Threshold values for the protection of marine ecosystems from NORM in subsea oil and gas infrastructure

This modelling study uses the ERICA Tool and Bateman ’ s equation to derive sediment threshold values for radiation protection of the marine environment relevant to NORM-contaminated products (radium-contaminated scales, 210 Pb films and 210 Po films) found in subsea oil and gas infrastructure. Threshold values are calculated as the activity concentration of the NORM-contaminated products ’ head of chain radionuclide (i.e., 226 Ra + 228 Ra, 210 Pb, or 210 Po) that will increase radiation dose rates in sediments by 10 μ Gy/h to the most exposed organism at a given release time. The minimum threshold value (corresponding to peak radiation dose rates from the ingrowth of progeny) were for radium-contaminated scales, 0.009 Bq/g of 226 Ra, 0.029 Bq/g of 228 Ra (in the absence of 226 Ra) or 0.14 Bq/g of 228 Ra (in the presence of 226 Ra), followed by 0.015 Bq/g for 210 Pb films, and 1.6 Bq/g for 210 Po films. These may be used as default threshold values. Added activity concentrations of the NORM-contaminated products to marine sediments below these threshold values implies a low radiological risk to organisms while exceedances imply that further investigation is necessary. Using contaminated product specific parameterisations, such as K d values derived for Ra from a BaSO 4 matrix in seawater, could greatly affect threshold values. Strong consideration should be given to deriving such data as part of specific radiological risk assessments for these products.


Introduction
Naturally occurring radionuclides are found at low activity concentrations in oil and gas reservoirs around the world (Smith, 2011).As fluids are extracted from reservoirs, radionuclides may become concentrated in contamination products including inorganic salt scales, films, sludges, and sands within infrastructure such as production pipelines (Nelson et al., 2016;Schmidt, 2000).These naturally occurring radioactive materials (NORM) may be recalcitrant and remain in oil and gas infrastructure at the cessation of operations.
A significant inventory of offshore oil and gas infrastructure is approaching the end of its productive life (Wood, 2017).Pipelines represent a large component of this infrastructure, with hundreds of thousands of kilometres of subsea pipelines laid around the world (Kaiser, 2018).Decommissioning options for this infrastructure include complete or partial removal or leaving it in situ.The reported benefits of leaving some infrastructure in situ include cost savings, improved health and safety outcomes for workers, and ecological benefits from the provision of productive artificial reef habitat (Bull and Love, 2019;McLean et al., 2022).However, questions remain about the long-term environmental consequence of contaminants, including NORM, in such infrastructure (MacIntosh et al., 2021;Melbourne-Thomas et al., 2021;Schläppy et al., 2021).
The disposal of residual NORM-contaminated products in oil and gas infrastructure to the marine environment via in situ abandonment is typically subject to regulatory oversight.For example, nations that are signatories to the Convention on the Prevention of Marine Pollution by Dumping of Wastes and Other Matter (London Convention) (IMO, 1972) have a requirement to ensure that any infrastructure approved to be abandoned in the marine environment has "de minimis" radioactivity levels (IAEA, 2016).Understanding the potential dose rates from NORM-contaminated products to marine organisms is therefore important to help clarify regulatory requirements.
Three important NORM-contaminated products have been identified in oil and gas systems, including radium-contaminated mineral scales, 210 Pb films, and 210 Po films (Koppel et al., 2022).Radium-contaminated scales occur when radium is precipitated as an inorganic salt, typically co-precipitated with barium sulfate (Ba(Ra)SO 4 ) (Grandia et al., 2008).These are precipitated from water so may occur wherever water from production fluids is transported.Two radium isotopes, 226 Ra and 228 Ra, contribute to these scales.Both isotopes have a common chemical behaviour; however, they have different decay chains and progeny, emission types and energies, and half-lives, so their radioecological impacts may be different. 210Pb films originate from either deposition from the decay of 222 Rn throughout treatment and transportation systems or electrodeposition of 210 Pb from fluids to surfaces such as wet parts of gas production systems (Schmidt, 2000;Smith, 2010).Unsupported 210 Po films are rarely reported but likely to originate where polonium partitions to lighter fractions such as ethane which are then separated and cooled in pipelines (Schmidt, 2000;Smith, 2010).More comprehensive reviews on the chemistry of NORM-contaminated product formation are provided by Nelson et al. (2015) and Schmidt (2000).
The risk of radionuclides to non-human biota is dependent on the amount of radiation that is absorbed by the organisms, termed the absorbed dose, and the radiosensitivity of the organism (ICRP, 2007).Radiation assessment tools have been developed to simplify calculations of absorbed dose rates to non-human biota and include the ERICA (Environmental Risk from Ionising Contaminants: Assessment and Management) Tool (Brown et al., 2008(Brown et al., , 2016) ) and RESRAD (Residual Radioactive) Biota code (US DoE, 2004).Calculated absorbed dose rates can be compared to various environmental dose rate reference levels that describe levels above which potential radiation effects in organisms may occur.Common dose rate reference levels include 10 μGy/h for initial screening and ecosystem-wide protection (Garnier-Laplace et al., 2008), 400 μGy/h benchmark dose rate for marine organisms (UNSCEAR, 2008), or the derived consideration reference levels (DCRL) of the International Commission on Radiological Protection (ICRP) of 40400 μGy/h for reference flatfish and seaweed or 400-4000 μGy/h for reference crab (ICRP, 2008).For more information on these values and their application see the review by Real and Garnier-Laplace (2020).
The 10 μGy/h screening value was derived using the same approach used to derive environmental quality standards for aquatic ecosystems in jurisdictions including Australia and New Zealand (ANZG, 2018), the United States of America (United States Environment Protection Agency, 1985), and the European Union (Directorate-General for Health and Food Safety (European Commission), 2017).In short, dose rates that elicit a 10% effect to an organism's health, for a range of species, are aggregated into a species sensitivity distribution.The dose rate protecting 95% of all species is then taken and an additional safety factor of 2 applied to account for limitations in input data (Garnier-Laplace et al., 2010).
The ERICA Tool uses the 10 μGy/h reference level as its default screening value.At Tier 1, the ERICA Tool defines Environmental Media Concentration Limits (EMCL), which are the activity concentration of a radionuclide in an environmental media (i.e.soil, sediment, freshwater, or seawater) that will lead to a 95th percentile dose rate of 10 μGy/h to the most exposed organism (Brown et al., 2008).These EMCL values allow for a rapid screening of radionuclide risk with minimal data needs.Scenarios where the summed quotient of radionuclide activity concentrations in the relevant environmental medium by their respective EMCL values is less than one are considered of negligible environmental concern with a high degree of confidence, due to the conservatism of their input parameters and comparison to dose rates where effects are known to occur (Larsson, 2008).
NORM-contaminated products in oil and gas infrastructure are unique because they form in disequilibria with their progeny, become unsupported when the extraction of oil and gas products cease, and are isolated from marine receptors until the containing infrastructure corrodes (Nelson et al., 2015).This means that the ingrowth of the progeny of the products needs to be considered on timescales commensurate to the corrosion of the infrastructure in the marine environment, which may be hundreds to thousands of years (Melchers, 2021).
This study derives NORM-contaminated product-specific threshold values for radium ( 226 Ra + 228 Ra) contaminated scale, 210 Pb films, and 210 Po films.Threshold values can be used to inform initial screening assessments of radiological impacts to marine organisms in ecological risk assessments of subsea oil and gas infrastructure.These thresholds are calculated as the maximum initial activity concentration of the NORM-contaminated product's head of chain radionuclides that when released to sediments will not exceed the screening dose rate criterion of 10 μGy/h.The influence of the NORM release time to the marine environment (i.e., radionuclide ingrowth and decay), and in the case of radium-contaminated scales the contribution of both 226 Ra and 228 Ra is investigated to account for temporal aspects of contaminant risk.These values may be used as screening values to demonstrate that levels of radioactivity in infrastructure are de minimis and thus suitable for in situ decommissioning, or whether further assessments are required.

Calculation of threshold values for NORM-contaminated products in marine sediment
A modelling approach was applied to calculate the activity concentrations of NORM-contaminated products that when released into marine sediments would result in a dose rate of 10 μGy/h to the most exposed marine organism, accounting for radionuclide ingrowth and decay over time.These activity concentrations are termed threshold values and were determined by combining ERICA Tier 1 calculations (and their related simplifications in dosimetry) with the Bateman Equation.
All decay constants were taken from ICRP Publication 107 nuclear decay data (Eckerman and Endo, 2008).Default EMCL values and the associated underlying parameter values including partition coefficients (K d ), concentration ratios (CR), organism occupancy factors, dose conversion coefficients (DCCs), and radiation weighting factors implemented in ERICA Tool 2.0 were used (Brown et al., 2016;Copplestone et al., 2013;ERICA Consortium, 2021).The most exposed organisms for the radionuclides investigated in this study were Phytoplankton ( 226 Ra, 210 Pb, 228 Ra, and 228 Th), and Sea anemones & True corals ( 210 Po).The EMCL values for these radionuclides are given in Table 1.
Calculations were conducted in the open source statistical software R version 4.0.5 (R Core Team, 2016), using the Tidyverse extension packages (Wickham and RStudio, 2021).R scripts implementing Equations (3), ( 6) and ( 8) and the production of risk quotient figures are provided in Supplementary Information S1.A table of the half-lives and decay constants are given in Supplementary Information S2.
ERICA Tier 1 EMCLs are calculated by Equation (1) where SDR is the A risk quotient was calculated following the Tier 1 approach in the ERICA Tool using Equation (3), where RQ t is the risk quotient for the NORM-contaminated product at time t, A n,t is the activity concentration of radionuclide n in the NORM-contaminated product at time t, and EMCL n is the EMCL for radionuclide n at a dose rate reference level of 10 μGy/h. (3) In this study the marine sediment EMCLs were used to reflect the solid-phase nature of NORM-contaminated products.When inputting radionuclide activity concentrations into the ERICA Tool, decay chains are truncated where progeny radionuclides' half-lives exceed 10 days based on the assumption that they are in equilibrium with their parent.This means that ERICA considers the decay for the head of chain radionuclides in NORM-contaminated products as shown in Fig. 1.While the input of short-lived progeny are simplified in this manner, the output of ERICA assessments includes their individual contribution to dose rates to organisms following the updated approach from ICRP Publication 136 (ICRP, 2017).
To calculate NORM-contaminated product threshold values for marine sediments, the radioactive decay and ingrowth equations were solved as follows.
The radioactive decay equation, Equation ( 4), gives the activity concentration of radionuclide n at time t: where A n,t is the activity concentration of radionuclide n at time t, A n,0 is the initial activity concentration of radionuclide n, and λ n is the decay constant for radionuclide n.Substituting Equation (4) into Equation (3) for a risk quotient of 1 and rearranging for the initial activity concentration gives Equation ( 5).This is the solution for a one-component decay chain, i.e., 210 Po films.
The Bateman Equation can be used to calculate the activity concentration of any radionuclide in a decay chain at time t given the starting activity concentration of its head of chain radionuclide and their decay constants (Bateman, 1910).The general derivation shown in Equation ( 6) indicates that radionuclide i decays into i+1 at the rate of λ i .J indicates the daughter radionuclide of i when n > 1. Equation ( 6) assumes that at time 0 only the head of chain radionuclides are present and that all decay chains are linear (i.e., the equation does not allow for branched decay).
The solution of Equation ( 6) for the second radionuclide in a decay chain at time t is given in Equation ( 7).
Substituting Equation ( 4) and Equation ( 7) into Equation ( 1) for a risk quotient of 1 and rearranging for the initial activity concentration of the head of chain radionuclide gives Equation ( 8), the function to calculate the threshold value a two-component decay chain, i.e. 210 Pb films.chain ( 226 Ra to 210 Pb to 210 Po, where short-lived progeny, i.e., with halflives <10 d, are assumed to be in equilibrium with these chain segment parents, Fig. 1).The third component can again be calculated using the Bateman Equation and combined with expressions for the first and second component, in Equation (3).This is shown in Equation ( 9).
The 228 Ra chain is a two-component decay chain and is solved using Equation ( 8).The risk quotient for the radium-contaminated scale can thus be expressed as shown in Equation ( 10).Equation ( 10) cannot be solved to calculate the initial activity concentration of 226 Ra and 228 Ra that would not exceed a risk quotient of 1 (as was done in Equation ( 5) and Equation ( 8)) because they can occur in different proportions in scale and have different decay rates.Therefore, threshold values were first calculated independently. 226Ra and 228 Ra will ingrow progeny and reach a peak in radioactivity at approximately 122 years and 4 years, respectively.Therefore, the maximum starting activity concentration of 228 Ra was calculated to give a risk quotient equal to 1 minus the risk quotient of 226 Ra at year 4 starting at its minimum head of chain activity.As 228 Ra decays faster, its contribution to the risk quotient of the radium scale will be insignificant at 122 years (corresponding to peak ingrowth of 226 Ra), meaning its contribution could be ignored.For a given time, the bracketed component of Equation (10) can be solved to give a single value for 226 Ra and 228 Ra components of the equation.This is shown for t = 4 years in Equation ( 11).The accuracy of these equations was confirmed using online calculators and the ERICA Tool (v2.0).The solutions of Equations ( 3), ( 6) and (7) for a given time were inputted to the World Information Service on Energy Uranium Project universal decay calculator (https://www.wise-uranium.org/rcc.html)to determine the activities of radionuclides in their decay chains at time t.These values were inputted to the ERICA Tool for a Tier 1 assessment and shown to give a risk quotient of 1.

Sensitivity analysis
The contribution of each radionuclide to the total dose rate to the two most exposed organisms for the NORM-contaminated products' head of chain radionuclides (Table 1), Phytoplankton and Sea Anemone & True corals, was investigated using a Tier 2 assessment in the ERICA Tool.Note that the output of ERICA assessments includes the contribution of all radionuclides, including those with half-lives <10 days.A sensitivity analysis was conducted using the Tier 3 functions of the ERICA Tool to investigate which parameter estimates contribute the greatest uncertainty to the total dose rate to each organism.To investigate how the provision of site-specific data affects threshold values, K d values from seawater leach experiments with radium-contaminated scale collected from a subsea oil and gas pipeline measured by Cresswell et al. (2021) and reported in Macintosh et al. (2022) were used.The K d values for radium, 1.5 × 10 6 L/kg, and polonium, 1.5 × 10 6 L/kg, were used to derive new EMCL values for radium-contaminated scales following the probabilistic Tier 1 approach specified in ERICA (Equation (1) and Equation ( 2)).The scale-specific EMCL values were inputted into Equation (10) and Equation ( 11) to calculate product-specific threshold values.
The K d value for polonium is based on measured solid and aqueous activity concentrations whereas the limit of detection for the analytical approach was used for the radium aqueous concentrations.This means that the radium K d value is a conservative estimate (i.e., will likely overestimate aqueous radium concentrations).Despite this, it is still 1000 times greater than the default value in ERICA of 2 × 10 3 L/kg.K d values were also reported for lead and thorium; however, all values were below the limit of detection which did not have the sensitivity to detect the expected aqueous concentrations based on their default K d values.For that reason, they were not used.

NORM-contaminated products and assessment assumptions
The three NORM-contaminated products assessed here are based on the types of NORM contamination reported in by-products in oil and gas systems (Ali et al., 2019;Koppel et al., 2022;Schmidt, 2000).In the assessment, the following assumptions are made about the products: 1. Contamination products are comprised of only their head of chain when oil and gas extraction has stopped.This assumption is used as a simplification as contamination products form to varying extents throughout the operational life of the oil and gas system (Yang et al., 2020).2. There is no loss of radionuclide progeny.This is a conservative assumption that is likely to be true for closed pipe decommissioning scenarios following the cessation of operations, but unlikely to be true for contaminated material in the marine environment where soluble components may emanate or leach.
D.J. Koppel et al. 3. The speciation of NORM-contaminated products does not change between operations ceasing and its release to the marine environment, such as following corrosive breakthrough of pipelines.Decommissioned pipelines that are filled with seawater and capped may lead to reducing conditions that could promote speciation changes of NORM-contaminated products, such as reductive dissolution of sulfate minerals.However, this is not well understood, and the speciation of the contaminants will change to reflect its local receiving environment.

Threshold values for NORM-contaminated products in marine sediments
Threshold values are defined as the initial head of chain radionuclide activity concentration in the NORM-contaminated product that will result in a 10 μGy/h dose rate to the most exposed organism in the marine environment at a given time of exposure (i.e., at a given age of the NORM-contaminated product when released into the marine environment).As radionuclides ingrow and decay over time in NORMcontaminated products, their corresponding dose rate to organisms will increase and decrease, as visualised in Fig. 2 for 226 Ra and 228 Ra.Therefore, it is important to consider radiation risks over time, particularly where the release of radionuclides will occur well after any decision about their disposal.
The results presented here provide for the ability to calculate appropriate threshold values for a given release time.Alternatively, a set of minimum threshold values are provided that are calculated from the highest radiation dose rate that NORM contamination will deliver to a marine organism at any time, reflecting the balance between the increase in dose rate from radionuclide in-growth and the decrease in dose rate from the decay of the parent radionuclide (i.e., the peak of the sum curve in Fig. 2).

Radium-contaminated scale
The minimum threshold value for 226 Ra was 0.009 Bq/g and for 228 Ra was 0.029 Bq/g, when considered independently (Table 2).A key difference in the temporal extent of the risk of 226 Ra and 228 Ra relates to their half-lives.The ingrowth and decay of the 228 Ra series is controlled by the 5.8-year half-life of 228 Ra, which is much shorter than the 1600year half-life of 226 Ra.This means that the 228 Ra series peaks and decays before the 210 Pb component of the 226 Ra series comes to equilibrium.As a result, 226 Ra reaches the peak of its ingrowth at 122 years compared to 4 years for 228 Ra (Fig. 2).
It is likely that 226 Ra and 228 Ra will co-occur in a radiumcontaminated scale, given their identical chemical behaviour (Nelson et al., 2015).As 226 Ra has a longer half-life and radioactive peak from the ingrowth of 210 Pb (Fig. 2) the joint risk quotient was determined to ensure that 228 Ra when in the presence of 226 Ra will not exceed a risk quotient of 1. 226 Ra with an activity concentration of 0.009 Bq/g will contribute a risk quotient of 0.52 after 4 yearscorresponding to peak radioactivity from 228 Ra ingrowth.Therefore, the maximum initial  a Note that the contribution of 228 Ra is minimal at this stage due to its relatively short half-life of 5.8 years.
D.J. Koppel et al.  228 Ra activity concentration that would give a risk quotient of 0.48 (calculated as 1 minus 0.52, the contribution of 226 Ra) was calculated to be 0.014 Bq/g.The impact of 228 Ra on the radioactivity peak from 226 Ra ingrowth is not necessary because of the relatively short half-life of 228 Ra.I.e., a starting 228 Ra activity concentration of 0.014 Bq/g would contribute a risk quotient <0.001 after 122 years.For release scenarios where 226 Ra activity concentrations will be less than 0.009 Bq/g, a greater amount of 228 Ra may be permissible than 0.014 Bq/g.More details on these results can be found in the code provided in Supplementary Information S1.

228
Ra and its progeny are decayed >95% after 25 years meaning that on the time scale of pipeline corrosion rates (10s to >1000 of years (Melchers, 2021)) it is unlikely to be present in significant quantities to meaningfully contribute much radiation dose rate if it is isolated from environmental receptors in infrastructure (Fig. 2).In comparison, it will take approximately 8000 years for 226 Ra to decay >95% of its initial activity meaning that the risk to the environment will be relevant on timescales relevant to pipeline corrosion.The consequence of these differences depends on the different 226 Ra to 228 Ra ratio and age of the NORM-contaminated product when released to the marine environment.At proportions of 226 Ra <0.4, or release times <20 years the 228 Ra ingrowth peak will result in a higher sum of risk quotient peak than the 226 Ra ingrowth peak (see Fig. 2).This will reduce the 'acceptable' 226 Ra activity concentration to <0.009 Bq/g, and result in the 226 Ra ingrowth peak being well below a risk quotient of 1.These temporal risk considerations will have to be considered if NORM-contaminated products are to be released at or shortly after operations cease, such as if they are stored open to the marine environment (e.g., uncapped or cut pipelines).

210 Pb and 210 Po films
The minimum threshold value for 210 Pb films was 0.015 Bq/g at 0 years due to the low EMCL of 210 Pb (0.015 Bq/g) and high EMCL of its progeny 210 Po (1.6 Bq/g) (Table 2).That is, the ingrowth of 210 Po was less impactful to the risk quotient compared to the decay of 210 Pb.This means that the risk from 210 Pb films decreases at a rate proportional to the half-life of 210 Pb (22.2 years).210 Po films have the highest EMCL of the radionuclides investigated in this study, a short half-life (t 1/2 = 138 d) relative to 210 Pb, 228 Ra, or 226 Ra, and no radioactive progeny.The minimum threshold value was equal to its EMCL at 1.6 Bq/g which increases at a rate proportional to the half-life of 210 Po.This means that unsupported 210 Po is unlikely to be a radionuclide of concern in decommissioning contexts where operations have ceased for more than a few years, as shown in Fig. 3.This does not diminish the need to consider the risk of 210 Po to human health and safety where public or occupational exposures are possible.

Sensitivity of radioecology parameters
The dose rate to the organisms Phytoplankton and Sea anemones & True corals were explored by a Tier 2 assessment.All NORMcontaminated products were investigated, and input activity concentrations were equal to those present in the NORM-contaminated products at the age that results in the minimum threshold value as defined in Table 2. Dose rates by NORM-contaminated product, organism, and radionuclide are given in Supplementary Information S3.In short, external dose rates are a negligible contribution to the total.For phytoplankton, 228 Ac contributes the greatest dose in scale containing 226 Ra and 228 Ra (~75% of the total dose rate) followed by 228   D.J. Koppel et al. approximately equivalent dose rates to phytoplankton.
A sensitivity analysis was undertaken using the Tier 3 function of the ERICA Tool.Input radionuclides were the 226 Ra decay series ( 226 Ra, 210 Pb, 210 Po) or both the 226 Ra and the 228 Ra series ( 228 Ra and 228 Th) and both Phytoplankton and Sea anemones & True Coral were investigated.The parameter and radionuclide pairs with the greatest Pearson correlation coefficients (positive or negative) are tabulated in Supplementary Information S4.Generally, CR values contribute the greatest variability with correlations between the parameter value and total dose rate to organism greatest for phytoplankton.For example, for phytoplankton exposed to scale with 226 Ra and 228 Ra, the Th CR was the most impactful parameter with a positive coefficient of 0.33 followed by the K d for Th with a negative coefficient of − 0.11.For exposures of 226 Ra in scale to phytoplankton, the CR of Pb was the most impactful parameter with a coefficient of 0.45 followed by K d for Pb at − 0.07.The relatively low sensitivity of K d values relative to CR values likely reflects the approach adopted to create a distribution around the K d estimatewhere the 5th and 95th percentile values of the distribution are set as 10x lower and higher than the point estimate, respectively (Brown et al., 2016;IAEA, 2004a).This is a much narrower range of potential values than exists for CR values.It is unlikely that this range fully accounts for the possible partitioning of radionuclides from NORM-contaminated products.For example, mineral scales and their associated radionuclides may become more soluble in anoxic conditions (Phillips et al., 2001).
To understand how using NORM-contaminated product specific K d values will affect EMCLs and TVs, radium-contaminated scale specific Ra and Po K d values were used.The scale-specific K d values were derived from a 30-day seawater leaching experiment of scale retrieved from a subsea oil and gas pipeline.These K d values were used to derive new EMCL values using the same probabilistic Tier 1 approach in ERICA (all distribution parameters used to recalculate EMCL values are given in Supplementary Information S5).The use of scale-specific K d values greatly increased the EMCL values for 226 Ra and 228 Ra, increasing from 0.02 Bq/g to 16.3 Bq/g for 226 Ra and increasing from 0.094 Bq/g to 70.2 Bq/g for 228 Ra.The EMCL for 210 Po decreased from 1.6 Bq/g to 0.105 Bq/g (Table 3).The distribution of F values is given in Fig. 4.
The minimum TVs recalculated using the scale-specific K d values for Ra and Po did not have the same substantial changes as the recalculated EMCL values (Table 4).For example, the minimum TV for 226 Ra increased from 0.009 Bq/g to 0.015 Bq/g whereas the EMCL increased from 0.02 to 16.3 Bq/g (Table 4).This is likely because 210 Pb and 228 Ac contribute the largest doses from 226 Ra and 228 Ra decay chains, respectively, meaning that the ingrowth of progeny may be more important to the dose rate than the radium isotopes themselves (Supplementary Information S3).K d values from the MacIntosh et al. ( 2022) study could not be calculated for all radionuclides detected in radium-contaminated scale due to the detection limits of the analytical approach (Cresswell et al., 2021).More detail is provided in Supplementary Information S5.However, the change in EMCL values for radium isotopes and 210 Po demonstrate the benefit of using parameterisations relevant to the exposure scenario.
Unexplored in this study are other contaminant and site-specific considerations that may increase or decrease NORM bioavailability and mobility.Default K d and CR parameter values assume radionuclide activities in the environment at an equilibrium between sediments, waters, and organisms.This assumption may not be true for point sources of contamination with different chemistries to environmental matrices such as radionuclides originating from NORM-contaminated products (Periáñez et al., 2018).This is clearly the case for radium, which is known to be highly insoluble in oxic seawater due to the low solubility product of RaSO 4 (log 10 K sp of − 10.24) (Brown et al., 2022).This gives greater context to the high K d value calculated from seawater leach experiments of radium in radium-contaminated barite scale, 1.5 × 10 6 L/kg, which is ~800 times greater than the default Ra K d value of 2 × 10 3 (IAEA, 2004a).
Radioecology and radioecotoxicology data are limited for marine organisms compared to terrestrial and freshwater organisms.Seawater leaching tests to calculate K d values are inexpensive and so should be actively considered by those managing infrastructure containing NORMcontaminated products.

Comparison to published radiation criteria and background activities
A comparison of the threshold values derived here to activity concentrations in NORM-contaminated products, background sediments (Koppel et al., 2022), and other criteria for radiation protection (Real and Garnier-Laplace, 2020), suggests that these derived threshold values are conservative.However, these values: are applied in addition to background radionuclide activity concentrations; account for the ingrowth of progeny over the radiological life of the NORM-contaminated product; do not incorporate dispersion or dilution of the NORM-contaminated products in the marine environment; and are proposed to be used as part of a graded approach to radiological protection.
The derived threshold values for radium-contaminated scales and 210 Pb films are much lower than the commonly applied exemption criterion for NORM of 1 Bq/g (IAEA, 2014a).Various national jurisdictions apply the 1 Bq/g criterion to NORM-contaminated products from oil and gas extraction to identify material subject to regulatory control (Loy, 2015).This study suggests that the 1 Bq/g criterion may not be suitably protective of the marine ecosystem for this exposure scenario.For example, the peak risk quotient for NORM-contaminated product with a starting activity concentration of 1 Bq/g in sediments will be 111 for 226 Ra at 122 years or 68 for 210 Pb at 0 years.For 226 Ra, risk quotients will persist above a value of 1 for 100-1000s of years (Fig. 3).
Threshold values that reflect release scenarios commensurate with corrosion times for pipelines would be much higher, particularly for 210 Po and 228 Ra decay series, which will have undergone significant decay.For example, for a release time 100 years after contaminatedproduct formation and oil and gas operations ceasing there may be no need to consider 210 Po films or 228 Ra in scale as many half-lives of decay will substantially reduce their activity concentration (Table 2).Where the corrosion of pipelines is expected to take >200 years, there may be no need to consider the risk of 210 Pb films.However, this should be considered carefully against expected corrosion timeframes.

Environmental management of NORM-contaminated products in subsea oil and gas infrastructure
The approach adopted in this study solves a challenge faced by the oil and gas industry and their regulators around the need to understand the radiological risk of NORM-contaminated products to the marine environment over their radiological life.Coupling risk and ingrowth models for exposure scenarios where the radioactive source is not being regenerated, and is at disequilibria from its progeny, has been adopted in  (Galloway et al., 2020).

Applying the threshold values
The threshold values derived here represent the head of chain activity concentration for NORM radionuclides in marine sediments that will prevent exceedance of the screening dose rate of 10 μGy/h and thus pose a negligible radiological risk to the marine ecosystem.These values represent activity concentrations that may be added to the natural background activity concentration of marine sediments, rather than limits on the activity concentrations of the NORM-contaminated products themselves.As such, threshold values could be applied to predicted sediment activity concentrations following the products' dispersion and mixing in the environment.
Exceedances of threshold values do not indicate that there will be radiological effects to organisms.Rather, exceedances should warrant a more detailed investigation.This may include incorporating site-specific data that can be used to better quantify radionuclide partitioning and bioavailability for a specific environment.This could then be assessed against other environmental dose rate reference levels such as the ICRP DCRL bands of 40-400 μGy/h for flatfish and seaweed (ICRP, 2008) or the UNSCEAR marine benchmark dose rate of 400 μGy/h (UNSCEAR, 2008).
For the context of offshore decommissioning, the use of the lower level of a relevant DCRL band (e.g.40 μGy/h for flatfish and seaweed) may be particularly appropriate given the definition of de minimis radiation levels from IAEA (2016) for the London Convention.De minimis subsumes the IAEA criteria of exemption and exclusion (IAEA, 2004b).As it relates to NORM contamination in subsea oil and gas infrastructure, a release of NORM-contaminated products that would substantially increase radioactivity at the site requires a specific assessment for marine flora and fauna protection.In the specific assessment, DCRL are used as the radiological criteria for marine biota.Threshold values derived here can be recalculated for a DCRL of 40 μGy/h by multiplying by 4.

The need for a holistic approach to radiological assessments
Only exposure scenarios where NORM is released to the marine environment were considered here.NORM-contaminated products may also pose a risk to sessile organisms by external-only radiation exposures.For example, the radionuclides 226 Ra, 214 Bi, 214 Pb of the 226 Ra decay chain and 228 Ac, 212 Bi, 212 Pb, 224 Ra, and 208 Tl from the 228 Ra decay chain have gamma emissions with emission probabilities >1% and energies >100 keV (Supplementary Information S6), which may result in a radiological dose rate to organisms colonising the external surfaces of contaminated pipelines (MacIntosh et al., 2022).
The derived threshold values represent a single line of evidence, radiological contamination of sediments, based on an impact of dose rates to marine organisms.A holistic understanding of all impacts and risks from a decommissioning scenario should be considered in an ecological risk assessment (Chapman et al., 2002).This should include the risk of other contaminants, such as mercury (Kho et al., 2022), plastics (Testoff et al., 2022), steel corrosion by-products, and mixtures therein (Koppel et al., 2018).Nonetheless, the derived threshold values provide a quick assessment method to determine whether NORM-contaminated products from oil and gas infrastructure pose a negligible radiological risk for a given release scenario.Importantly, this assessment aligns with IAEA and ICRP recommendations for radiological protection of the environment in a planned exposure scenario (IAEA, 2014b;ICRP, 2014).This approach may also be useful in determining the suitability of disposing NORM-contaminated infrastructure at sea under the 'de minimis' standard of the London Convention and Protocol.

Table 4
Minimum threshold values re-calculated using scale-specific K d values for Ra and Po.Age at minimum threshold value (corresponding to the peak radiological dose to the most exposed organism balancing radionuclide ingrowth and decay) is given in brackets.

Case study
To illustrate how these threshold values can be applied, two fictional case studies are considered here: (1) an export gas pipeline containing films of 210 Pb at 10 Bq/g; and (2) a flexible flowline containing radiumcontaminated scale in a barite matrix at an activity concentration of 5 Bq/g 226 Ra and 5 Bq/g 228 Ra.In these hypothetical examples, an oil and gas operator is investigating the possibility of decommissioning their pipeline containing residual NORM contamination by leaving them in situ.The proposed assessment approach is visualised in Fig. 5.

Lead film in gas export pipeline
Gas export pipelines may carry traces of 222 Rn which decay into the longer lived 210 Pb (t 1/2 = 22 years) leading to the formation of thin lead films on the internal surface of pipes.For this hypothetical case study, a 10 Bq/g 210 Pb film has been measured in a pipeline that will experience corrosive breakthrough (the point where pipeline corrosion leads to the ingress of seawater) after 200 years.
Step 1: The TV d for 210 Pb films is 0.015 Bq/g (Table 2).As the activity concentration of 210 Pb in the film is greater than the TV d the assessment moves to step 2.
Step 2: Gas-export pipelines are often made of carbon steel.Typical corrosion rates mean that corrosive breakthrough may occur after approximately 200 years.This is the point where corrosion penetrates the pipe wall allowing seawater/sediment to contact the scale contamination.For a real scenario, an understanding of the pipeline material and its corrosion rates in its environment would be necessary.
The TV R for 210 Pb with a release time of 200 years is 7.5 Bq/g (i.e., the solution for Equation (8) at t = 200 years).The 210 Pb film activity concentration is above the TV R the assessment moves to step 3.
Step 3: A 10 Bq/g film requires a dilution factor of 1.3 to reduce to 7.5 Bq/g.A highly conservative mixing factor could be 1:1 pipeline to sediment, with the lead film being a small component of overall pipeline mass.This suggests that the resulting release to the marine environment is unlikely to lead to dose rates >10 μGy/h.A persuasive argument could thus be made that the radiological component of the ecological risk assessment for the release of NORM from this pipeline is low.

Radium-contaminated scale in production flowlines
Flexible production flowlines can connect subsea wells to floating production storage and offloading vessels.They carry raw well fluids and so may accumulate NORM residues over their production life.
Step 1: The TV d for 226 Ra is 0.009 Bq/g and 228 Ra is 0.014 Bq/g (Table 2).As the activity concentration of the radium scale is greater than the TV d the assessment moves to step 2.
Step 2: Flexible flowlines often contain coatings and corrosion resistant steels.So corrosive breakthrough is expected after approximately 1000 years.
The release-time adjusted screening value, TV R , is 0.013 Bq/g for 226 Ra and >1000 Bq/g for 228 Ra due to its 5.8-year half-life (i.e. the solutions to Equation (9) and Equation (8), respectively).This means that the time-adjusted screening value TV R is now 0.013 Bq/g for 226 Ra and 228 Ra no longer needs to be considered.
As the activity concentration of 226 Ra in the radium scale is greater than the TV R the assessment moves to step 3. Fig. 5.A proposed implementation of a graded approach to assess NORM risk in the marine environment using threshold values derived in this study.Risk refers to the likelihood of dose rate exceedances of a dose screening level to marine organisms.NORM may have other impacts to marine ecosystems that should be considered holistically in an ecological risk assessment.IAEA (2016) Specific Assessment refers to the assessment provided by IAEA-TECDOC-1759 Determining the Suitability of Materials for Disposal at Sea under the London Convention 1972 and London Protocol 1996: A Radiological Assessment Procedure.

Fig. 1 .
Fig.1.Decay chains for NORM-contaminated products based on the ERICA Tool.The dose contribution of progeny radionuclides with a half-life of less than 10 days is incorporated into the DCC of its parent radionuclide based on an assumption of secular equilibrium.This is shown for each component of the decay chains for 226 Ra, 228 Ra, 210 Pb, and 210 Po.Halflives are given in years (y), days (d), hours (h), minutes (m) and seconds (s).Note that 220 Rn and 222 Rn are not included in the ERICA assessment as their primary exposure route is by inhalation.

Fig. 2 .
Fig. 2. The changing risk profile of (a) 226 Ra, (b) 228 Ra, and (c) 226 Ra + 228 Ra following their ingrowth and decay over time.The contribution of long-lived progeny (t 1/2 >10 d) are shown while short-lived progeny (t 1/2 <10 d) are accounted for by their first parent with a t 1/2 >10 d as per Brown et al. (2008).The starting activity concentration is 0.009 Bq/g for 226 Ra and either 0.029 Bq/g for 228 Ra in the absence of 226 Ra or 0.014 Bq/g for 228 Ra in the presence of 226 Ra. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) each).In the absence of 228 Ra, 210 Pb and 226 Ra contribute

Fig. 3 .
Fig. 3. Temporal patterns of risk for (a) 226 Ra, (b) 228 Ra, (c) 210 Pb, and (d) 210 Po at different initial activity concentrations and times.Risk quotients were determined as the sum of the quotient of each radionuclide in the products' decay chains at a given time by their EMCL for a 10 μGy/h dose rate.Default parameters were used to calculate EMCLs in the ERICA Tool (v2.0).Note the different ranges for the age of the NORM contaminant (y axis).

Fig. 4 .
Fig. 4. Probability density distributions of F values calculated using default parameters and radium-contaminated scale specific K d values for Ra and Po taken from MacIntosh et al. (2022).Dashed lines indicate the 95th percentile of the lognormally distributed F values calculated based on Monte Carlo simulations of the K d and CR values lognormal distributions.The resulting EMCL from this distribution is given in text adjacent to the lines (calculated by Equation (1)).

Table 1
The most exposed organism and EMCL values for the head of chain radionuclides in NORM-contaminated products.
4. The values derived are for sediment concentrations and do not incorporate any mixing of the NORM-contaminated product into the sediment.Application of these values to real world scenarios should account for a conservative dilution of contaminated material in the environment.Dispersion modelling or other approaches should be considered to justify a selected mixing scenario.5.The ERICA Tool includes default model parameters for radionuclide partitioning between sediments and seawater (K d values) and from waters to biota (CR values).There is no CR value for 226 Ra, 210 Pb, or 210 Po to Sea anemones and True coral, so values from a taxonomically similar organism (polychaete worm) were used in line with the recommendations of Hosseini et al. (2008).Additionally, the default K d value for polonium is based on a 'periodically adjacent element' (IAEA, 2004a).More information about the limitations of default ERICA parameter values for this exposure scenario can be found in recent publications from Koppel et al. (2022) and Macintosh et al. (2022).

Table 2
Threshold values (TV) for the head of chain radionuclides of different NORMcontaminated products in sediments that would result in the most exposed marine organism receiving a dose rate of 10 μGy/h.All values reflect initial activity concentrations, and do not consider mixing and dispersion processes.

Table 3
Default EMCLs, recalculated EMCLs using default parameters, and recalculated EMCLs for 226 Ra, 210 Po, and 228 Ra using Ra and Po K d values from radiumcontaminated scale leach experiments.Lognormal distribution parameters are given in Supplementary Information S5.Scale-specific EMCLs were not calculated for 210 Pb and 228 Th because appropriate updated K d values were not available.