- Top of page
- CONCLUSIONS AND FURTHER REFLECTIONS
Recent developments within the field of radiological protection have seen the expansion of the system for humans to one that includes protection of the environment itself against the detrimental effects of radiation exposure. This is evidenced by the latest recommendations of the International Commission on Radiological Protection (ICRP) that provide for the explicit aim of environmental protection by preventing or reducing the frequency of deleterious radiation effects to a level where they would have a negligible impact on the maintenance of biological diversity, the conservation of species, or the health and status of natural habitats, communities, and ecosystems (ICRP 2007). For human health, the aims of the ICRP remain the management and control of exposures to ionizing radiation such that deterministic effects are prevented, and the risks of stochastic effects are reduced to a reasonably achievable level (ICRP 2007). Although a fully developed framework for integration of human and ecological risk assessment exists for chemicals (WHO 2001; Suter et al. 2003), development of such an approach for radionuclides is at an early stage (Pentreath 2002; Brownless 2007; Copplestone et al. 2010).
As a further step in moving toward assessing the practicability of this integration goal, it is appropriate to examine, compare, and critically contrast various facets of the methodologies applied in human and environmental assessments, issues that are elaborated further on in the present work. The ERICA Integrated Approach (Larsson 2008) has been widely applied in environmental impact assessments required for authorized discharges (Hosseini et al. 2011; Nedveckaite et al. 2011) and radioactive waste disposal (Robinson et al. 2010). Within ERICA, environmental risk is quantified by calculating exposure for nonhuman biota using data on environmental transfer and dosimetry. Calculated exposures are then compared, at the screening tiers, to a predicted-no-effect dose rate to remove from further investigation those situations where the environmental impact is likely to be of negligible concern. The necessarily large data sets underpinning this assessment approach have led to the development of a supporting computer based tool—the ERICA Tool (see Brown et al. 2008). Components of the ERICA Tool (transfer and dosimetry) have been evaluated through comparison with other models and by limited direct comparison with measured data, results of these evaluations being found elsewhere (Beresford et al. 2008a, b2008b). These exercises have included a comparison of the ERICA Tool with the equally comprehensive RESRAD-Biota methodology (USDOE 2002) and demonstrate that for certain assessment components, particularly those considering transfer, differences in output can be substantial. Such observations highlight the uncertainties associated with these types of environmental impact modeling approaches.
The manner in which internal dose–rates are derived differs substantially between the current methodologies applied in human and environmental radiological protection. To introduce more precision in the subsequent text, the environmental approach that is being explicitly referred to throughout is the one that has been outlined for terrestrial animals as used by the ERICA Integrated Approach and the ICRP (2008). The methodologies for human radiological protection presented relate to the guidance of the ICRP (2007) unless otherwise stated. For the human approach, elaborated biokinetic models are used for a well-defined “Reference person” to simulate the intake route and retention of a given radionuclide within the human body. For example, for an (oral) ingestion of a given radionuclide, a compartmental model characterized by various exchange rates is set up to simulate the transport along the gastrointestinal tract. Other radionuclide specific compartmental models are used to simulate for a particular radionuclide the retention within and depuration from the body over protracted timescales. Coupled with dosimetric models, (based around, for example, Monte Carlo [MC] radiation transport codes) and using the elemental composition of the reference person, radiation, and tissue weighting factors to account for the quality of radiation (in causing biological damage) and the differential sensitivity of various organs to irradiation, coefficients mapping an intake of activity (in Bq) onto an effective committed dose (in Sv) are derived (ICRP 1996).
For the determination of exposures in plants and animals, and using ERICA as an example, a simplified methodology is typically used whereby concentration ratios (CRs) are used to characterize the radionuclide transfer, which is considered to be aggregated over all transfer pathways. Concentration ratio values for particular organisms are, as far as possible, empirically based, and the generally accepted approach has been not to differentiate between organs within the organism. In truth, this approach allows calculations to be simplified and is a valid reflection of a lack of knowledge about distributions of radionuclides in many organisms. Nonetheless, it also fits appropriately with the observation that most of the pertinent radiation effects data, on which the system relies, relate to whole body, external exposures. Dose rates, essentially considered to be instantaneous or under equilibrium and not accounting for dose commitment over subsequent years, are then derived from the starting point of an activity concentration in the plant and/or animal tissue and have been based on MC radiation transport simulations using tissue compositional data (Taranenko et al. 2004) with further elaborations involving interpolation functions (Ulanovsky et al. 2008).
Similarly, there are differences with regard to the external exposure models used in human and environmental radiological protection systems. Although both human and environmental systems are based on MC radiation transport modeling and are based initially on absorbed dose, the human system differentiates between exposure to organs and, once again, introduces factors such as the differential sensitivity of organs to derive dose conversion factors in units of effective dose (Sv). In contrast, environmental dosimetric calculations relate everything to a whole body absorbed dose in units of absorbed dose (Gy).
Differences between the methodologies for assessing risk from radiation exposure of humans and the environment are evident and without elaborating in detail it should be clear from the above why this is so. The cruder models used by ERICA are (arguably) adequate for environmental assessment where there is a requirement to consider a wide variety of organisms. This can be considered as also being pertinent to situations where this is limited to a small group of families as exemplified by the ICRP approach of reference animals and plants, as opposed to a specific species as is the case for human radiological protection. Likewise, a more simplified assessment and methodology would seem to be acceptable when it is intended to achieve protection at the population level (Copplestone et al. 2007) as opposed to reducing the probability of occurrence of a carcinogenic endpoint for an individual, as is a stated aim of protection for humans.
These considerations have instigated the present elaboration, the aim of which was to explore the possible implications of the simplifications made for the exposure component of the environmental radiological protection system in terms of the efficacy and robustness of (dose–rate) predictions. It was considered that this could be facilitated by conducting a comparison between human radiological assessment and environmental radiological assessment for a given case. For this purpose, an anthropomorphic surrogate was used within the environmental system with subsequent results for the surrogate, produced by both the environmental- and human-oriented radiological assessment systems, being critically compared and contrasted.
By comparing the transfer and dosimetric calculation methodologies for human radiological assessment (HRA) and environmental radiological assessment (ERA) using the same assessment endpoint, it will be possible to identify where close similarities and substantial discrepancies exist in output. Where there are differences, there is clearly a dilemma in identifying the source of this as, potentially, either one of the models could be aberrant or, plausibly, both models could be producing erroneous results. Nonetheless, in view of their longer history of development, more comprehensive refinement and specificity for the endpoint of this assessment, human radiological assessment approaches have been assumed to be the more robust. In a similar vein, where close similarities exist in model results, this provides a degree of reassurance that the ERA methods are producing sensible outputs and allows us to contemplate whether integration is necessary or indeed practicable. What we are not trying to do is to establish whether the protective measures that are in place for humans are equally protective as those in place for the environment. This has been the discussion of extensive earlier debate in the published literature (NCRP 1991; IAEA 1992) the resolution of which would seem, self-evidently, to rest on the assumptions made regarding the actual scenarios considered and the (environmental and human) benchmarks applied (Copplestone et al. 2010). Finally, the very specific assumptions regarding parameter selection, etc. that were imposed in this case study for the sake of allowing comparison between HRA and ERA should in no way be extrapolated to more general situations.
- Top of page
- CONCLUSIONS AND FURTHER REFLECTIONS
In the interests of clarity, the calculations and subsequent discussion of the results were split into several parts, these being 1) physical transfer in an ecosystem, 2) transfer to humans, 3) internal doses to humans, and 4) external doses to humans with the calculations being carried out using both a human radiological assessment and ecologically based radiological assessment approaches and the methods inherent in each.
The IMBA Professional Plus software (Birchall et al. 1998) was used to calculate the human whole body activity concentrations and internal doses in the present work. This software implements the latest ICRP dosimetric and biokinetic models to estimate intakes and doses from bioassay measurements for selected radionuclides (Marsh et al. 2005). Fractional absorption in the gastrointestinal tract (GIT), f1, factors of 1, 0.3, 0.3, 0.1, and 1E–-03 were used for 137Cs, 90Sr, 60Co, 210Po, and 239Pu, respectively, with all other models (GIT, bioassay, biokinetic, radiation weighting, and tissue weighting factors) using the default ICRP values available in the IMBA Professional Plus software. Values for f1 were taken from ICRP (1979) for 137Cs, 90Sr, 60Co, 210Po and from ICRP (1988) for 239Pu. These fractional absorption values are essentially for workers but with the exception of Po, which will be discussed in more detail below, provide relatively conservative estimates of transfer when compared to values generally used for the general public. This inherent conservatism provided the main rationale for the selection of these values.
For external doses for humans, recourse was made to the United States Environmental Protection Agency (USEPA) Federal Guidance Report No. 12 (FGR-12) where Eckerman and Ryman (1993) provide a compendium of dose conversion factors that have, because of the format they are presented in, high relevance to the comparison of external dose–rates. The ERICA approach (Larsson 2008) supplemented using the ERICA Tool (Brown et al. 2008) was used for the ERA. This required the use of guidance given in the approach, such as how one attains transfer parameters for cases where no specific empirical information is available. For dosimetric calculations an α-radiation weighting factor of 20 was applied to increase compatibility with the HRA.
Physical transfer in an ecosystem
In cases where an assessor has no access to a site specific “transport” model, reference can be made to the generic models that have been compiled and presented by the IAEA (2001, 2010). The models in IAEA's SRS-19 (IAEA 2001) allow activity concentrations in soil, sediment, and water to be derived with only basic information on release rates from the source, in terms of Bq per unit time, and various parameters. For example, the small lake model requires only 3 parameters characterizing the flow rate of the river flowing into the lake, the lake volume, and expected life of the facility that is discharging the radioactivity.
The models were originally developed with human radiological protection in mind, with an intrinsic level of conservatism such that “hypothetical critical group doses are generally likely to be overestimated” and that “under no circumstances would doses be underestimated by more than a factor of ten” (IAEA 2001). However, in view of the consideration that the physical and chemical processes leading to the advection and dispersion of radionuclides in the environment are identical regardless of whether the assessment endpoint is a human or wildlife, these models have obvious application to ecological risk assessments. The various SRS-19 models have thus been incorporated as the “front end” of the ERICA Tool as considered by Brown et al. (2008). Nevertheless, consideration of the matter indicates that there is a clear point of divergence in the application of the generic transport models. Typically, in the process of model implementation, a source to receptor distance requires specification. For humans, this distance will be determined primarily by resource usage. Examples could include the determination of the source to receptor distance in a river based on the extraction point for potable water or in a coastal environment based on the location of the commercial fishery nearest the discharge point. For wildlife, the identification of a unique distance is less evident because we are often concerned with various populations (Copplestone et al. 2007), characterized by different types of organisms exhibiting distinctive life histories. As elaborated on by Hosseini et al. (2011), the selection of a distance should ideally be organism-specific (cf. sessile organism such as mollusk to a mobile organism such as fish) and should also account for the spatial averaging in exposures that may occur as the organism moves within the environment being considered. Furthermore, in view of the focus on impacts on populations, this spatial averaging should account for the integration of exposure for a defined group of individuals.
Whereas the application of identical generic transport models is an obvious step toward integration of ERA and HRA methodologies, there are limits on how far this unification can be extended. For some environmental pathways, such as external exposure based on derivation of activity concentrations in soil, sediment, and water, the degree of unification can be quite extensive as long as careful consideration is afforded the location of the various “receptors” under consideration. For other environmental pathways, such as inhalation of radionuclides, integration may be less apparent. Whereas a typical HRA might, for example, tend to use air concentration results from generic models directly, an ERA like ERICA tends not to do so with internal concentrations being derived, via deposition calculations, from soil CRs.
Transfer to humans
In modeling the transfer of radionuclides in the environment the ERICA Tool draws on empirical databases, comprised of CR values, wherever possible. In cases where empirical data are not available, the ERICA approach provides guidance for the purpose of generating surrogate values. These values can be derived using various methods, including taxonomic and/or similar species analogues, biogeochemical analogues and various transfer modeling approaches (Beresford et al. 2008c).
For this particular case, it is assumed that data is available for a mammal that constitutes the main food source for a human deriving their primary sustenance from natural or seminatural environments. The guidance, where one is required to make predictions of transfer to a human based on such a limited data set and using ERA guidance, might be to use a “taxonomic analogue.” This would essentially mean the adoption of appropriate mammalian transfer (or by association activity concentration) data as a surrogate for the phylogenetically related organism the hypothetical human is assumed to be. In other words, the mammalian data will provide the surrogate values for the human whole body predictions made using the ERA. It should be noted that in this selection of taxonomic analogues, no effort has been made to evaluate the relative effectiveness of applying any other extrapolation approach to this problem.
Simple calculations can then be carried out for the sake of comparing HRA and ERA methodologies by assuming a unit concentration in the generic mammal and by assuming that this is the only foodstuff being ingested by the hypothetical human. In so doing, the human body activity concentration (Bq/kg) using the HRA methodology is derived by inputting the activity ingested per day, which is the product of this unit activity concentration and ingestion data for (generic mammal) meat into the selected HRA tool (i.e., IMBA Professional Plus). The selection of an appropriate ingestion rate is, of course, ambiguous because there are few, if any, contemporary human populations subsisting on a single animal product, but a possibly appropriate highly generalized value can be extracted from the compilations of the European Food safety Authority (EFSA 2012). Taking the 5 most populated western European countries, ingestion rates of 124 (mean), 267 (95th percentile), and 354 (99th percentile) g/d of meat and meat products can be derived. If a very crude assumption is made that the 99th percentile consumers could represent individuals deriving most of their daily nutritional requirements from a meat protein source, comparative calculations can be made. The results are shown in Table 1.
Table 1. Activity concentrations in a human using ERA and HRA methods
|Radionuclide||Whole body (Bq kg−1 f.w.)|
|ERA approach||HRA approach|
With respect to the values provided in Table 1, it appears that the ERA approach is a poor predictor for Pu and Po activity concentrations in the hypothetical human that could, given the context within which this thought experiment is being developed, be reasonably considered as representing a top predator in the ecosystem. It may be pragmatic, however, to view this observation in relation to the limitations of the underlying assumptions made in the ERICA Tool in their application to this specific case: first in relation to how mammals have been categorized in the ERICA approach and their transfer data have been treated (data for all mammals have been pooled together, assuming full comparability) and the second assumption with respect to “gap-filling” guidance for the ERICA approach previously mentioned. Allowing for even higher ingestion rates of meat that might arguably be more representative of an animal deriving its entire nutritional requirements from this ecosystem, the ERA approach still, substantially (by more than a factor of 10), overpredicts for these radionuclides. In contrast, 90Sr and 137Cs predictions are comparatively good. For these 2 isotopes there is no substantive need to adjust ingestion rates much before the values for the 2 assessment types correlate very closely. An appropriate general conclusion in this context however would be that consideration of dynamics is undoubtedly important with respect to a number of radionuclides, and that this is a matter warranting some attention in consideration of integration of human and environmental assessment. This might, for example, take the form of exploring the application of mechanistic based models, e.g., biokinetic models, to mammals, drawing on previous simulations tools for humans, to better capture the evolution of activity concentrations following specified intake regimes. Nonetheless, the ERA methodology for this particular application provides more pessimistic predictions than those derived using the HRA methodology, a result that is reassuring in view of the often intended use of the ERA transfer value in screening assessments where conservatism is a prerequisite (Beresford, Hosseini et al. 2010).
Internal doses to humans
IMBA Professional Plus was used to derive the effective (committed) dose to humans when a whole body activity concentration of 1 Bq/kg was attained using a specified chronic ingestion regime (Bq/d). Potentially, a problem exists in this regard in that when using the human radiological model, the initial whole body concentration is zero and a constant whole body concentration is not attained before a certain period of time has elapsed. In most cases, a constant whole body concentration would not be expected even over several decades. Indeed this is observed when applying the intakes to the bioassay function implemented in IMBA for radionuclides such as 239Pu and 90Sr. In other words, finding a unique value for an intake of certain radionuclides giving a 1 Bq/kg f.w. whole body concentration is problematic at best. For radionuclides with relatively short biological and/or physical half-lives (e.g., 137Cs and 210Po equilibrium is attained relatively rapidly), the activity concentration in the whole body attains a stable value in a short period). It is therefore possible to find a unique value for an intake. Conversely, for radionuclides exhibiting long biological and/or physical half-lives in body organs, selection of such a unique value is not possible. Therefore, with a view to consistency, an arbitrary reference time of 5000 days was selected for radionuclides where biological half-lives in various bodily compartments were protracted (of the order of years). An ingestion rate resulting in a whole body activity concentration of 1 Bq/kg at 5000 days was thus derived and thereafter used to estimate equivalent doses. The use of an arbitrary reference time was pertinent for 90Sr and 239Pu calculations only. The influence of the selection of 5000 days compared to an equilibrium (or 70 year time interval) value on the whole body activity concentrations of these radionuclides is shown in Table 1.
The doses reported by IMBA are not derived via the body activity concentrations with time. Instead, a chronic intake of “n” Bq/d over a given number “d” of days is treated as a calculation involving an acute intake of n × d Bq. This has the effect of removing a potential problem associated with having to account for a nonconstant activity concentration in the body with time when comparing with the results from the ERA dose models. Even so, the issue is not entirely circumvented but is rather transferred to the selection of an appropriate ingestion rate as described above. Furthermore, the dose for the human radiological assessment is “committed” for a default of 50 years (when using the IMBA Tool). Again, for radionuclides with long biological half-lives in various bodily organs this might be expected to create an anomaly compared to an equilibrium dose-rate model based on a whole body activity concentration (i.e., the ERA-based model). In contrast to the equilibrium model a large fraction of the dose is treated as being received in the years following the discrete moment in time for which the determination was made by the HRA approach.
IMBA Professional Plus provides estimates of effective doses. For the sake of comparison with the ERA, these values were converted to a dose rate by dividing the doses by the exposure time. This has the effect of providing an average dose-rate over the simulation period noting that dose rates will in reality be nonconstant over time. For radionuclides that attain equilibrium rapidly, this will not be an issue (i.e., average dose rates reflect those expected at equilibrium and are constant regardless of the choice of simulation times exceeding hundreds of days). For radionuclides attaining equilibrium slowly, or not at all, the choice of simulation time will affect the average dose rate to some degree thus the data derived in this way should be treated only as being indicative of equilibrium dose rates.
Within the ERICA Tool, the “Add-organism” functionality was used such that a new geometry might be generated representing the anthropomorphic surrogate. Necessary information concerning the mass (in kg) and the dimensions (of height, width, and length in meters) of this reference individual that was used by the ERICA Tool are provided in Table 2. The results for internal dose rates to the anthropomorphic surrogate as obtained from running IMBA Professional Plus and the ERICA Tool are given in Table 3.
Table 2. Morphogenic parameters used for the surrogate geometry in ERICA
| ||Mass (kg)||Height (cm)||Length (cm)||Width (cm)|
Table 3. Comparison of internal dose rates for the surrogate as derived from the human risk assessment tool (IMBA) and the environmental risk assessment tool (ERICA)
|Radionuclide||Input (Bq/d)||IMBA Effective dose rate (μSv/h)||ERICAa Weighted absorbed dose rate (μGy/h)|
|239Pu||1.1E + 01||2.3E–01||5.9E–02|
|210Po||9.5E + 00||9.8E–02||6.1E–02|
|60Co||2.3E + 00||6.6E–04||6.2E–04|
Tempered by the limitations discussed above in comparing 2 models that require 2 different input values and the consideration that the results of the ERA are in units of μGy/h and those for HRA in units of μSv/h, only tentative observations can be made. The (weighted) absorbed dose–rates derived from ERICA correspond closely with the effective dose–rates from IMBA for 137Cs, 90Sr, and 60Co. The similarity in dose–rates was expected for 137Cs reflecting the knowledge that this radionuclide reaches a steady state between activity per unit time ingested and activity in the body rapidly and becomes distributed relatively homogeneously. This means that an intake based model, like IMBA, can be set up to mimic a stable activity concentration in the whole body, as assumed in ERICA, in a convincing manner. The correspondence was less expected for 90Sr and 60Co because these radionuclides do not tend to distribute homogeneously (hence large differences might be expected between particular organs and body structures leading to pronounced discrepancies between weighted absorbed [i.e., essentially an “equivalent”] dose–rates and effective dose–rates). The selection of fractional absorption, f1, in the gastrointestinal tract is one of several factors that influence the effective dose derived using IMBA Professional Plus as transfer-biokinetics and dosimetry are interlinked. Altering the f1 value, of course, also necessitates changes to the ingestion rate (Bq/d) required to attain a unit whole body concentration at specified time “t.” For 239Pu, if an f1 value of 5E–04 is selected (as used as default by ICRP  as opposed to the 1E–03 used above) the ingestion rate must be altered to 19 Bq/d and the resulting effective dose becomes 2E–01 μSv/h. The difference between the HRA prediction and the ERA prediction becomes slightly less pronounced but is still substantial.
The results for 210Po present a particular conundrum. Selection of appropriate f1 has been the topic of some debate with ICRP increasing their recommended value from 0.1 in 1979 to 0.5 in 1993 (ICRP 1993). However, large variations have been reported in the literature as exemplified by Hunt and Allington (1993) who presented values in the range of 0.6–0.94. The use of the higher end of the range from the aforemenioned publications (with a concomitant lower ingestion rate to attain a unit whole body concentration at time “t”), however, has the effect of increasing the discrepancy between HRA and ERA results.
In corroboration of the views expressed above concerning whole body activity concentrations, it seems that further integration might take a path involving a more detailed analysis of the coupled biokinetic–dosimteric models for humans with a view to their adaptation to some mammals. This suggestion is further promoted by the observation that biokinetic data for humans have been derived by not only using results from studies involving people directly but also from animal experiments. Although the ERA models for 137Cs perform more than satisfactorily, this cannot be said for 239Pu and 210Po that might provide a radionuclide-based focus for initial developmental work.
External doses to humans
The FGR-12 report (Eckerman and Ryman 1993) provides dose coefficients for exposure to soil contaminated to a depth of 15 cm on a radioisotope by radioisotope basis. These data were selected as being nearest to the external dose coefficient in ERICA that relate to a 10 cm volumetric source. To be consistent with the internal dose calculations, the effective dose has been used within this discussion noting that it uses tissue-weighting factors that are specific to humans and stochastic endpoints. Only 137Cs and 60Co have been retained from Table 3 as the external doses from the other radionuclides considered previously are trivial, reflecting the low mean free path or low range of the β-particles emitted by 90Sr and the α-particles emitted by 239Pu and 210Po. To supplement the comparison, 3 other radionuclides that can be important contributors to external dose have been considered. The dose arising from a 1 Bq/kg soil concentration was derived using both human and environmental assessment systems (Table 4).
Table 4. Comparison of external dose estimates for the surrogate between human and environmental assessment estimates
|Radionuclide||FGR-12 Dose–rate (µSv/h)||ERICA Dose–rate (µGy/h)|
The external dose-rates derived using the FGR-12 are a factor of 1.3 above the corresponding values using ERICA in all cases except 234Th, where the FGR-12 value is 1.14 times the ERICA value. The difference is primarily explained by the depth of the source used in the 2 models. To demonstrate this, data pertaining to photon dose–rates from distributions of sources with depth in soil as compiled by Kocher and Sjoreen (1985) are used. The importance of the contribution to external dose rate for radionuclides at depth in soil, in the increment 10–15 cm for the given example, will depend on the characteristic energies of emitted photons from the considered radionuclides. Taking the energy range 0.5–3 MeV that covers a large proportion of the emitted photons for the selected radionuclides, the dose rate at 1 m above ground for a volumetric source of 15 cm depth would be larger than a volumetric source of 10 cm depth by a factor in the range 1.15 (for 0.5 MeV photons) to 1.22 (for 3 MeV photons) according to data tabulated by Kocher and Sjoreen (1985). Clearly this consideration largely accounts for the discrepancy between FGR-12 and ERICA dose–rates. Two less important factors plausibly explain the remaining difference. First, the FGR-12 uses a human phantom in the dosimteric set-up which might be expected to yield slightly different outputs than those provided by a human equivalent ellipsoid as used in our ERA set-up. Second, the calculations for FGR are, in fact, effective dose–rates derived via tissue weighting factors, wT, that account for the differential sensitivity of organs in the body to radiation. Although the sum of wT = 1 when deriving (whole body) effective doses, the calculation methodology will place more “weight” on absorbed dose–rate derivation for certain organs. This differs from the ERICA dosimetry where no such weighting is introduced. From this and other work where comparisons have been made with external exposure in the field (Beresford, Barnett et al. 2010), it seems that the ERICA external dose models perform reasonably well. Although Pentreath (2009) advocated further integration of ERA and HRA dosimetric methodologies possibly involving the application of voxel phantoms to wildlife, as the current state of the art methods do so for humans, it is a moot point, in view of the apparently minor “errors” introduced by the simplification used in current wildlife dosimetric methods, whether further integration is necessary.
CONCLUSIONS AND FURTHER REFLECTIONS
- Top of page
- CONCLUSIONS AND FURTHER REFLECTIONS
From the discussions and results presented above, it is apparent that identifying ERA and HRA exposure methodologies where integration is necessary is fraught with difficulties. The initial stage transport models simulating advection, dispersion, and deposition of radionuclides in the environment would be an obvious assessment component amenable for further unification, and in some cases developments have already moved in this direction. However, complete integration is arguably unattainable as the differences between ERA and HRAs in terms of the assessment endpoint mean that the relevant outputs from the models will not be the same. Comparing the transfer and dosimetry components of 2 typical ERA and HRA methodologies in making predictions of whole body activity concentrations and internal doses for humans, it seems that the efficacy of the ERA is rather radionuclide-dependent. Whereas, in the given example, the ERICA methodology provides predictions that are very similar to those provided by IMBA Professional Plus for humans, the predictions given by ERICA for 90Sr and 60Co are less satisfactory and those for 239Pu and 210Po are evidently poor. From this simple analysis, and bearing in mind the limited comparability of some of the values generated, it would seem that the equilibrium and homogeneous distribution-based models used in ERA methods like ERICA may have limited applicability to cases where radionuclides become associated with different organs after intake and where retention in these organs is protracted as might be observed for 239Pu in mammals. Integration might therefore take the form of exploring the adaptability of the biokinetic models developed for humans with regards to selected animals and radionuclides. External dose models developed for ERA provide results for humans that correspond quite closely with those provided by standard HRA methods. Although further integration of the methods could be attempted through the application of voxel phantoms in ERA, for example, the requirement to do so is a moot point.
Although human and ecosystem radiological assessments share many similarities, there are a number of important differences between them in terms of both the aims of the systems and the means by which they are conducted. The connection to the natural environment is more apparent for wild plants and animals than for humans. This contrast of “distance” between the receptor to source in the environment, in terms of both spatial distance and food chain length for biota and humans, implies that ambient concentrations of radionuclides in the environment are likely to have a more direct influence on the exposures of wild biota through direct physical contact with the contaminated media and via transfer through short, more consistent food chains than they will for humans. Consider, for example, the case of a population of freshwater fish that are continually in direct contact with a given contaminant in lake water compared to a human living on the lake shore. Humans may be exposed intermittently to the same contaminant but additional exposures may occur via ingestion of contaminated foostuffs from other locations or external exposures via working activities or lifestyle. Predator–prey relationships may lead to complex responses to any given environmental stressor, as exemplified by the modeling work of Bartell et al. (1999). Because the human environment is largely controlled, these types of relationships need not be considered explicitly when considering human health endpoints. Because of the inherent complexity of the processes involved and the enormous variability of organisms and their natural habitats, more simplifications and approximations necessarily have to be made when undertaking environmental radiological assessments compared to human radiological assessment. All these introduce extra uncertainty into the estimated results. The various tools that have been developed in human radiological risk assessment are generally more advanced than those available for environmental risk assessment, reflecting the fact that humans have been generally more studied than most species of wild plants and animals. This imbalance can also be extended to the degree of data coverage in these 2 fields. However, there are still many advantages of integrating these 2 assessment types. These include achieving a more consistent and coherent basis for regulation and greater efficiency and quality of assessments by sharing of data, models, and insights (Suter 2007).