![Weekly consumption of seafood (g/week) for different nations [72], each dot representing the data for a nation or the overall fish consumption of coastal indigenous communities [74].](https://www.researchgate.net/profile/Rebecca-Von-Hellfeld/publication/369589787/figure/fig5/AS:11431281132941976@1680251356833/Weekly-consumption-of-seafood-g-week-for-different-nations-72-each-dot-representing_Q320.jpg)
Content uploaded by Rebecca von Hellfeld
Author content
All content in this area was uploaded by Rebecca von Hellfeld on Mar 31, 2023
Content may be subject to copyright.
Journal of Hazardous Materials 452 (2023) 131298
Available online 27 March 2023
0304-3894/© 2023 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Research Paper
An approach to assess potential environmental mercury release, food web
bioaccumulation, and human dietary methylmercury uptake from
decommissioning offshore oil and gas infrastructure
Rebecca von Hellfeld
a
,
b
,
*
, Christoph Gade
a
,
b
, Darren J. Koppel
c
,
d
, William J. Walters
e
,
Fenny Kho
c
,
f
, Astley Hastings
a
,
b
a
School of Biological Sciences, University of Aberdeen, School of Biological Sciences, Aberdeen, UK
b
National Decommissioning Centre, Ellon, UK
c
Curtin Oil and Gas Innovation Centre, Faculty of Science and Engineering, Curtin University, Perth, WA, Australia
d
Australian Institute of Marine Science, Perth, Australia
e
Ken and Mary Alice Lindquist Department of Nuclear Engineering, Pennsylvania State University, PA, USA
f
Curtin Corrosion Centre, Curtin University, Perth, WA, Australia
HIGHLIGHTS GRAPHICAL ABSTRACT
•Calculated pipeline mercury threshold
to avoid exceeding environmental
guidelines.
•Novel application of sheries manage-
ment programme for contaminant
tracking.
•Marine mercury contamination could
impact future food webs.
•Offshore decommissioning plans could
benet from additional risk assessments.
ARTICLE INFO
Editor: J¨
org Rinklebe
Keywords:
Minamata convention
Threshold determination
Environmental mercury release
Offshore oil and gas decommissioning plans
Environmental risk assessment
ABSTRACT
Subsea pipelines carrying well uids from hydrocarbon elds accumulate mercury. If the pipelines (after
cleaning and ushing) are abandoned in situ, their degradation may release residual mercury into the envi-
ronment. To justify pipeline abandonment, decommissioning plans include environmental risk assessments to
determine the potential risk of environmental mercury. These risks are informed by environmental quality
guideline values (EQGVs) governing concentrations in sediment or water above which mercury toxicity may
occur. However, these guidelines may not consider e.g., the bioaccumulation potential of methylated mercury.
Therefore, EQGVs may not protect humans from exposure if applied as the sole basis for risk assessments. This
List of abbreviations: CA, Comparative Assessment; CR, Concentration Ratio; EFSA, European Food Safety Authority; ERA, Environmental Risk Assessment; EQGVs,
Environmental Quality Guideline Values; EWI, Estimated Weekly Intake of methylmercury; EwE, Ecopath with Ecosim; FAO, Food and Agriculture Organisation of
the United Nations; FS, Food Standards (for mercury); Hg
0
, Elemental mercury; IAEA, International Atomic Energy Agency; ICES, International Council for the
Exploration of the Sea; K
d
, Partitioning coefcient; M/O ratio, Model versus Observed ratio; TWI, Tolerable Weekly Intake of methylmercury; SQGV, Sediment
Quality Guideline Value; UNCLOS, United Nations Convention on the Law of the Seas; WHO, World Health Organisation; WQGV, Water Quality Guideline Value;
XRF, X-Ray Fluorescence.
* Correspondence to: 23 St Machar Drive, AB24 3UU Aberdeen, UK.
E-mail addresses: rebecca@vonhellfeld.de, rebecca.vonhellfeld@abdn.ac.uk (R. von Hellfeld).
Contents lists available at ScienceDirect
Journal of Hazardous Materials
journal homepage: www.elsevier.com/locate/jhazmat
https://doi.org/10.1016/j.jhazmat.2023.131298
Received 23 November 2022; Received in revised form 3 March 2023; Accepted 24 March 2023
Journal of Hazardous Materials 452 (2023) 131298
2
paper outlines a process to assess the EQGVs’ protectiveness from mercury bioaccumulation, providing pre-
liminary insights to questions including how to (1) determine pipeline threshold concentrations, (2) model
marine mercury bioaccumulation, and (3) determine exceedance of the methylmercury tolerable weekly intake
(TWI) for humans. The approach is demonstrated with a generic example using simplications to describe
mercury behaviour and a model food web. In this example, release scenarios equivalent to the EQGVs resulted in
increased marine organism mercury tissue concentrations by 0–33 %, with human dietary methylmercury intake
increasing 0–21 %. This suggests that existing guidelines may not be protective of biomagnication in all cir-
cumstances. The outlined approach could inform environmental risk assessments for asset-specic release sce-
narios but must be parameterised to reect local environmental conditions when tailored to local factors.
1. Introduction
1.1. Offshore oil and gas infrastructure decommissioning
Oil and gas infrastructure nearing the end of their life may be
decommissioned in different ways, where options include the complete
or partial removal, or in situ abandonment, of submerged structures [1].
The preferred approach depends on, among other factors, the environ-
mental impacts and benets associated with each option. In some ju-
risdictions, for example, the operator of the facility to be
decommissioned must undertake a comparative assessment (CA) of each
of the abandonment options for the infrastructure. These assessments
need to consider all receptors potentially impacted by the decom-
missioning activities in the form of environmental risk assessments
(ERA). The objective of CAs and ERAs is to determine decommissioning
scenarios with minimal impact on decommissioning operations and
ecosystems, and to mitigate potential human risk. For example, the EU
directives 2008/56/EU (Marine strategy framework directive [2]) and
2011/92/EU [3] outline the need to identify, describe, and assess the
direct and indirect impacts of each ‘project’ on biota, the environment,
material assets, and cultural heritage. Special focus should be given to
oil and gas pipelines, as well as contamination by oil or gas exploration.
Similar legislation exists in Australia, necessitating environment plans to
demonstrate acceptable environmental impacts/risk levels for any
decommissioning activity.
In situ decommissioning, where subsea infrastructure is cleaned and
abandoned in the marine environment, is considered a cost and time
effective decommissioning option for pipelines. The structures can be
abandoned entirely or in parts, and different cleaning protocols are
available for the pipelines’ inner surface before abandonment [4,5]. This
is important because contaminants are known to accumulate in subsea
infrastructure which may pose an unacceptable risk to the marine
environment [6,7]. A plethora of international treaties and legislations
have been developed over the years to ensure protection and sustainable
use of the marine environment that apply to in situ decommissioning
decisions. These include: (1) The ‘United Nations Convention on the Law
of the Sea’ (UNCLOS), which outlines the rights and responsibilities of
nations for the use of the ocean, and that any abandoned/disused in-
stallations must be removed with consideration for the environment [8].
It also states that the exploitation of continental shelf may not interfere
with navigation, shing, or conservation works, and that all abando-
ned/disused structures must be removed entirely, as previously stated
under the ‘Convention on the Continental Shelf 1958’ [9]; (2) the
‘Convention on the Prevention of Marine Pollution by dumping of
wasted and other matters 1972’ (London Convention) and the ‘London
Protocol’, which aim to prevent marine pollution by regulating the
dumping of wastes and other matters [10]; and (3) the ‘Minamata
Convention on Mercury’ which is designed to protect human health and
the environment from anthropogenic emissions and releases of mercury
and mercury compounds [11]. The regulatory threshold to justify in situ
abandonment of pipelines (e.g., in Australia) is currently governed by
the need to demonstrate equal or better environmental and health and
safety outcomes than complete removal, as well as meeting all further
regulatory and legal requirements [12].
1.2. Mercury in the marine environment
Mercury is a contaminant of concern for offshore industries, because
it is a naturally occurring heavy metal present in oil and gas reservoirs
[13,14]. Other processes associated with the anthropogenic release of
mercury include mining and smelting, coal burning, cement production,
and artisanal gold mining [15]. In the marine environment mercury can
form a range of chemical species [16], depending on local factors such as
the sediment redox potential, pore-water sulphide concentration, sedi-
ment organic matter content, pH, and sediment texture [17–20]. Many
mercury species are not very water soluble, but within abandoned
pipelines some (such as elemental mercury; Hg
0
) may leach into the
dissolved phase over time. Upon release, they will disperse and speciate
depending on the local environment leading to their partitioning be-
tween the water column and sediment phase [7]. In the sediment,
site-specic parameters will inuence its bioavailability by forming e.g.,
insoluble sulphide complexes in the presence of relevant pore water
sulphide concentrations, or more bioavailable species [17].
The speciation of mercury, together with environmental parameters,
affect its availability for methylation by certain bacterial communities.
If methylated, mercury forms toxic and bioaccumulative organomercu-
rials (e.g., mono- and di-methylmercury, herein methylmercury) [21].
Methylmercury has a higher propensity to bioaccumulate and bio-
magnify than other mercury species [22], along with greater toxicity to
marine organisms [23]. Exposure to methylmercury can lead to devel-
opmental neurotoxicity in foetuses, as well as various adverse effects
after birth [24], Due to its biomagnication, it may adversely affect high
trophic level organisms such as marine mammals and large predators in
marine food webs [25].
Mercury is a contaminant of global concern, as outlined by the
Minamata Convention on Mercury [26]. The convention’s goal is the
reduction of mercury emissions from sources such as the industrial
processess, artisanal small-scale gold mining, medical and municipal
waste incineration, and sulphide ore roasting [27,28]. Oil and gas
extraction is also identied as an activity that may release mercury.
During hydrocarbon extraction, mercury remains in the produced pe-
troleum derivatives (mostly as Hg
0
and sulphide-bound mercury) [29,
30]. Mercury accumulates in the production infrastructure through
deposition to pipeline surfaces via condensation, adsorption to steel and
into corrosion products, as well as in e.g., sludges and produced water
[4,31,32]. Historically, studies assessing the impact of oil and gas pro-
duction on the marine ecosystem have focused on contamination
occurring during the infrastructures’ operational life, rather than the
potential environmental contamination of in situ decommissioned
pipelines. If pipelines are to be abandoned in situ, guidelines on the
acceptable levels of residual contaminants, such as Hg, must be devel-
oped. A detailed review of subsea pipeline associated mercury and the
ecological risk has recently been published, providing further insights
into these aspects [7].
1.3. EQGVs and consumer protection
Environmental quality guideline values (EQGVs) are values below
which there is expected to be a low probability that a pollutant will have
R. von Hellfeld et al.
Journal of Hazardous Materials 452 (2023) 131298
3
a negative impact on the environment, based on a particular measure of
impact [33]. EQGVs for mercury in the marine environment exist for a
range of jurisdictions to support contaminant ecological risk assess-
ments. For the sediment compartment, the most common guidelines
(referred to as sediment quality guideline values; SQGVs) range from
0.13 to 0.7 mg/kg dry sediment [34–36], although some nations are
more stringent, with a SQGV of 0.07 mg/kg dry sediment [37,38]. The
water quality guideline values (WQGVs) ranged from 0.016 to 1.4 µg/l
[35,39], with more variation between the nations than was the case for
the SQGVs. An overview of the minimum EQGVs of various nations can
be found in Table S1. These values are based on toxicity data obtained
from laboratory and eld studies of species from a range of different
taxonomic groups [40]. However, these values do not necessarily pro-
tect against the potential long-term impacts arising from mercury bio-
magnication, affecting marine organisms or seafood consumers.
Seafood consumers are protected by regulations on mercury in
landed sh. These include the food standards (FS) for total mercury in
tissue and the tolerable weekly intake (TWI) for methylmercury. The FS
describe the acceptable levels of mercury in sh muscle tissue for
commercially sold species and recommended sh consumption for
humans per week. Due to the bio-accumulating and -magnifying nature
of mercury, these FS often account for the trophic position of the species
in question and provide separate values for species that are more or less
likely to accumulate high amounts of mercury (Table S1). The TWI aims
to limit human dietary exposure to methylmercury due to its highly toxic
nature, and because most of the mercury measured in sh tissue is
methylmercury [41]. Most countries implement the TWI of 1.3 µg
methylmercury/kg bodyweight (herein µg/kg) as outlined by the Eu-
ropean Food Safety Authority (EFSA) [42], which is more conservative
than the previously determined TWI of 1.6 µg/kg provided by the
FAO/WHO [43].
1.4. Aims and objectives
The sustainable development principle of intergenerational equity,
that the present generation should ensure that the health, diversity, and
productivity of the environment are maintained or enhanced for the
benet of future generations, requires that the long-term fate and im-
pacts of mercury are considered in any decommissioning activity [44].
The current EQGVs were developed to protect the local ecosystem from
potential toxic effects but do not necessarily protect against bio-
magnication impacts. They are also not intended to be used as the sole
point of comparison for environmental impact and risk assessment
studies that consider human health.
This paper aims to demonstrates an approach to assess the potential
biomagnication impacts from mercury releases to marine food webs.
We use existing EQGVs as input mercury concentrations to explore how
protective EQGVs are of mercury biomagnication impacts in the ma-
rine food web. In this study, EQGVs were rst used to back-calculate
pipeline mercury threshold concentrations that describe the concen-
tration of mercury in pipelines that when released to the environment
will not exceed EQGVs for sediments and seawater. Then, a hypothetical
marine food web was used to outline the applicability of the modelling
programme Ecopath with Ecosim (EwE) for the determination of mer-
cury bioaccumulation in a food web using generic parameterisation,
upon which site-specic assessment can be built. Four trials describing
hypothetical release scenarios were selected to determine the applica-
tion to calculate future biota mercury concentrations. The derived biota
mercury concentration data was then used to showcase the potential of
this method to assess human risk by assessing the possible exceedance of
the FS for mercury in foods by calculating dietary exposure to methyl-
mercury, using published global weekly sh consumption data. The
example implementation of this study’s approach does not intend to
draw conclusions on the direct health implications or to discern the
actual intake of methylmercury of future generations of seafood con-
sumers. Rather, the data needs for each step for future site-specic
parametrisations are discussed, to inform the local implementations of
such methods.
2. Methods
To determine pipeline mercury threshold concentrations that would
not exceed the sediment and water EQGVs under conservative release
scenarios (dened as estimates that err on the side of caution, also
referred to as a ‘worst case scenario’ [45]) a simple calculation is pro-
posed (part A). Modelling is then proposed to determine the bio-
accumulation potential of mercury in a marine food web and the
potential for future exceedance of FS and TWIs (part B). The following
steps were taken:
A. The pipeline mercury threshold concentrations are calculated using
different approaches, to provide output values in area, length, and
mass-based units, allowing for the comparison to data derived from
different measurement techniques such as acid digest, pipeline
pigging, or X-ray uorescence (XRF) assessment of pipeline coupons.
This simplies direct comparison between actual measurements and
threshold values with the same unit of measurement, standardising
the approaches.
B. A hypothetical food web designed in EwE, using the Ecotracer
contaminant tracking tool, was used to assess the mercury accumu-
lation of different contamination scenarios. The output was con-
verted to muscle tissue concentrations and compared to current FS
for mercury in sh. The muscle tissue concentrations then informed
the estimated weekly ingestion of methylmercury (EWI) for humans
from seafood to determine the potential future human risk of
increased methylmercury via dietary uptake.
Where possible, realistic parameterisations and considerations are
given for each step in the outlined approach. However, these are not
specic to a particular location or food web, Additionally, various data
gaps necessitated the use of parameterisations that may be unrealisti-
cally simplistic. For the purpose of demonstrating the approach, these
were selected to ensure they represented a conservative exposure sce-
nario (i.e., assumed that all mercury was bioavailable). The assumptions
are highlighted throughout the methods section and further re-
quirements are discussed in more detail in the discussion.
2.1. A: pipeline mercury threshold calculation
To simplify the approach, a 1 kg sediment compartment (hereafter
called the sediment box) underlying a pipeline is postulated, to allow for
the calculation of a pipeline mercury threshold (Fig. 1). This assumes
that at the end of pipe decay all the mercury falls in this sediment box.
After determining the mercury concentration in the box, the result-
ing water contamination can be calculated using a partitioning coef-
cient (K
d
value). A K
d
value can be used to approximate the behaviour of
Fig. 1. Area-based approach to determining the pipeline mercury threshold
according to the SQGV and WQGV.
R. von Hellfeld et al.
Journal of Hazardous Materials 452 (2023) 131298
4
mercury in the marine environment, with a sediment-water K
d
value
describing the relationship between the solid and dissolved phase [46],
assuming an equilibrium between the contaminant fraction in the
sediment and the water column:
Kd(l/kg) = Hg concentration per unit sediment mass(kg/kg)
Hg concentration per seawater volume(kg/l)(1)
The recommended K
d
value for open ocean of 4 ×10
3
by the IAEA
[47] was applied in the following work, a value below commonly re-
ported empirical values (typically >10
4
) [48,49]. This value will most
likely overestimate the amount of mercury in the dissolved phase, but
follows the approach recently discussed in an environmental plan for the
decommissioning of a gas export pipeline [50]. As recommended by the
IAEA, K
d
values higher and lower than the recommended value were
used (here 4 ×10
2
and 4 ×10
4
), to determine the percentage change
between the K
d
values. The determined difference for the IAEA recom-
mended K
d
values was between 1 % and 10 % from the 4 ×10
3
K
d
value
(data not shown) indicating that a tenfold change in K
d
value induced
only a 10 % change in resulting maximum permissible pipeline con-
centration and thus supporting the use of the chosen K
d
value. No change
in steel density due to corrosion is considered, and the calculations used
the API 5L pipeline dimensions (Table S2). The calculations were done
in Excel (Microsoft, Version 2201) and the Tidyverse package for R [51,
52] was used for data visualisation.
2.1.1. Area/Length-based calculation
The diameter of the pipeline (A) is xed to represent the width of the
underlying sediment box. Pipelines used for subsea oil and gas in-
frastructures range from 2 ′′ diameter for gas pipelines to 14 ′′ for export
pipelines. Additionally, the wall thickness varies, typically being be-
tween 6 and 40 mm for frequently used pipelines (see Table S2 for the
pipeline dimensions considered here). The depth of the sediment box (C)
varied, with 5 cm used to derive the minimum and 10 cm the maximum
permissible concentration. 5–10 cm is a common depth for taking
sediment samples (shown as ‘min’ and ‘max’ in the results), and the
density of marine sediment/water (δ) is known. Thus, the length of the
sediment box (B):
B= (A∗C)/δ(2)
B also represents the length of the pipeline section. Using this value,
and the inner diameter of the pipeline, the inner surface area (
α
)of the
pipeline can be determined:
α
=2
π
rB (3)
The resulting information can then be applied to determine either the
area- or the length-based pipeline threshold values for mercury
contamination considering the guideline value for either the sediment or
water compartment.
2.1.2. Mass-based calculation
Given the known mass of the pipeline material, Eqs. (1)–(3) derived
above can also be converted to provide an output value as weight/length
unit (kg/m) in the mass-based approach for comparison with data of
mercury concentration obtained from e.g., pigging (Fig. 2). This
approach considers the pipeline parameters to compute the weight of
the pipeline segment.
The output derived from length-based approach was divided by the
weight of the steel pipe:
Compartmentmax(mg/kgsteel) = Compartmentmax (mg/m)/Steelweight(kg/m)
(4)
2.2. B: food web modelling
EwE is a mass balance food web modelling programme, which allows
for a static mass balance (Ecopath), as well as temporal (Ecosim) and
spatiotemporal modelling (Ecopath). In addition, a tool for the tracing of
contaminants has been developed (Ecotracer), using already established
models and equations. A detailed description of the programmes main
equations and underlying parameters can be found in the publication by
Christensen and Walters [53].
The Ecotracer tool within EwE simulates the transport of any
contaminant through the food web, solving the contaminant dynamic
equation simultaneously with the outlined EwE equation [54]. It allows
for a varied contaminant inux over time and considers different
decomposition/outow methods. The underlying assumption is that the
contaminant is either within the environment (the water compartment),
or the species (functional group). Each functional group is considered as
a compartment and can thus have different contaminant concentrations.
2.2.1. ‘Anchovy Bay’
A hypothetical food web called ‘Anchovy Bay’ was used for this study
(Fig. 3) which was previously part of a course taught at the University of
British Columbia [55]. This food web has been veried, is
mass-balanced, and spans 4 trophic levels, which can be seen as repre-
sentative of most marine food webs. The model was selected to
demonstrate a potential method for the determination of mercury bio-
accumulation in marine food webs that can be used to assess environ-
mental risks related to activities, such as offshore oil and gas
decommissioning. Due to a lack of mercury concentration and decom-
missioning data for locations globally, the hypothetical ’Anchovy Bay’
model was thus chosen, to demonstrate the applicability of Ecotracer, as
it has been used in past contaminant tracking examples (see Walters and
Christensen [54]) and will now be applied to decommissioning-related
contaminant accumulation in marine food webs. Input data, as well as
the dietary matrix, output parameters and statistical assessments of the
model can be found in Table S3.
Four different trials were run with the ‘Anchovy Bay’ model, to assess
Fig. 2. Mass-based approach to determining the pipeline mercury threshold according to the WQGV (left) and SQGV (right).
R. von Hellfeld et al.
Journal of Hazardous Materials 452 (2023) 131298
5
the impact of different release scenarios on the bioaccumulation of
mercury in the food web over a total of 200 years via Ecosim. All con-
centrations of mercury inux are based on literature-derived values,
converted to the model environment. Here it was assumed that all
mercury that was added directly entered the water compartment and
was available for incorporation into the model food web (i.e., conser-
vatively assuming its complete availability to be methylated and bio-
accumulate). Although this is not environmentally representative, as
only approximately 15 % of the total mercury in the water column is
accounted for by methylmercury [56–59], and with most of the
methylation occurs in marine sediments [60,61]. Direct uptake rates for
lower trophic level organisms were calculated from real world examples
(see Eqs. (5)–(7) for details), which was then validated against actual
biota concentrations (see Eqs. (8) and (9)). One 1 km long representative
pipeline with a diameter of 35.56 cm, and a wall thickness of 1.91 cm
was used for all calculations of equivalent releases in the trials. The food
web was dened as a closed system with an area of 1 km
2
and a depth of
100 m. All input values are listed in Table S4.
Trial 1 – Background exposure: This trial represents the current
environmental conditions of existing oceanic mercury concentration,
providing an insight into the potential future mercury bioaccumulation
if the current rate of anthropogenic and natural mercury released into
the environment is unchanged in the next 200 years. The initial envi-
ronmental concentration was calculated from the global ocean mercury
mass of 68 million metric tons, resulting in an average concentration of
0.19 t/km
2
provided by Neff et al., [62]. They further determined an
annual global oceanic anthropogenic inux of up to 8800 t/year
(0.000024 t/km
2
/year). No release of pipeline-associated mercury is
considered in this trial.
Trial 2 – Release of 11 mg mercury/m pipeline (mg/m) (0.005 t)
over 1 year: In addition to the trial 1 assumptions, it is assumed that a
1 km pipeline segment in the modelled ecosystem is contaminated with
11 mg mercury/m pipeline (herein mg/m). When considering the K
d
value of 4 ×10
3
kg/l [47],this results in a dissolved mercury concen-
tration equivalent to the most stringent WQGV listed in Table S1 of
0.05 µg/l, herein referred to as ‘low WQGV’. This amount of mercury is
released into the modelled ecosystem in a single-release event after 50
years as an input of 0.005 t/km
2
, before returning to the previous inux
value.
Trial 3 – Release of 22 mg/m mercury (0.01 t/km
2
) over 1 year:
In addition to the trial 1 assumptions, it is assumed that a 1 km pipeline
segment in the modelled ecosystem is contaminated with 22 mg/m
mercury [63], which is equivalent to 0.01 t dissolved mercury in the
food web model, or 0.1 µg/l considering a K
d
value of 4 ×10
3
[47]. This
is a WQGV for mercury for various nations (Table S1). This amount of
mercury is released into the modelled ecosystem in a single-release
event after 50 years as an input of 0.01 t/km
2
/y, before returning to
the previous inux value.
Trial 4 – Release of 86 mg/m mercury (0.04 t/km
2
) over 1 year:
In addition to the trial 1 assumptions, it is assumed that a pipeline
segment in the modelled ecosystem is contaminated with 86 mg/m
mercury, which is the equivalent of the least stringent WQGV listed in
Table S1 of 0.4 µg/l, herein referred to as ‘high WQGV’. This amount is
released into the modelled ecosystem in a single-release event after 50
years as an input of 0.04 t/km
2
/year, before returning to the previous
inux value.
Trials 2–4 represent a concentration series as all underlying as-
sumptions remain the same as outlined in trial 1 with the addition of
varying mercury concentrations in the single release event. This was
done to determine the rate at which increased mercury emissions led to
increased biota mercury concentrations. Incremental release trials
(where an equivalent amount of mercury to trials 2–4 was released over
100 years) were also run (data not shown) but the nal accumulated
mercury concentration did not vary. To this end, the percentage increase
in muscle mercury concentration between trial 1 and trials 2–4 were
examined, rather than the actual mercury accumulation. It must be
noted that the food web used here is a simplied representation. Mer-
cury biomagnication will be highly sensitive to local environments and
ecosystems and so these results should not be interpreted as being pre-
dictive of impacts from real mercury releases. For the derivation of the
equivalent pipeline mercury concentration of trials 2–4, the sediment
water partitioning of mercury was taken into consideration, and a
complete distribution of mercury in the respective overlaying water
column was assumed. The additional mercury concentrations modelled
to be released into the ‘Anchovy Bay’ environment are within the range
of those measured in oil and gas pipelines of different diameters [7].
To derive the initial concentrations for the functional groups, the
literature-derived value mercury concentration of 83 ng/g dry weight
for phytoplankton by Bełdowska and Kobos [64] was used, converting it
to 0.004 mg/kg wet weight by applying the dry-to-wet weight conver-
sion factor of 20 [65], as well as the initial environmental concentration
of 0.19 t/km
2
[62]. Then, the uptake rate was computed following the
calculations discussed by Booth et al., [66], to match the modelled
concentration to the literature derived organism mercury concentration.
This approach allows one to tailor the model more accurately to envi-
ronmental conditions and can be done with more robust data for
regions/systems of interest, where present. Briey, the primary pro-
ducer direct uptake rate (ui) is derived by:
Fig. 3. ‘Anchovy Bay’ food web structure. The blue lines indicate food web interaction, the line thickness indicates the proportion of biomass transferred between the
functional groups.
R. von Hellfeld et al.
Journal of Hazardous Materials 452 (2023) 131298
6
ui=CRi× (Pi/Bi+mi+di)(5)
Based on the concentration ratio CRi, the production over biomass
ratio P
B, the excretion rate mi, and the physical decay rate di. The con-
centration ratio is computed with the contaminant concentration within
the group Ai, the biomass of the group B, and the initial environmental
concentration C0:
CRi= (Ai/Bi)/C0(6)
To derive the initial concentration for all other functional groups in
the ‘Anchovy Bay’ model, the model was allowed to equilibrate with
only the above-mentioned parameters (see Table S4). The thus derived
initial concentrations for the remaining groups were then used for all
further trials and for the calculation of the direct uptake rate of the
zooplankton and benthos functional groups. For consumer functional
groups, the uptake rates are derived from the predation losses and
consumption uptakes as follows:
ui=Ci,eq ×
j=predator
Zi+mi+di−AEi×
j=prey
QijAj
Bj
Bi×C0
(7)
Here the losses from predation include the concentration of the
functional group of interest Ci and sum of the group’s mortality,
excretion, and physical decay rates. The gains due to consumption of
other functional groups includes the assimilation efciency of the
functional group of interest AEi and the sum of uptake as consumption
rate of j prey Qij with a concentration Aj and biomass. In addition, the
demethylation rate by whales and seal was set to 15 % and 25 %,
respectively [67].
To determine the validity of the model-derived initial concentrations
from the direct uptake rate of phytoplankton as outlined above,
literature-based muscle concentrations were examined taking species
not directly listed in the model but of the same trophic level into ac-
count. Additionally, data from publicly accessible databases was used,
creating an average concentration for each species observed. The dietary
matrix based trophic levels for all species were derived from FishBase
and SeaLifeBase (Table S5) [68,69]. Only studies published in English
were reviewed for this purpose, where the trophic level of the examined
species could be determined, muscle samples were analysed, it was
evident whether the concentration was given in dry or wet weight, and
where clear distinctions were made between the data obtained in the
study and data taken from literature for comparative purposes. Addi-
tionally, studies were excluded that examined mercury concentrations
in food products, as no clear indication was given on the number of
freeze-thawing cycles and overall sample preparation/handling. Where
dry weight mercury concentrations were used, the values were divided
by the wet-to-dry conversion factors published by Cinnirella et al. [70]
and necessary unit conversion were conducted. In addition, the
modelled (M) versus mean observed (O) mercury concentration ratio
was calculated for each trophic level, following the example of Li et al.
[71] where:
M/O ratio =M/O(8)
The ratio between the modelled (M) and observed (O) muscle tissue
concentration for a given species allows one to assess how comparable
the model derived initial mercury concentrations are to biota samples.
The closer to one the M/O ratio value is, the more similar the modelled
and the observed concentration are and thus the more accurately the
model initial concentrations represent the concentration measured in
biota samples. The M/O value was used to determine when the modelled
value output is predictive of actual environments and where the model
over- or under-estimated the biota concentration. Additionally, the
normalised mean bias of the M/O ratio, as outlined by Li et al. [71] was
determined:
Normalised mean bias =
n
1
(M−O)
n
1
O(9)
As a hypothetical marine ecosystem has been used, and sh tissue
concentrations were taken from multiple studies representing multiple
ecosystems, the M/O ratio was used to validate that the EwE and Eco-
tracer model outputs were adequately parameterised to give represen-
tative outputs reecting generic biota tissue concentrations (see Section
4.2 for further details).
2.2.2. Methylmercury EWI
To determine the actual weekly sh consumption for each nation in
2019, the FAO database was used [72]. All searches were based on the
following selection criteria: Countries – “Select All”; Elements – “Food
supply quantity (kg/capita/yr)”; and Years – “2019”. The items were
grouped into three (Items – “Demersal Fish”, “Marine Fish, Other”, and
“Pelagic Fish” OR “Crustaceans”, “Cephalopods”, “Molluscs, Other” OR
“Meat, Aquatic Mammals”, “Aquatic Animals, Other”) and will be
referred to as “sh”, “benthos” and “mammals” herein. It should be
noted, however, that the database does not allow for the accurate sep-
aration between freshwater and marine species in all cases (e.g., for
demersal and pelagic sh and crustaceans). This is not thought to affect
the accuracy of the present ndings as average consumption values are
used for this study, and the intention was to demonstrate an assessment
method that should be tailored to individual exposure scenarios. The
respective consumption rate was then applied to the functional group it
is associated with, assuming that no seafood from other sources was
consumed. To account for the dietary diversity in terms of sh species
consumed, anchovy and mackerel were dened as oily sh, with the
remaining sh species dened as white sh in accordance with the food
standards agency denition [73] when assessing the weekly consump-
tion of methylmercury via dietary intake. To account for the higher
reliance on marine resources of coastal indigenous communities, the sh
consumption of such communities was derived from
Cisneros-Montemayor et al. [74]. The publication, however, made no
distinction between different types of seafood and thus the data was
applied to all functional groups. However, the muscle tissue concen-
trations published in this work can be used to derive the weekly meth-
ylmercury intake for communities with a different dietary matrix.
Comparing the EwE output to e.g., the FS or the TWI, more accu-
rately, literature derived data regarding the distribution of total mercury
in the functional groups was used. Of the total body burden of mercury,
the following percentages were noted to accumulate in muscle tissue for
each functional group in the ‘Anchovy Bay’ food web: 2 % for whales
[75], 5 % for seals [76], 50 % for all sh species [77], 15 % for shrimp,
and 100 % for the benthic [78–80] and planktonic species [81,82], and
detritus. This approach assumes that the mercury released into the
marine environment disperses homogenously within the model, and that
no migration occurs. The hypothetical model used here was selected for
the simplicity of the food web, lending itself as a proof-of-concept model
for the method presented. The results of trial 1 was then used as baseline,
deriving the percentage increase in muscle tissue mercury concentration
for trials 2–5.
The output obtained from the Ecotracer trials was further used to
calculate the methylmercury EWI based on the weekly sh consumption
data. The fraction of total mercury in muscle tissue that represented
methylmercury was 95 % [41]. Additionally, the bioaccessibility, the
fraction of ingested methylmercury that is released from the food matrix
into a soluble form within the gastrointestinal tract, for the species (or
related species) modelled in the ‘Anchovy Bay’ food web was obtained
from Bradley et al. [83] (Table 1). The data here refers to raw samples,
as gaps in the published data did not allow for the assessment of other
preparation methods. Moreover, the data for cooked samples frequently
presented with higher bioaccessibility rates of methylmercury than raw
samples, thus using raw sample bioaccessibility data provided a degree
R. von Hellfeld et al.
Journal of Hazardous Materials 452 (2023) 131298
7
of conservatism to the study.
Cysteine is the most abundant form of protein- or peptide-bound
thiol in biological systems [84]. It is the major complexing agent in
(sh) muscle tissue [85], including for methylmercury [86]. The EWI
was calculated based on the average sh consumption of women of
childbearing age and children, using the mercury accumulation in
muscle tissue (AC; µg/kg), the fraction of total mercury that is methyl-
mercury (F; %), the weekly intake (WI; kg), the absorption rate though
stomach cells (AB; %, known to be 79 % for cysteine bound methyl-
mercury [83]), and the persons weight (W; kg). In this case, the weight
of the average person used (62 kg), was in accordance with Walpole
et al. [87]:
ConsumedMeHg = (ACxFxWIxAB)W(10)
To account for the difference in white and oily sh consumption, it
was assumed that ≤280 g (two portions) per week of oily sh were
consumed [88], with the rest of each nations sh consumption being
made up of white sh. In the case of nations consuming too little sh a
week to account for 2 portions of oily sh, only one portion (140 g) was
assumed. Where this was not feasible, 50 % of the consumed weekly sh
was assumed to be oily. For shellsh or marine mammal intake rates, the
FAO database derived consumption rate was split equally between the
respective functional groups from Ecotracer. As no distinction in con-
sumption rates of different types of seafood were provided for the
indigenous communities [74], an even split between all functional
groups was assumed. In addition, the overall mean intake values were
used to generalise the mean global methylmercury intake via seafood
using the following consumption rates: Crustaceans 6 g, sh 23.5 g,
marine mammals 1 g, and for the indigenous community a mean con-
sumption rate of 1.42 kg was used. Here, too, the results derived from
trial 1 were then used as baseline, determining the percentage increase
in weekly ingested methylmercury for trials 2–5.
2.3. Limitations of the modelling approach
The outlined approach relies on simplications of complex envi-
ronmental processes, due to the lack of data and the use of a hypothetical
food web. These simplications were selected for the purpose of
demonstrating the approach and providing perspective on the data re-
quirements for local implementation. The assumptions, some of which
may be unrealistic for local implementations, are highlighted below for
transparency and to outline where future research focus is needed, in
order to make the method presented suitable for site-specic use. As-
sumptions made in this work include:
1. Calculating seawater mercury concentrations from solid-phase
mercury concentrations using an open-ocean K
d
value.
2. Assuming there is no sediment mixing and dilution post-pipeline
release, and the potential for dispersion of aqueous mercury.
3. Representing all aqueous mercury as bioavailable for uptake by
marine biota.
4. Using a closed food web model that does include mercury removal
mechanisms by ocean currents or allow for the migration of organ-
isms in and out of the food web.
5. That all consumed seafood originated from the modelled food web.
These limitations are described to encourage proponents seeking to
adapt this approach to their context to apply an appropriate complexity
to ensure an appropriate level of condence in the modelling outcome.
3. Results
Pipeline mercury threshold concentrations were computed for the
highest and lowest published guideline values (Table S1) to give a range
representing the different EQGVs adopted by different nations. The
EQGVs applied to the calculations are 0.05 µg/l [37] and 0.4 µg/l [89]
for the water compartment, and 0.13 mg/kg [89] and 0.15 mg/kg [36]
for the sediment compartment, representing the lowest applied EQGVs
for different nations. A sampling depth of 5–10 cm was used to param-
eterise the sediment box below the pipeline, and the previously dis-
cussed K
d
value of 4000 l/kg used [47]. The WQGV of 0.4 µg/l was also
used to inform the potential future mercury accumulation in marine
food webs (as outlined in Section 2.2.1, trial 4).
3.1. A: Pipeline mercury threshold calculation
This calculation investigated the pipeline mercury threshold con-
centrations to remain within the EQGVs. Overall, the low SQGV
(0.13 mg/kg) was the most sensitive threshold value when deriving
residual pipeline mercury concentrations. The area-based pipeline
mercury threshold concentrations derived here were not affected by the
pipeline parameters and was purely determined by changes in the
referenced EQGV. The pipeline threshold concentration before
exceeding the low SQGV was 7 mg/m
2
, compared to 108.2 mg/m
2
before exceeding the low WQGV (Table 2). These differences likely
reect the choice of K
d
value.
For the derivation of the length-based threshold concentration, the
thinner the wall thickness and the larger the pipe diameter, the higher
the permissible mercury concentration (mg/m). The reverse relationship
was observed for the mass-based pipeline mercury threshold concen-
tration (mg/kg). This can be seen in Fig. 4. For the length- and mass-
based calculations, the pipeline diameter and wall thickness inu-
enced the pipeline mercury threshold concentration (Fig. 4). Here, the
derived pipeline threshold concentration to remain below the low
SQGVs were 2.00 mg/m (for a pipeline with 4 ′′ diameter and 1.1 cm
wall thickness) or 0.003 mg/kg (14–18 ′′ diameter, 3.5–4.5 cm wall
thickness), for the length- and mass-based calculation, respectively,
considering a sediment sampling depth of 5 cm. The pipeline threshold
Table 1
Bioaccessibility (%) of methylmercury from raw
samples from the species of the ‘Anchovy Bay’ food
web model to human stomach epithelial cells as
published by Bradley et al. [83].
Species Bioaccessibility
Whale 98
a
Seal 98
a
Cod 77
Whiting 100
b
Mackerel 80
Anchovy 100
b
Shrimp 100
Benthos 100
c
a Values refer to tuna.
b data refers to meagre.
c data refers to scallops.
Table 2
Pipeline mercury threshold concentrations to remain below the EQGVs
(Table S1; low WQGV: 0.05 µg/l, high WQGV: 0.4 µg/l, low SQGV: 0.13 mg/kg,
high SQGV: 0.15 mg/kg) for different pipeline congurations (Table S2) and
sampling depths.
Low EQGV High EQGV
Area-
based
[mg/
m
2
]
Length-
based
[mg/m]
Mass-
based
[mg/
kg]
Area-
based
[mg/
m
2
]
Length-
based
[mg/m]
Mass-
based
[mg/
kg]
WQGV Min 108.20 31.28 0.39 865.8 250.00 3.13
Max 108.20 151.13 2.45 865.8 1290.00 19.59
SQGV Min 7.00 2.00 0.03 8.1 2.30 0.03
Max 7.00 9.80 0.16 8.1 11.30 0.18
R. von Hellfeld et al.
Journal of Hazardous Materials 452 (2023) 131298
8
concentrations derived for remaining below the low WQGV were
31.28 mg/m (4 ′′ diameter, 1.1 cm wall thickness) and 0.39 mg/kg (18 ′′
diameter, 4.5 cm) for the length- and mass-based output, respectively.
3.2. B: food web modelling
3.2.1. Muscle tissue mercury accumulation
To determine the representativeness of initial mercury concentration
used in the ‘Anchovy Bay’ model, publicly available mercury biota
concentrations were obtained from literature, as well as databases for
the North Sea [90,91] and US coasts [92]. ‘Anchovy Bay’ modelled
mercury concentrations were within the distribution of measured mer-
cury concentrations for species of comparable trophic levels (Fig. 5). The
M/O ratio for whiting and shrimp were closest to one. All derived M/O
values were between 0.13 and 3.0, with the model derived values for
zooplankton, benthos, shrimp, mackerel, cod, and whale exceeding the
literature derived ones (M/O ratio >1) and the modelled concentrations
for phytoplankton, anchovy, whiting, and seal remaining under those
provided by literature (M/O ratio <1). This was further supported by
the determined normalised mean bias (Table S5).
The nal mercury muscle tissue concentration in the different
functional groups of the ‘Anchovy Bay’ ecosystem increased with
increasing inux concentrations (i.e., from trial 1 to trials 2–4, Table 3).
In trial 1, mackerel and both marine mammal species, exceed the
0.5 mg/kg value. Cod and whiting were found to exceed the 1 mg/kg FS
in trial 1. Relative to background input (trial 1), the additional inux of
mercury modelled in trials 2–4 is 0.5 %, 1%, and 4 %, respectively,
eliciting increased biota concentrations of up to 33 %, compared to trial
1. Due to the different feeding behaviours of the functional groups, as
well as the modelled detoxication process for whales and seals, the
species with highest muscle tissue concentration in the scenarios are
cod, whiting, and mackerel. Due to the short generation time and
feeding habits of the lower trophic level organisms, the overall accu-
mulation of mercury here does not increase signicantly, even with
increased annual mercury inux. For trials 2–4, the concentrations are
shown as percentage increase from trial 1 (actual concentrations in
Fig. S1).
3.2.2. Hypothetical weekly dietary intake of methylmercury
The average seafood consumption rates of non-indigenous commu-
nities were: 1 g/week aquatic mammals, 235 g/week sh, 59 g/week
benthic and non-sh species. These values were used for all further
calculations of methylmercury intake for non-indigenous communities.
The average seafood consumption of coastal indigenous communities
was 1.43 kg/week (Fig. 6).
The EWI in trial 1 was calculated for the consumption rates of sea-
food by non-indigenous communities, and only the consumption of
white sh led to an exceedance of the 1.3 µg/kg TWI [42] (Table 4). The
EWI from all seafood groups would also exceed the TWI. No increase in
EWI was determined between trial 1 and 2. Between trials 1 and 3, an
increase in EWI from white sh and total seafood consumption (of 6.7 %
and 4.2 %, respectively) was determined. Between trials 1 and 4, an
increase in ingested methylmercury of 16.7 % overall was calculated,
and both oily and white sh led to an increased intake in this trial
(14.3 % and 20 %, respectively). Considering the seafood consumption
rates of coastal indigenous communities, only the overall seafood con-
sumption data could be assessed, highlighting an exceedance of the TWI
[42]. An increase in EWI for trials 2–4, compared to trial 1, was noted (of
2.5 %, 5.2 %, and 20.5 %, respectively).
Fig. 4. Pipeline mercury threshold value calculation for WQGV (top) and SQGV (bottom). The length- (left) and mass-based (right) mercury pipeline threshold values
for the low (continuous lines) and high (dashed lines) EQGVs for different pipeline schedules (colours).
R. von Hellfeld et al.
Journal of Hazardous Materials 452 (2023) 131298
9
4. Discussion
The work conducted here demonstrated an approach to assess the
potential biomagnication impacts of mercury released into the marine
environment and explores how protective existing EQGVs are of food
web impacts. This work has focussed on remaining below the predened
EQGVs to outline the pipeline mercury threshold values that would
require no additional testing or analysis in the context of environmental
protection.
4.1. A: Pipeline mercury threshold calculation
This work has focussed on determining threshold values for pipeline
mercury concentrations which, upon release from a corroding pipeline
decommissioned in situ would not exceed existing guideline values. It
should be noted that the exceedance of the EQGVs only indicates that
further investigation is required to better quantify and understand the
environmental risk [93]. Local parameterisations reecting site-specic
exposures should be adopted to calculate tailored threshold values.
The pipeline mercury threshold data provided output comparable to
different measurement techniques currently used in the industry, such as
X-Ray uorescence, pigging dust analysis, or acid digests [4,94,95].
Previous publications measuring mercury in uncleaned pipelines have
found concentrations of 10 g/kg in pigging dust, 10–100 g/kg concen-
tration of scale on steel pipeline surfaces, or <5 g/m pipeline [31,96,
97]. Comparing these literature derived values to the threshold con-
centrations presented in Table 2, it may seem as though the published
values far exceed the threshold values used for the calculations pre-
sented. However, this does not take into consideration that the
literature-derived values were obtained from contamination product
analysis (such as measuring the mercury concentration in pigging dust
or cleaning solution), whilst computed values here are indicative of the
Fig. 5. Density distribution of mercury concentration in trophic levels of marine organisms sampled in the global oceans (data can be found in Table S6), compared
to the model-derived initial concentration in ‘Anchovy Bay’ (reference lines and text). Trophic levels listed in the facet labels on the right.
R. von Hellfeld et al.
Journal of Hazardous Materials 452 (2023) 131298
10
maximum allowable pipeline-associated mercury concentration
following cleaning.
A recent environmental plan submitted for the decommissioning of
the Grifn gas export pipeline on the Northwest shelf of Australia
calculated a pre-cleaning pipeline mercury concentration of 98 mg/kg.
Their studies investigating the efcacy of cleaning options suggests that
post-cleaning the pipelines will have a concentration range of
0.26–0.010 mg/kg. Based on the approach outlined in Section 2.1, this
would be unlikely to exceed the WQGVs or the higher SQGV [50]. These
ndings highlight the importance of mitigation actions such as pipeline
cleaning prior to in situ abandonment, where such actions are planned.
4.1.1. Requirements for the determination of site-specic threshold values
The approach outlined here highlights the gaps and limitations that
need to be addressed to improve the environmental relevance of the
derived threshold values. Key gaps include the consideration of existing
levels of contamination, the potential for dispersion and mixing of
released mercury, the extent to which mercury in bioavailable or
methylated, and the partitioning of mercury between the solid and
aqueous phase.
This approach assumes an uncontaminated environment for the
calculation of the pipeline threshold value. This may be more applicable
to regions of low environmental contamination levels such as the Baltic
Sea (0.6 ng/l water mercury concentration [98]) or the North Sea
(0.5 ng/l water mercury concentration [99]). Regions with elevated
background concentrations may require an adjusted approach to ensure
relevant environmental concentrations are below levels at which it may
cause harm.
Mixing and aqueous dispersion should be included in site-specic
applications of this approach. Quantifying sediment transport and
settling rates provide an estimate of solid-phase mixing, whilst the rate
of mercury dissolution should be based on laboratory experiments and
considered against ocean current-based dispersion.
Speciation is known to affect toxicity, and research on mercury
speciaition in marine waters suggests that only up to ~ 22 % of total
mercury is present as methylmercury [100], with the remainder being
inorganic mercury adsorbed to organic or particulate matter [60]. The
fractions could be estimated through laboratory experiments or in situ
measurements in scenarios reecting local conditions. It should be
considered, however, that mercury speciation may change over longer
time frames, so environmental transformations should be considered to
better understand exposures.
Describing the partitioning of mercury between solid and aqueous
phases is an important consideration to ensure the protection of marine
ecosystems. This was achieved in the present study by applying a K
d
Table 3
Total mercury concentration (mg/kg) in muscle tissue in all ’Anchovy Bay’
function groups in trial 1, and the percentage increase in muscle tissue mercury
concentration in trials 2–4 (cf Fig. S1).
Trial 1
Background
(mg/kg)
Trial 2
0.005 t/
km
2
/year (%
increase)
Trial 3
0.01 t/km
2
/
year (%
increase)
Trial 4
0.04 t/km
2
/
year (%
increase)
Whale 0.89
a
2.25 4.49 20.22
Seal 0.61
a
1.64 4.92 19.67
Cod 1.15
b
2.61 5.22 20.00
Whiting 1.23
b
3.25 5.69 21.14
Mackerel 0.95
a
3.16 5.26 21.05
Anchovy 0.22 4.55 9.09 22.73
Shrimp 0.06 0.00 16.67 16.67
Benthos 0.16 0.00 6.25 18.75
Zooplankton 0.06 0.00 16.67 33.33
Phytoplankton 0.00 0.00 0.00 25.00
a exceeds 0.5 mg/kg FS for low accumulation species.
b exceeds 1 mg/kg FS for high accumulation species.
Fig. 6. Weekly consumption of seafood (g/week) for different nations [72], each dot representing the data for a nation or the overall sh consumption of coastal
indigenous communities [74].
R. von Hellfeld et al.
Journal of Hazardous Materials 452 (2023) 131298
11
value. Future research should consider the use of regional and mercury-
species specic values where available. Other marine mercury K
d
values
published in the literature range from 10
4
to 10
6
l/kg [17,48,49,101], or
higher for species such as metacinnabar. Alternative approaches, such as
describing the ux and dispersion of mercury from sediment the over-
laying waters offer a more sophisticated approach but may have higher
data and experimental requirements [102].
4.2. B: food web modelling
The background mercury concentration presented here for some of
the species in the ‘Anchovy Bay’ were found to already exceed the FS.
Although this may not be representative of all marine food webs
currently, this is supported by e.g., the United Nations most recent
report on global mercury assessment, where the mean mercury con-
centration in almost all marine species examined in the published case
studies exceeded the 0.5 mg/kg FS. This held true for higher trophic
level sh, such as swordsh, when examined in different oceanic basins.
Additionally, all shark species examined in the report exceeded the
0.5 mg/kg FS, and 60% also exceeded the higher (1.0 mg/kg) FS [100].
This was supported by literature derived data from different environ-
ments (see Table S6). Moreover, the ICES data centre for North Sea
species found that species such as cod (Gadus morhua), common dab
(Limanda limanda), common mussel (Mytilus edulis) and manila clam
(Lajonkairia lajonkairii) presented with mercury concentrations that
exceeded at least one of the FS [103], with similar observations being
made by others. [23,91,104]. In the modelled approach, comparison of
modelled initial concentrations and literature derived data showed that
whilst some degree over- or under-estimation occurred within the
modelled species (Table S6), the concentrations still fell within the range
observed in biota samples (Fig. 5). Thus, the focus was placed on relative
changes in mercury biota concentrations with increasing contaminant
inux in the model environment rather than the actual concentration
itself.
The ‘Anchovy Bay’ model showed that with an increased mercury
inux of 0.5 %, 1 %, and 4 % into the model environment between trials
2–4, the functional groups accumulated, an average of 2.6 %, 5.1 %, and
20.5 % additional mercury in muscle tissue. This example highlights
that certain species with a population average below the FS, may exceed
them in the future, with mercury releases equivalent to the EQGVs.
Considering the EWI calculated in this example, the percentage in-
crease with increasing mercury inux into the system was more variable
than that observed for the muscle concentration (Table 4). Between
trials 2 and 4, the EWI increased by 0 %, 4.2 %, and 16.7 % for the non-
indigenous community, but followed the muscle tissue trajectory more
closely for coastal indigenous communities around the world (2.5 %,
5.2 %, and 20.5 %, respectively). The increased EWI by the non-
indigenous community was inuenced only by white sh (in trials 3
and 4) and oily sh (in trial 4), whilst the consumption of crustaceans
and marine mammals did not affect the overall EWI. Studies assessing
the EWI of different populations has found that parts of the Amazonas
[105], Finnish [106], Italian [107], Spanish [108], and Taiwanese [109]
populations may already be exceeding the TWI though the consumption
of seafood. This is unsurprising given the wide distribution of organism
mercury tissue concentrations (Fig. 5). There is a linear relationship
between the mercury added to the Anchovy Bay food web and the
percent increase in the sh functional groups but not the shrimp or
benthos. This increased to a maximum of 20–30 % increased methyl-
mercury intake compared to background levels.
It should be noted that in the present approach, the maximum po-
tential uptake of (methyl-)mercury by marine organisms from the
environment and humans through seafood consumption were calcu-
lated, thus presenting a conservative, i.e., ‘worst case scenario’. This
does not necessarily imply that a specic decommissioned offshore asset
poses a risk currently, but that evidence suggests the potential for it to do
so in the future. Such ndings further underline the need to address the
potential implications of environmental mercury release for both cur-
rent and future communities. While the environmental guideline values
may be protective of immediate adverse effects, already increasing
background concentrations of mercury highlight the need to further
determine the long-term accumulation and biomagnication potential
of mercury in different marine food webs.
4.2.1. Requirements for site-specic food web modelling to assess mercury
biomagnication and dietary exposure of humans
The approach described in this study uses a generic food web model
that does not represent any specic local ecosystem. Applications of this
method should thus incorporate site-specic data to improve the
representativeness of the food web and mercury input components [54].
This includes updating the food web structure and improving the mer-
cury uxes to and from the ecosystem (e.g., considerations discussed in
Section 4.1.1, along with mercury bioavailability, sediment sequestra-
tion, and biota excretion rates) [110,111]. Food web structure will have
an impact on the biomagnication of contaminants, but it is not clear
how different food webs will affect the transfer of mercury among or-
ganisms. Moreover, sub-acute chronic exposure to mercury is likely to
lead to adverse health effect on the organisms before reaching lethal
concentrations. In sh, such effects may be developmental alterations
during embryogenesis, as well as impeding the larval predator avoid-
ance and prey-capture abilities [112–116]. A summary of sub-acute ef-
fects observed in laboratory experiments with aquatic species can be
found in Table S7. However, only few publications examine adverse
effects in wild-caught species, thus highlighting the need for a better
understanding of how laboratory-derived results translate into envi-
ronmentally applicable insights. When assessing mercury in marine
mammals, most publications focus on measuring concentrations and
species in various organs and tissues [117]. Although marine mammals
accumulate the majority of methylmercury in the liver [118], it will also
accumulate in muscle tissue and the brain in the long run [119]. Inter-
estingly, these organisms can detoxify methylmercury by binding it to
selenium, forming the insoluble and non-toxic mercuric selenide crystals
[120]. However, this detoxication pathway can induce selenium de-
ciency, which in turn has deleterious health effects if it occurs over a
longer period [121]. Such interactions are not considered in many
models, including Ecotracer, further affecting the representative food
web biomass.
The toxicity of the accumulated mercury on the marine organisms
should be considered [54,122], and spatial considerations should be
incorporated, to reect the effects of e.g., foraging ranges or dispersion
and dilution from the contaminant source, to parameterise the model for
a specic site. This may include more rened calculations describing
Table 4
Weekly methylmercury intake based on the trial 1model output (µg/kg), and the
average seafood consumption for non-indigenous [72] and indigenous com-
munities [74]. For trials 2–4 the percentage increase in weekly methylmercury
intake from trial 1 is given.
Trial 1
Background
(µg/kg)
Trial 2
0.005 t/
km
2
/year
(% increase)
Trial 3
0.01 t/km
2
/
year (%
increase)
Trial 4
0.04 t/km
2
/
year (%
increase)
Crustacean 0.01 0.00 0.00 0.00
Oily sh 0.07 0.00 0.00 14.29
White sh 0.15
a
0.00 6.67 20.00
Marine
mammals
0.01 0.00 0.00 0.00
Total non-
indigenous
consumption
0.24
a
0.00 4.17 16.67
Total
indigenous
consumption
10.29
a
2.53 5.15 20.51
a exceeds TWI of 1.3 µg/kg [42].
R. von Hellfeld et al.
Journal of Hazardous Materials 452 (2023) 131298
12
mercury partitioning based on its transformation in local sediment,
sediment binding or sequestration, and demethylation. Such processes
may affect mercury biomagnication beyond the possible scope of the
generic example applied in this study. Some of these processes may be
addressed through integrating the spatial scale into Ecotracer by using
Ecospace in EwE, but other parameters must be addressed outside of the
model.
When calculating dietary methylmercury exposure, the potential
diversity of seafood sources and their preparation should also be
considered. Seafood is a globally traded commodity and thus species
caught from a contaminated food web may not affect the local com-
munity, or locally consumed sh may originate from a polluted food
web, even if the local ecosystem is pristine. Moreover, the work of
Bradley et al. [123] demonstrates that mercury bioaccessibility in sea-
food and its absorption by humans varies between species and method of
preparation. Thus, factors such a local preparation methods and species
preferences should be considered for local release scenarios. An attempt
of this has been made for the present example, but gaps remain in the
data for many commercially relevant seafood species.
Validating food web modelling is important, given the approach in
this study aims to predict future release scenarios. These include
deriving the concentration ratio between internal and environmental
contaminant concentration or used the median concentration value
where no concentration ratio could be derived. These values were then
assessed for distribution normality, and Ecotracer was used to predict
the observed versus predicted contaminant root mean square deviation
to understand whether Ecotracer successfully reproduced real world
concentration scenarios [66]. Others t the model through restricted
maximum likelihood, obtaining coefcient estimates and marginal
(representing the variance explained by xed factors) and conditional
(representing the variance explained by xed and random effects) R
2
values [124].
5. Summary
With many offshore petroleum production assets nearing the end of
their operational life, decommissioning activities will only increase.
Discussions about the way to decommission these assets need to consider
the impacts and risks to the environment, human safety, and cost.
Environmental guidelines are typically based on ecotoxicological data
from laboratory and eld-based studies. However, impacts such as
biomagnication and dietary exposure to humans also need to be
considered.
This desktop study aimed to demonstrate an approach to assess the
potential biomagnication impacts from mercury releases equivalent to
the EQGVs. In this demonstration, increased mercury input equivalent to
pipelines contaminated with mercury below levels that would exceed
EQGVs were found to increase marine organism tissue concentrations by
0–33 % (Table 3) and corresponding dietary methylmercury intake by
0–20 % (Table 4). This implies that further research is needed to be able
to better characterise biomagnication risks from residual mercury in
offshore oil and gas infrastructure decommissioned in situ. Environ-
mental managers and regulators should consider the range of recom-
mendations provided to tailor the approach outlined in this study to
local environmental and release scenarios. Considering the bio-
accumulation insights gained from this study, the applied software was
found to be an applicable and easy-to-use tool in marine contaminant
tracing and should be considered for site-specic assessments as a
complementary line of evidence to existing EQGVs.
Environmental implications
Past research has shown that mercury can accumulate in offshore
hydrocarbon pipelines, deeming such materials ‘hazardous’. The focus
of environmental risk assessments, however, is predominantly on direct
environmental impact and potential health implications for workers.
This research paper has highlighted the importance of also focussing on
potential future implications for food webs and the resulting impact this
might have on seafood consumers.
CRediT authorship contribution statement
Rebecca von Hellfeld: Conceptualization, Data curation, Formal
analysis, Investigation, Methodology, Software, Validation, Visualiza-
tion, Writing – original draft, Writing – review & editing. Christoph
Gade: Methodology, Writing – original draft, Writing – review & edit-
ing. Darren J. Koppel: Conceptualization, Investigation, Validation,
Visualization, Writing – review & editing. William J. Walters: Formal
analysis, Methodology, Software. Fenny Kho: Conceptualization,
Writing – review & editing. Astley Hastings: Conceptualization,
Funding acquisition, Investigation, Methodology, Project administra-
tion, Resources, Supervision, Writing – review & editing.
Funding
This research was funded by the National Decommissioning Centre
through University of Aberdeen. We also acknowledge the in-kind sup-
port from the Net Zero Technology Centre. Astley Hastings and Rebecca
von Hellfeld are further funded by the UK Research and Innovation
Energy Programme under Grant no. EP/S029575/1.
Declaration of Competing Interest
The authors declare the following nancial interests/personal re-
lationships which may be considered as potential competing interests:
Some authors have undertaken research funded by the oil and gas in-
dustry. These funding sources have not inuenced the approach or
conclusions of this study. The authors have no further potential conicts
of interest to disclose.
Data Availability
Data will be made available on request.
Acknowledgements
The authors also acknowledge the contribution of: (1) the Interna-
tional Council for the Exploration of the Sea (ICES) through the provi-
sion of the contaminants dataset [103], (2) the Centre for Environment,
Fisheries, and Aquaculture Science (CEFAS) through the provision of the
time trends and status of cadmium, mercury, and lead in sh and
shellsh dataset [91], and (3) the United States Environmental Protec-
tion Agency (US EPA) for the provision of the mercury in marine life
dataset [92]. The authors hereby acknowledge all institutions that have
contributed to the creation of the datasets used in the present
publication.
Appendix A. Supporting information
Supplementary data associated with this article can be found in the
online version at doi:10.1016/j.jhazmat.2023.131298.
References
[1] Bull, A.S., Love, M.S., 2019. Worldwide oil and gas platform decommissioning: a
review of practices and reeng options. Ocean Coast Manag 168, 274–306.
https://doi.org/10.1016/j.ocecoaman.2018.10.024.
[2] EU. 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); 2008. 〈https://eur
-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32008L0056&from=E
N〉.
[3] EU. Directive 2011/92/EU of the European parliament and the council of 13
December 2011 on the assessment of the effects of certain publica and private
R. von Hellfeld et al.
Journal of Hazardous Materials 452 (2023) 131298
13
projects on the environment; 2012. 〈https://eur-lex.europa.eu/legal-content/EN/
TXT/PDF/?uri=CELEX:32011L0092&from=EN〉.
[4] 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. Abu Dhabi Int Pet Exhib Conf
SPE. https://doi.org/10.2118/188801-MS.
[5] Baker S, Andrew M, Kirby M, Bower M, Walls D, Hunter L, et al. 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: Day 1 Tue, November 30,
2021. SPE; 2021. 〈https://doi.org/10.2118/208475-MS〉.
[6] Koppel, D.J., Kho, F., Hastings, A., Crouch, D., MacIntosh, A., Cresswell, T., et al.,
2022. Current understanding and research needs for ecological risk assessments
of naturally occurring radioactive materials (NORM) in subsea oil and gas
pipelines. J Environ Radiol 241, 106774. https://doi.org/10.1016/j.
jenvrad.2021.106774.
[7] Kho, F., Koppel, D.J., von Hellfeld, R., Hastings, A., Gissi, F., Cresswell, T., et al.,
2022. Current understanding of the ecological risk of mercury from subsea oil and
gas infrastructure to marine ecosystems. J Hazard Mater 438, 129348. https://
doi.org/10.1016/j.jhazmat.2022.129348.
[8] UN. United Nations convention on the law of the sea. Treaty Ser; 1982. 〈htt
ps://www.un.org/depts/los/convention_agreements/texts/unclos/unclos_e.pdf〉.
[9] United Nations, 1958. Convention of the continental shelf. Treaty Ser 499,
311–316.
[10] IMO. London Convention on the prevention of dumping of wastes and other
matter; 1972. 〈https://www.epa.gov/sites/production/les/2015–10/document
s/lc1972.pdf〉.
[11] UN. Minamata convention on mercury; 2015, 72. 〈https://www.mercuryconvent
ion.org/Portals/11/documents/Booklets/COP3-version/Minamata-Convention
-booklet-Sep2019-EN.pdf〉, [Accessed 23 December 2022].
[12] NOPTA. Guideline: offshore petroleum decommissioning; 2022. 〈https://www.
nopta.gov.au/_documents/guidelines/decommissioning-guideline.pdf〉,
[Accessed 24 August 2022].
[13] Wilhelm, S.M., Bloom, N., 2000. Mercury in petroleum. Fuel Process Technol 63,
1–27. https://doi.org/10.1016/S0378-3820(99)00068-5.
[14] Ryzhov, V. v, Mashyanov, N.R., Ozerova, N.A., Pogarev, S.E., 2003. Regular
variations of the mercury concentration in natural gas. Sci Total Environ 304,
145–152. https://doi.org/10.1016/S0048-9697(02)00564-8.
[15] EA. Mercury: Sources, pathways, and environmental data; 2019. 〈https://consult.
environment-agency.gov.uk/++preview++/environment-and-business/challen
ges-and-choices/user_uploads/mercury-pressure-rbmp-2021.pdf〉, [Accessed 24
August 2022].
[16] Gworek, B., Bemowska-Kałabun, O., Kije´
nska, M., Wrzosek-Jakubowska, J., 2016.
Mercury in marine and oceanic waters — a review. Water Air Soil Pollut 227.
https://doi.org/10.1007/s11270-016-3060-3.
[17] Benoit, J.M., Gilmour, C.C., Mason, R.P., Heyes, A., 1999. Sulde controls on
mercury speciation and bioavailability to methylating bacteria in sediment pore
waters. Environ Sci Technol 33, 951–957. https://doi.org/10.1021/es9808200.
[18] He, M., Tian, L., Braaten, H.F.V., Wu, Q., Luo, J., Cai, L.-M., et al., 2019.
Mercury–organic matter interactions in soils and sediments: angel or devil. Bull
Environ Contam Toxicol 102, 621–627. https://doi.org/10.1007/s00128-018-
2523-1.
[19] Amato, E.D., Simpson, S.L., Remaili, T.M., Spadaro, D.A., Jarolimek, C.V.,
Jolley, D.F., 2016. Assessing the effects of bioturbation on metal bioavailability in
contaminated sediments by diffusive gradients in thin lms (DGT). Environ Sci
Technol 50, 3055–3064. https://doi.org/10.1021/acs.est.5b04995.
[20] Xu, J., Bland, G.D., Gu, Y., Ziaei, H., Xiao, X., Deonarine, A., et al., 2021. Impacts
of sediment particle grain size and mercury speciation on mercury bioavailability
potential. Environ Sci Technol 55, 12393–12402. https://doi.org/10.1021/acs.
est.1c03572.
[21] Dai, S.-S., Yang, Z., Tong, Y., Chen, L., Liu, S.-Y., Pan, R., et al., 2021. Global
distribution and environmental drivers of methylmercury production in
sediments. J Hazard Mater 407, 124700. https://doi.org/10.1016/j.
jhazmat.2020.124700.
[22] NRC. Toxicological effects of methylmercury. Washington, D.C., D.C.: National
Academies Press; 2000. 〈https://doi.org/10.17226/9899〉.
[23] 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. https://doi.org/10.1007/s11270-015-2735-5.
[24] Kitamura, S., Miyata, C., Tomita, M., Date, S., Kojima, T., Minamoto, H., et al.,
2020. A central nervous system disease of unknown cause that occurred in the
minamata region: results of an epidemiological study. J Epidemiol 30, 3–11.
https://doi.org/10.2188/jea.JE20190173.
[25] Das, K., Debacker, V., Pillet, S., Bouquegneau, M., 2002. Heavy metals in marine
mammals. Toxicol Mar Mamm 135–167. https://doi.org/10.1201/
9780203165577.ch7.
[26] UN. Minamata Convention on Mercury; 2009, p. 1–30.
[27] Bank, M.S., 2020. The mercury science-policy interface: history, evolution and
progress of the Minamata Convention. Sci Total Environ 722, 137832. https://
doi.org/10.1016/j.scitotenv.2020.137832.
[28] Mason, R.P., Sheu, G.-R., 2002. Role of the ocean in the global mercury cycle.
Glob Biogeochem Cycles 16, 40-1–40–14. https://doi.org/10.1029/
2001GB001440.
[29] Wilhelm, S.M., 1999. Generation and disposal of petroleum processing waste that
contains mercury. Environ Prog 18, 130–144.
[30] Enrico, M., Mere, A., Zhou, H., Loriau, M., Tessier, E., Bouyssiere, B., 2020.
Methods for total and speciation analysis of mercury in the petroleum industry.
Energy Fuels 34, 13307–13320. https://doi.org/10.1021/acs.
energyfuels.0c02730.
[31] Wilhelm, S.M., Nelson, M., 2010. Interaction of elemental mercury with steel
surfaces. J Corros Sci Eng 13, 1–14.
[32] Yang, Y., Luo, X., Elsayed, Y.E.A.N., Hong, C., Yadav, A., Rogowska, M., et al.,
2021. Characteristics of scales and their impacts on under-deposit corrosion in an
oil production well. Mater Corros 72, 1051–1064. https://doi.org/10.1002/
maco.202012095.
[33] US EPA. Environmental factor guideline: marine environmental quality; 2016.
[34] Taljaard S. The development of a common set of water and sediment quality
guidelines for the coastal zone of the BCLME – Project BEHP/LBMP/03/04.
Stellenbosch; 2006. 〈https://www.ais.unwater.org/ais/aiscm/getprojectdoc.
php?docid=1728〉.
[35] ANZECC. ARMCANZ, Australian and New Zealand guidelines for fresh and
marine water quality. Volume 2. Aquatic ecosystems – rationale and background
information (chapter 8). National Water Quality Management Strategy; 2000.
[36] US EPA. Ambient water quality criteria for mercury [EPA/5-80-058].
Washington, D.C.; 1980. 〈https://www.epa.gov/sites/default/les/2019-03/d
ocuments/ambient-wqc-mercury-1980.pdf〉, [Accessed 19 January 2023].
[37] EC. Directive 2008/56/EC of the European Parliament and of the Council of 17
June 2008 establishing a framework for community action in the eld of marine
environment policy (Marine Strategy Framework Directive); 2008.
[38] SEPA. Supporting guidance (WAT-SG-53) Environmental quality standards and
standards for discharges to surface waters; 2020. 〈https://www.sepa.org.uk/medi
a/152957/wat-sg-53-environmental-quality-standards-for-discharges-to-surface
-waters.pdf〉.
[39] RSA DEA. South African water quality guidelines for coastal marine waters –
natural environment and mariculture use; 2018. 〈https://www.dffe.gov.za
/sites/default/les/docs/waterqualityguideline2018.pdf〉.
[40] ANZECC. ARMCANZ, Australian and New Zealand guidelines for fresh and
marine water quality. Volume 2. Aquatic ecosystems – ratinoale and background
information (chapter 8); 2000.
[41] Bloom, N.S., 1992. On the chemical form of mercury in edible sh and marine
invertebrate tissue. Can J Fish Aquat Sci 49, 1010–1017. https://doi.org/
10.1139/f92-113.
[42] EFSA, 2012. Scientic opinion on the risk for public health related to the presence
of mercury and methylmercury in food. EFSA J 10, 241. https://doi.org/
10.2903/j.efsa.2012.2985.
[43] FAO. WHO, Joint FAO/WHO expert committee on food additives, sixty-rst
meeting. In: Joint FAO/WHO Expert Committee on Food Additives. Rome; 2003.
〈ftp://ftp.fao.org/es/esn/jecfa/jecfa61sc.pdf〉.
[44] Brundtland. Report of the world commission on environment and development.
Oslo; 1987.
[45] EFSA. Conservative assumption. European Food Safety Authority; 2022.
〈https://www.efsa.europa.eu/en/glossary/conservative-assumption#:~:text=An
%20estimat%20that%20tends%20to,possible%20is%20taken%20into%20acco
unt〉, [Accessed 6 February 2023].
[46] Lyon, B.F., Ambrose, R., Rice, G., Maxwell, C.J., 1997. Calculation of soil-water
and benthic sediment partition coefcients for mercury. Chemosphere 35,
791–808. https://doi.org/10.1016/S0045-6535(97)00200-2.
[47] IAEA. Sediment distribution coefcients and concetration factors for biota in the
marine environment. Vienna; 2004. 〈https://www-pub.iaea.org/MTCD/Publ
ications/PDF/TRS422_web.pdf〉.
[48] Turner, A., Millward, G.E., Roux, S.M.Le, 2001. Sediment–water partitioning of
inorganic mercury in estuaries. Environ Sci Technol 35, 4648–4654. https://doi.
org/10.1021/es010933a.
[49] Fitzgerald, W.F., Lamborg, C.H., Hammerschmidt, C.R., 2007. Marine
biogeochemical cycling of mercury. Chem Rev 107, 641–662. https://doi.org/
10.1021/cr050353m.
[50] Advisian. Grifn gas export pipeline decommissioning environmental plan; 2022.
[51] Wickham, H., Averick, M., Bryan, J., Chang, W., McGowan, L., François, R., et al.,
2019. Welcome to the Tidyverse. J Open Source Softw 4, 1686. https://doi.org/
10.21105/joss.01686.
[52] R Core Team. R: a language and environment for statistical computing; 2019.
[53] Christensen, V., Walters, C.J., 2004. Ecopath with ecosim: methods, capabilities
and limitations. Ecol Model 172, 109–139. https://doi.org/10.1016/j.
ecolmodel.2003.09.003.
[54] Walters, W.J., Christensen, V., 2018. Ecotracer: analyzing concentration of
contaminants and radioisotopes in an aquatic spatial-dynamic food web model.
J Environ Radio 181, 118–127. https://doi.org/10.1016/j.jenvrad.2017.11.008.
[55] Christensen V, de la Puente S. Fish 501: Ecosystem modelling with Ecopath with
Ecosim; 2021.
[56] Mason RP, Fitzgeraldt WF. The distribution and biogeochemical cycling of
mercury in the equatorial Pacic Ocean; 1993.
[57] Mason RP, Sullivan KA. The distribution and speciation of mercury in the South
and equatorial Atlantic; 1999.
[58] Sunderland, E.M., Krabbenhoft, D.P., Moreau, J.W., Strode, S.A., Landing, W.M.,
2009. Mercury sources, distribution, and bioavailability in the North Pacic
Ocean: insights from data and models. Glob Biogeochem Cycles 23. https://doi.
org/10.1029/2008GB003425.
[59] Cossa, D., Heimbürger, L.E., Lannuzel, D., Rintoul, S.R., Butler, E.C.V., Bowie, A.
R., et al., 2011. Mercury in the Southern Ocean. Geochim Cosmochim Acta 75,
4037–4052. https://doi.org/10.1016/j.gca.2011.05.001.
R. von Hellfeld et al.
Journal of Hazardous Materials 452 (2023) 131298
14
[60] Gworek, B., Bemowska-Kałabun, O., Kije´
nska, M., Wrzosek-Jakubowska, J., 2016.
Mercury in marine and oceanic waters—a review. Water Air Soil Pollut 227.
https://doi.org/10.1007/s11270-016-3060-3.
[61] Paranjape, A.R., Hall, B.D., 2017. Recent advances in the study of mercury
methylation in aquatic systems. FACETS 2, 85–119. https://doi.org/10.1139/
facets-2016-0027.
[62] Neff, J.M., 2002. Mercury in the ocean. Bioaccumul Mar Org 103–130. https://
doi.org/10.1016/b978-008043716-3/50007-5.
[63] Zettlitzer, M., Kleinitz, W., 1997. Mercury in steel equipment used for natural gas
production - amounts, speciation and penetration depth. Oil Gas Eur Mag 23,
25.30.
[64] Bełdowska, M., Kobos, J., 2018. The variability of Hg concentration and
composition of marine phytoplankton. Environ Sci Pollut Res 25, 30366–30374.
https://doi.org/10.1007/s11356-018-2948-4.
[65] Cinnirella, S., Bruno, D.E., Pirrone, N., Horvat, M., ˇ
Zivkovi´
c, I., Evers, D.C., et al.,
2019. Mercury concentrations in biota in the Mediterranean Sea, a compilation of
40 years of surveys. Sci Data 6. https://doi.org/10.1038/s41597-019-0219-y.
[66] Booth, S., Walters, W.J., Steenbeek, J., Christensen, V., Charmasson, S., 2020. An
Ecopath with Ecosim model for the Pacic coast of eastern Japan: describing the
marine environment and its sheries prior to the Great East Japan earthquake.
Ecol Model 428, 109087. https://doi.org/10.1016/j.ecolmodel.2020.109087.
[67] 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. https://doi.org/10.1289/ehp.7603.
[68] Palomares MLD, Pauly D. SeaLifeBase; 2022. 〈www.sealifebase.org〉, [Accessed 5
September 2022].
[69] Froese R, Pauly D. FishBase; 2022. 〈www.shbase.org〉, [Accessed 5 September
2022].
[70] Cinnirella, S., Bruno, D.E., Pirrone, N., Horvat, M., ˇ
Zivkovi´
c, I., Evers, D.C., et al.,
2019. Mercury concentrations in biota in the Mediterranean Sea, a compilation of
40 years of surveys. Sci Data 6, 1–11. https://doi.org/10.1038/s41597-019-0219-
y.
[71] Li, M.-L., Gillies, E.J., Briner, R., Hoover, C.A., Sora, K.J., Loseto, L.L., et al., 2022.
Investigating the dynamics of methylmercury bioaccumulation in the Beaufort
Sea shelf food web: a modeling perspective. Environ Sci Process Impacts. https://
doi.org/10.1039/d2em00108j.
[72] FAO. Food balances (2010-). Licence CC BY-NC-SA-3/0-IGO; 2021.
〈https://www.fao.org/faostat/en/#data/FBS〉.
[73] Food Standards Agency. What’s an oily sh?; 2004. 〈https://webarchive.nation
alarchives.gov.uk/ukgwa/20101210005807/http://www.food.gov.uk/new
s/newsarchive/2004/jun/oilyfishdefinition〉, (Accessed 31 August 2022].
[74] Cisneros-Montemayor, A.M., Pauly, D., Weatherdon, L.V., Ota, Y., 2016. A global
estimate of seafood consumption by coastal indigenous peoples. PLoS One 11,
e0166681. https://doi.org/10.1371/journal.pone.0166681.
[75] Frodello, J.P., Rom´
eo, M., Viale, D., 2000. Distribution of mercury in the organs
and tissues of ve toothed-whale species of the Mediterranean. Environ Pollut
108, 447–452. https://doi.org/10.1016/S0269-7491(99)00221-3.
[76] Sergeant, D.E., Armstrong, F.A.J., 1973. Mercury in seals from Eastern Canada.
J Fish Res Board Can 30, 843–846. https://doi.org/10.1139/f73-142.
[77] Wang, R., Wong, M.-H., Wang, W.-X., 2010. Mercury exposure in the freshwater
tilapia Oreochromis niloticus. Environ Pollut 158, 2694–2701. https://doi.org/
10.1016/j.envpol.2010.04.019.
[78] Taylor, D.L., Calabrese, N.M., 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.
https://doi.org/10.1016/j.marpolbul.2017.10.089.
[79] Roveta, C., Pica, D., Calcinai, B., Girolametti, F., Truzzi, C., Illuminati, S., et al.,
2020. Hg levels in marine porifera of Montecristo and Giglio Islands (Tuscan
Archipelago, Italy). Appl Sci 10, 4342. https://doi.org/10.3390/app10124342.
[80] Rivera-Hern´
andez, J.R., Fern´
andez, B., Santos-Echeandia, J., Garrido, S.,
Morante, M., Santos, P., et al., 2019. Biodynamics of mercury in mussel tissues as
a function of exposure pathway: natural vs microplastic routes. Sci Total Environ
674, 412–423. https://doi.org/10.1016/j.scitotenv.2019.04.175.
[81] Bełdowska, M., Kobos, J., 2016. Mercury concentration in phytoplankton in
response to warming of an autumn – winter season. Environ Pollut 215, 38–47.
https://doi.org/10.1016/j.envpol.2016.05.002.
[82] Bełdowska, M., Kobos, J., 2018. The variability of Hg concentration and
composition of marine phytoplankton. Environ Sci Pollut Res 25, 30366–30374.
https://doi.org/10.1007/s11356-018-2948-4.
[83] Bradley, M.A., Barst, B.D., Basu, N., 2017. A review of mercury bioavailability in
humans and sh. Int J Environ Res Public Health 14, 169. https://doi.org/
10.3390/ijerph14020169.
[84] Wang, F., Lemes, M., Khan, M.A.K., 2011. Metallomics of mercury: role of thiol-
and selenol-containing biomolecules. In: Environmental chemistry and toxicology
of mercury. John Wiley & Sons, Inc., Hoboken, NJ, USA, pp. 517–544. https://
doi.org/10.1002/9781118146644.ch16.
[85] Lemes, M., Wang, F., 2009. Methylmercury speciation in sh muscle by HPLC-
ICP-MS following enzymatic hydrolysis. J Anal Spectrom 24, 663–668. https://
doi.org/10.1039/b819957b.
[86] Roos, D.H., Puntel, R.L., Lugokenski, T.H., Ineu, R.P., Bohrer, D., Burger, M.E.,
et al., 2010. Complex methylmercury-cysteine alters mercury accumulation in
different tissues of mice. Basic Clin Pharm Toxicol 107, 789–792. https://doi.org/
10.1111/j.1742-7843.2010.00577.x.
[87] Walpole, S.C., Prieto-Merino, D., Edwards, P., Cleland, J., Stevens, G., Roberts, I.,
2012. The weight of nations: an estimation of adult human biomass. BMC Public
Health 12, 439. https://doi.org/10.1186/1471-2458-12-439.
[88] NHS. Fish and shellsh; 2018. 〈https://www.nhs.uk/live-well/eat-well/food-type
s/sh-and-shellsh-nutrition/〉.
[89] UNEP. NCS, CSIR, Guidelines for the establishment of environmental quality
objectives and targets in coastal zones of the Western Indian Ocean (WOI) region.
Nairobi; 2009. 〈https://wedocs.unep.org/handle/20.500.11822/8760;jse
ssionid=A95BD988066FED6200E4DD0A53728D49, (Accessed 19 January
2023].
[90] ICES. Contaminants and biological effects; 2021.
[91] Nicolaus, E.E.M., Lyons, B., Miles, A., Robinson, C.D., Webster, L., Fryer, R., 2018.
Time trend and status for cadmium, mercury and lead in sh and shellsh.
https://doi.org/10.7489/12111-1.
[92] US EPA. Mercury in marine life database; 2003. 〈https://cfpub.epa.gov/si/si_publ
ic_record_report.cfm?Lab=OST&dirEntryId=58213〉, [Accessed 23 September
2022].
[93] ANZECC. ARMCANZ, National water quality management strategy – paper no. 4 –
Australian and New Zealand guidelines for fresh and marine water quality –
volume 1 – the guidelines. Australian Water Association, New South Wales; 2000.
〈https://www.waterquality.gov.au/sites/default/les/documents/anzecc-a
rmcanz-2000-guidelines-vol1.pdf〉.
[94] US EPA. Method 6200: eld portable x-ray ourescence spectrometry for the
determination of elemental concentration in soil and sediment; 2007.
[95] US EPA. Method 7473 (SW-846): mercury in solids and solutions by thermal
decomposition, amalgamation, and atomic absorption spectrophotometry; 1998.
[96] Silakorn, P., Chanvanichskul, C., Siangproh, W., Chailapakul, O.,
Boonyongmaneerat, Y., 2019. Alternative method for mercury detection using
square wave anodic striping voltammetry. In: SPE symposium: decommissioning
and abandonment. SPE. https://doi.org/10.2118/199215-MS.
[97] O’rear D, Grice KJ, Pradhan V, Dassey AJ, Hakam A, Tiwary A. Process, method,
and system for removing mercury from pipelines [W0/2016/154394]; 2016.
[98] Pempkowiak, J., Cossa, D., Sikora, A., Sanjuan, J., 1998. Mercury in water and
sediments of the southern Baltic sea. Sci Total Environ 213, 185–192. https://doi.
org/10.1016/s0048-9697(98)00091-6.
[99] Schmidt, D., 1992. Mercury in Baltic and North Sea waters. Water Air Soil Pollut
62, 43–55. https://doi.org/10.1007/BF00478452.
[100] UNEP. Global mercury assessment 2018; 2018, p. 1–81. 〈https://wedocs.unep.
org/bitstream/handle/20.500.11822/25462/GMA 2018-ReviewDraft_2505
18_CLEAN_SEC.pdf?sequence=1&isAllowed=y〉.
[101] Schartup, A.T., Balcom, P.H., Mason, R.P., 2014. Sediment-porewater
partitioning, total sulfur, and methylmercury production in estuaries. Environ Sci
Technol 48, 954–960. https://doi.org/10.1021/es403030d.
[102] Clarisse, O., Dimock, B., Hintelmann, H., Best, E.P.H., 2011. Predicting net
mercury methylation in sediments using diffusive gradient in thin lms
measurements. Environ Sci Technol 45, 1506–1512. https://doi.org/10.1021/
es102730n.
[103] ICES. Contaminants dataset. CEMP Assessment; 2022. 〈https://gis.ices.dk/geone
twork/srv/metadata/a5058fef-19fb-4ce9-8552-1b74e9199b9d〉, [Accessed 18
October 2022].
[104] Baeyens, W., Leermakers, M., Papina, T., Saprykin, A., Brion, N., Noyen, J., et al.,
2003. Bioconcentration and biomagnication of mercury and methylmercury in
North Sea and Scheldt Estuary sh. Arch Environ Contam Toxicol 45, 498–508.
https://doi.org/10.1007/s00244-003-2136-4.
[105] da Silva, S.F., Pereira, J.P.G., Oliveira, D.C., de, M., Lima, O., 2020.
Methylmercury in predatory and non-predatory sh species marketed in the
Amazon Triple Frontier. Bull Environ Contam Toxicol 104, 733–737. https://doi.
org/10.1007/s00128-020-02862-5.
[106] Karjalainen, A.K., Hallikainen, A., Hirvonen, T., Kiviranta, H., Knip, M.,
Kronberg-Kippil¨
a, C., et al., 2013. Estimated intake levels for Finnish children of
methylmercury from sh. Food Chem Toxicol 54, 70–77. https://doi.org/
10.1016/j.fct.2012.02.074.
[107] Giandomenico, S., Cardellicchio, N., Spada, L., Annicchiarico, C., Di Leo, A.,
2016. Metals and PCB levels in some edible marine organisms from the Ionian
Sea: dietary intake evaluation and risk for consumers. Environ Sci Pollut Res 23,
12596–12612. https://doi.org/10.1007/s11356-015-5280-2.
[108] Ortega-García, J.A., Rodriguez, K., Calatayud, M., Martin, M., V´
elez, D.,
Devesa, V., et al., 2009. Estimated intake levels of methylmercury in children,
childbearingage and pregnant women in a Mediterranean region, Murcia, Spain.
Eur J Pedia 168, 1075–1080. https://doi.org/10.1007/s00431-008-0890-z.
[109] Lin, P., Nan, F.H., Ling, M.P., 2021. Dietary exposure of the Taiwan population to
mercury content in various seafood assessed by a total diet study. Int J Environ
Res Public Health 18, 4–6. https://doi.org/10.3390/ijerph182212227.
[110] Rajeshkumar, S., Li, X., 2018. Bioaccumulation of heavy metals in sh species
from the Meiliang Bay, Taihu Lake, China. Toxicol Rep 5, 288–295. https://doi.
org/10.1016/j.toxrep.2018.01.007.
[111] Annabi, A., Said, K., Messaoudi, I., 2013. Cadmium: bioaccumulation,
histopathology and detoxifying mechanisms in sh. Am J Res Commun 1, 60–79.
[112] Weis, J.S., Weis, P., 1995. Effects of embryonic exposure to methylmercury on
larval prey-capture ability in the mummichog, Fundulus heteroclitus. Environ
Toxicol Chem 14, 153–156. https://doi.org/10.1002/etc.5620140117.
[113] Weis, J.S., 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. https://doi.org/10.1139/
f95-809.
[114] Weis, J.S., Weis, P., 1977. Effects of heavy metals on development of the killish,
Fundulus heteroclitus. J Fish Biol 11, 49–54. https://doi.org/10.1111/j.1095-
8649.1977.tb04097.x.
R. von Hellfeld et al.
Journal of Hazardous Materials 452 (2023) 131298
15
[115] del, M., Alvarez, C., Murphy, C.A., Rose, K.A., McCarthy, I.D., Fuiman, L.A., 2006.
Maternal body burdens of methylmercury impair survival skills of offspring in
Atlantic croaker (Micropogonias undulatus. Aquat Toxicol 80, 329–337. https://
doi.org/10.1016/j.aquatox.2006.09.010.
[116] Smith GM, Weis JS. Predator-prey relationships in mummichogs (Fundulus
heteroclitus (L.)): effects of living in a polluted environment; 1997.
[117] Evans, R.D., Hickie, B., Rouvinen-Watt, K., Wang, W., 2016. Partitioning and
kinetics of methylmercury among organs in captive mink (Neovison vison): a stable
isotope tracer study. Environ Toxicol Pharm 42, 163–169. https://doi.org/
10.1016/j.etap.2016.01.007.
[118] Kershaw, J.L., Hall, A.J., 2019. Mercury in cetaceans: exposure, bioaccumulation
and toxicity. Sci Total Environ 694, 133683. https://doi.org/10.1016/j.
scitotenv.2019.133683.
[119] Oliveira Ribeiro, C.A., Rouleau, C., Pelletier, ´
E., Audet, C., Tj¨
alve, H., 1999.
Distribution kinetics of dietary methylmercury in the Arctic Charr (Salvelinus
alpinus). Environ Sci Technol 33, 902–907. https://doi.org/10.1021/es980242n.
[120] Wagemann, R., Trebacz, E., Boila, G., Lockhart, W., 1998. Methylmercury and
total mercury in tissues of arctic marine mammals. Sci Total Environ 218, 19–31.
https://doi.org/10.1016/S0048-9697(98)00192-2.
[121] Ralston, N.V.C., Azenkeng, A., Raymond, L.J., 2012. Mercury-dependent
inhibition of selenoenzymes and mercury toxicity. In: Methylmercury and
neurotoxicity. Springer, US, Boston, MA, pp. 91–99. https://doi.org/10.1007/
978-1-4614-2383-6_5.
[122] Booth S, Steenbeek J, Charmasson S. ECOTRACER – a user’s guide to tracking
contaminants using the Ecopath with Ecosim (EwE) approach; 2020, 32.
[123] Bradley, M.A., Barst, B.D., Basu, N., 2017. A review of mercury bioavailability in
humans and sh. Int J Environ Res Public Health 14. https://doi.org/10.3390/
ijerph14020169.
[124] Buck, D.G., Evers, D.C., Adams, E., DiGangi, J., Beeler, B., Sam´
anek, J., et al.,
2019. A global-scale assessment of sh mercury concentrations and the
identication of biological hotspots. Sci Total Environ 687, 956–966. https://doi.
org/10.1016/j.scitotenv.2019.06.159.
R. von Hellfeld et al.
