arcgis.com/experience/eab843697e0a40d6991b5443c01799e0/
page/Home-Page/?views=Data
72. International Atomic Energy Agency (IAEA). 2004 Sediment distribution coecients and concentration factors for biota in the marine environment.
See https://www-pub.iaea.org/MTCD/Publications/PDF/TRS422_web.pdf
73. ICES. 2020 Denition and rationale for ICES ecoregions. See https://doi.org/10.17895/ices.advice.6014 (accessed 25 October 2022).
74. MMO. 2019 UK sea sheries statistics 2019. See https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_
data/le/920679/UK_Sea_Fisheries_Statistics_2019_-_access_checked-002.pdf (accessed 25 October 2022).
75. Bradley MA, Barst BD, Basu N. 2017 A review of mercury bioavailability in humans and sh. Int. J. Environ. Res. Pub. Health 14, 169. (doi:10.
3390/ijerph14020169)
76. NHS. 2018 Fish and shellsh. See https://www.nhs.uk/live-well/eat-well/food-types/sh-and-shellsh-nutrition/
77. PHE, FSA . National Diet and Nutrition Survey Results from Years 1, 2, 3 and 4 (combined) of the Rolling Programme (2008/2009 – 2011/2012).
See https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/le/594361/NDNS_Y1_to_4_UK_
report_full_text_revised_February_2017.pdf (accessed 29 March 2023).
78. Goumenou M, Tsatsakis A. 2019 Proposing new approaches for the risk characterisation of single chemicals and chemical mixtures: the source
related hazard quotient (HQS) and hazard index (HIS) and the adversity specic hazard index (HIA). Toxicol. Rep. 6, 632–636. (doi:10.1016/j.
toxrep.2019.06.010)
79. US EPA. 2005 Characterizing risk and hazard. Human health risk assessment protocol. See https://archive.epa.gov/epawaste/hazard/tsd/td/
web/pdf/05hhrap7.pdf
80. US EPA. 2001 Methylmercury (MeHg). CASRN 22967-92-2. See https://iris.epa.gov/static/pdfs/0073_summary.pdf
81. ICES. 2022 CEMP Assessment. ICES. See https://gis.ices.dk/geonetwork/srv/metadata/a5058fef-19fb-4ce9-8552-1b74e9199b9d
82. Nicolaus EEM, Lyons B, Miles A, Robinson CD, Webster L, Fryer R. 2018 Time trend and status for cadmium, mercury and lead in sh and
shellsh. See https://moat.cefas.co.uk/pressures-from-human-activities/contaminants/metals-in-biota/
83. Baeyens W, Leermakers M, Papina T, Saprykin A, Brion N, Noyen J, De Gieter M, Elskens M, Goeyens L. 2003 Bioconcentration and
biomagnication of mercury and methylmercury in North Sea and Scheldt estuary sh. Arch. Environ. Contam. Toxicol. 45, 498–508. (doi:10.
1007/s00244-003-2136-4)
84. Bełdowska M , Falkowska L. 2016 Mercury in marine sh, mammals, seabirds, and human hair in the coastal zone of the southern Baltic. Water
Air Soil Pollut. 227, 52. (doi:10.1007/s11270-015-2735-5)
85. Weis JS, Weis P. 1995 Eects of embryonic exposure to methylmercury on larval prey-capture ability in the mummichog, Fundulus heteroclitus.
Environ. Toxicol. Chem. 14, 153. (doi:10.1897/1552-8618(1995)14[153:EOEETM]2.0.CO;2)
17
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 11: 230943
86. Weis JS, Weis P . 1995 Swimming performance and predator avoidance by mummichog (Fundulus heteroclitus) larvae after embryonic or larval
exposure to methylmercury. Can. J. Fish. Aquat. Sci. 52, 2168–2173. (doi:10.1139/f95-809)
87. Weis JS, Weis P. 1977 Eects of heavy metals on development of the killish, Fundulus heteroclitus. J. Fish Biol. 11, 49–54. (doi:10.1111/j.1095-
8649.1977.tb04097.x)
88. Calatayud M, D evesa V, Virseda JR, Barberá R , Montoro R, Vélez D. 2012 Mercury and selenium in sh and shellsh: occurrence, bioaccessibility
and uptake by Caco-2 cells. Food Chem. Toxicol. 50, 2696–2702. (doi:10.1016/j.fct.2012.05.028)
89. FDA. 1990 Mercury levels in commercial sh and shellsh. See https://www.fda.gov/food/environmental-contaminants-food/mercury-levels-
commercial-sh-and-shellsh-1990-2012
90. Alvarez M del C, Murphy CA, Rose KA, McCarthy ID, Fuiman LA. 2006 Maternal body burdens of methylmercury impair survival skills of ospring
in Atlantic croaker (Micropogonias undulatus). Aquat. Toxicol. 80, 329–337. (doi:10.1016/j.aquatox.2006.09.010)
91. Smith GM , Weis JS. 1997 Predator-prey relationships in mummichogs (Fundulus heteroclitus (L.)): eects of living in a polluted environment. J.
Exp. Mar. Biol. Ecol. 209, 75–87. (doi:10.1016/S0022-0981(96)02590-7)
92. Bradley MA, Barst BD, Basu N. 2017 A review of mercury bioavailability in humans and sh. Int. J. Environ. Res. Pub. Health 14, 169. (doi:10.
3390/ijerph14020169)
93. Jacobs S, Sioen I, Jacxsens L, Domingo JL, Sloth JJ, Marques A, Verbeke W. 2017 Risk assessment of methylmercury in ve European countries
considering the national seafood consumption patterns. Food Chem. Toxicol. 104, 26–34. (doi:10.1016/j.fct.2016.10.026)
94. Sirot V, Guérin T, Mauras Y, Garraud H, Volatier JL, Leblanc JC. 2008 Methylmercury exposure assessment using dietary and biomarker data
among frequent seafood consumers in France CALIPSO study. Environ. Res. 107, 30–38. (doi:10.1016/j.envres.2007.12.005)
95. Dellatte E, et al. 2014 Individual methylmercury intake estimates from local seafood of the Mediterranean Sea, in Italy. Regul. Toxicol.
Pharmacol. 69, 105–112. (doi:10.1016/j.yrtph.2014.03.002)
96. Ortega-García JA et al. 2009 Estimated intake levels of methylmercury in children, childbearing age and pregnant women in a Mediterranean
region, Murcia, Spain. Eur. J. Pediatr. 168, 1075–1080. (doi:10.1007/s00431-008-0890-z)
97. Fuentes-Gandara F, Pinedo-Hernández J, Marrugo-Negrete J, Díez S. 2018 Human health impacts of exposure to metals through extreme
consumption of sh from the Colombian Caribbean Sea. Environ. Geochem. Health 40, 229–242. (doi:10.1007/s10653-016-9896-z)
98. Sar van I, Bürgelt M, Lindtner O, Greiner M. 2017 Expositionsschätzung von Stoen in Lebensmitteln. Bundesgesundheitsbl. 60, 689–696. (doi:
10.1007/s00103-017-2566-1)
99. Yu X, Khan S, Khan A, Tang Y, Nunes LM, Yan J, Ye X, Li G. 2020 Methyl mercury concentrations in seafood collected from Zhoushan Islands,
Zhejiang, China, and their potential health risk for the shing community: capsule: methyl mercury in seafood causes potential health risk.
Environ. Int. 137, 105420. (doi:10.1016/j.envint.2019.105420)
100. Murillo-Cisneros DA, Zenteno-Savín T, Harley J, Cyr A, Hernández-Almaraz P, Gaxiola-Robles R, Galván-Magaña F, O’Hara TM. 2021 Mercury
concentrations in Baja California Sur sh: dietary exposure assessment. Chemosphere. 267, 129233. (doi:10.1016/j.chemosphere.2020.129233)
101. Silva C da, Santos S de O, Garcia CAB, de Pontes GC, Wasserman JC. 2020 Metals and arsenic in marine sh commercialized in the NE Brazil: risk
to human health. Human Ecol. Risk Assessment: Int. J. 26, 695–712. (doi:10.1080/10807039.2018.1529552)
102. Jeevanaraj P, Hashim Z, Elias SM, Aris AZ. 2019 Risk of dietary mercury exposure via marine sh ingestion: assessment among potential
mothers in Malaysia. Expo. Health 11, 227–236. (doi:10.1007/s12403-017-0270-x)
103. Burger J, Gochfeld M, Alikunhi N, Al-Jahdali H, Al-Jebreen D, Al-Suwailem A, Aziz MAM, Batang ZB. 2015 Human health risk from metals in sh
from Saudi Arabia: consumption patterns for some species exceed allowable limits. Human Ecol. Risk Assess. Int. J. 21, 799–827. (doi:10.1080/
10807039.2014.934585)
104. Harding G, Dalziel J, Vass P. 2018 Bioaccumulation of methylmercury within the marine food web of the outer Bay of Fundy, G ulf of Maine. PLoS
One 13, e0197220. (doi:10.1371/journal.pone.0197220)
105. Bakke T, Klungsøyr J, Sanni S. 2013 Environmental impacts of produced water and drilling waste discharges from the Norwegian oshore
petroleum industry. Mar. Environ. Res. 92, 154–169. (doi:10.1016/j.marenvres.2013.09.012)
106. Potts LD, Perez Calderon LJ, Gubry-Rangin C , Witte U, Anderson JA. 2019 Characterisation of microbial communities of drill cuttings piles from
oshore oil and gas installations. Mar. Pollut. Bull. 142, 169–177. (doi:10.1016/j.marpolbul.2019.03.014)
107. Coquery M, Cossa D. 1995 Mercury speciation in surface waters of the north sea. Netherlands J. Sea Res. 34, 245–257. (doi:10.1016/0077-
7579(95)90035-7)
108. Schmidt D. 1992 Mercury in Baltic and North Sea waters. Water Air Soil Pollut. 62, 43–55. (doi:10.1007/BF00478452)
109. Qa3.2021 Further evaluation of steel pipeline coupons for mercury content (total, elemental, organic and Leachable mercury) and investigation
into the ecacy of Mercure as a decontamination solution. Winchester, UK.Repor t No.: Qa32011001.
110. Mark Wilhelm S. 1999 Avoiding exposure to mercury during inspection and maintenance operations in oil and gas processing. Process Safety
Progress 18, 178–188. (doi:10.1002/prs.680180311)
111. Qa3. 2021 Speciation of residual mercury in steel coupons after a 16-hour Mercure decontamination treatment (including mercury that leaches
into seawater over an extended period of time). Winchester, UK. Report No.: QA32104002.
112. Parkhurst D, Appelo C. Description of input and examples for PHREEQC version 3 - a computer program for speciation, batch-reaction, one-
dimensional transport, and inverse geochemical calculations US Geological Survey Techniques and Methods. See http://www.hydrochemistry.eu
113. Dai SS etal. 2021 Global distribution and environmental drivers of methylmercury production in sediments. J. Hazard Mater. 407, 124700. (doi:
10.1016/j.jhazmat.2020.124700)
18
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 11: 230943
114. Jennings S, et al. 2016 Aquatic food security: insights into challenges and solutions from an analysis of interactions between sheries,
aquaculture, food safety, human health, sh and human welfare, economy and environment. Fish Fisheries 17, 893–938. (doi:10.1111/faf.
12152)
115. Froese R, Pauly D. 2022 A spectral three-dimensional color space model of tree crown health. See www.shbase.org (accessed 05 September
2022).
116. Kelly BC, Ikonomou MG, Higgs DA, Oakes J, Dubetz C. 2008 Mercury and other trace elements in farmed and wild salmon from British Columbia,
Canada. Environ. Toxicol. Chem. 27, 1361–1370. (doi:10.1897/07-527)
117. Karimi R, Fitzgerald TP, Fisher NS. 2012 A quantitative synthesis of mercury in commercial seafood and implications for exposure in the United
States. Environ. Health Perspect. 120, 1512–1519. (doi:10.1289/ehp.1205122)
118. White T. 2000 Diet and the distribution of environmental impact. Ecol. Econom. 34, 145–153. (doi:10.1016/S0921-8009(00)00175-0)
119. BEIS. 2020 Strengthening the UK’s oshore oil and gas decommissioning industry government response to the call for evidence. See https://
assets.publishing.service.gov.uk/media/5fd731a1d3bf7f30657f97a1/strengthening-uk-decommissioning-cfe-govt-response-.pdf
120. von Hellfeld R, Hastings A. 2024 Data from: An approach to assessing subsea pipeline-associated mercury release into the North Sea and its
potential environmental and human health impact. Figshare. (doi:10.6084/m9.gshare.c.7095868)
19
royalsocietypublishing.org/journal/rsos R. Soc. Open Sci. 11: 230943
