Environmental Protection - Transfer parameters for Reference Animals and Plants

Draft document: Environmental Protection - Transfer parameters for Reference Animals and Plants
Submitted by Steven Shaffer, Nuclear Regulatory Commission
Commenting on behalf of the organisation

The U.S. Nuclear Regulatory Commission (NRC) staff would like to thank the International Commission on Radiological Protection (ICRP) for the opportunity to provide comments on the draft publication titled “Environmental Protection: Transfer Parameters for Reference Animals and Plants”. The NRC appreciates the opportunity to submit and review stakeholder comments on ICRP documents.


General Comments:


1. Nothing in the following comments should be construed as changing the NRC’s previous position and comments regarding the need and merit value of developing a separate radiation protection system for the protection of the environment that we provided for ICRP Reports 103 and 108.


2. We recognize this report is an initial step in developing a more consistent approach to

modeling of Reference Animals and Plants. Such a modeling approach, with additional data and development, along with recognizing and heeding important caveats or limitations in the data could become a useful tool in supporting the existing environmental assessment requirements legally mandated in the United States. We recognize that the assessment methodologies for humans has taken decades in development, and that we would fully expect that data to support assessment methodologies for environmental assessments will need to be reliable before use in regulatory applications and continuously refined and developed, as additional knowledge and understanding of the complex ecosystem interactions are gained.


3. Use of “lumped” parameter sets such as a concentration ratio (CR) has been previously used throughout environmental assessments for human exposures. Soil-to-Plant transfer factors are, of course, the most analogous. Other examples are Kds for interaction of elements in a soil/water matrix and mass loading factors for estimation of air concentrations based on soil concentrations. These gross level parameter values attempt to provide an “effective” result in one parameter value that is actually the resultant steady-state behavior of a large number of processes. This can be readily seen when attempts are made to provide more mechanistic models, for example, the tritium, fruit tree and forest models from the IAEA BIOMASS program that attempted to replace specific activity or CR approaches. Such lumped parameters are useful in cases where average steady-state behavior or results are desired, and appropriate, for modeling as peak concentrations will not be captured. Use of a CR would not capture variation in the concentrations between individual specific modeling points caused by a variation in the effectiveness, or presence, of each of the specific processes implicitly being modeled by the CR or the result of transient conditions.


As the purpose of the CRs for Reference Animals and Plants, based on ICRP 108, is on a population-level assessment and not an individual assessment, use of a CR approach could be appropriate for modeling purposes, especially with the lack of data to support more complex approaches. A CR approach is an adequate initial modeling approach, as long as the limitations and caveats of the supporting data are understood by the analyst and decision maker.


However, the draft publication highlights the need for a substantial amount of additional information and data. The approach needs significant research and model development to bolster the scientific basis, including gathering empirical data for most species and collocated environmental media. The uncertainty associated with the use of a CR for reference species derived from limited data or surrogate approaches is very great. As such, the document is a statement of a research program, and not particularly useful as a scientific reference, or a foundation for making biosphere modeling decisions.3. ICRP discusses the limitations on the application of the CRs in many different sections of the report on specific issues. However the document is not clear on how or why one should use the CR approach. We suggest adding an additional section summarizing ICRP’s environmental impact framework and where the CRs fit into that framework. ICRP could additionally strengthen the report by adding a specific section in Chapter 2 on the general limitations of the current data set. This should limit misuse or mischaracterization of assessment results that utilize the compiled CRs.


4. The report needs to discuss in more elaboration uncertainties arising from use of CR data for Reference Animals and Plants in dose calculations of Representative Organism, as the relationship between Reference Animals and Plants and Representative Organism is vague. In addition, approaches on how to identify and characterize the Representative Organism need to be discussed in more detail. In this context, we envisage additional large uncertainties associated with the concept of Representative Organism (e.g., due to influencing factors as such: age, species, allotropy, migration/range of habitat, climatic changes, currents, chemical conditions, etc).


5. In many instances, the report relied on unpublished (or in process) references to support the ICRP results. ICRP could strengthen the report if it delayed final publication until these critical IAEA and others reports become final to ensure that the references continue to support the ICRP results or approach.


6. The CR data focused on average values (e.g., arithmetic means or geometric means). Therefore, we believe its use is for deterministic analysis. NRC notes, that for probabilistic parameter analysis, modelers typically establish distribution function curves (e.g., probability versus parameter value) for each parameter to assess uncertainties in a more realistic fashion in the dose exposure calculations. Therefore, the report would benefit from establishing probabilistic distribution function curves particularly if sufficient data are available for the specific parameter, especially as the standard deviation for many of the species is greater than the range of data for the element: species combination.


7. The report may benefit from having a glossary section since ICRP uses numerous biological and ecological terms without explanation and the standard ICRP audience may not be familiar with all terms.


8. The report acknowledges its limitation for use only under planned and existing exposure situations. The report indicated that the CR approach is inappropriate for use under accidental or emergency exposure situations. We note that most significant radiological contaminations that indeed may require special attention for protection of the ecology and environment are related to accidents or unplanned exposures (e.g., Chernobyl). It would be useful to discuss possible alternate strategies to estimate impacts in such situations.


9. Is there a point at which the lack of specific data is enough to keep the animal from being considered a representative species? If the majority of the data for a species is derived from another species or generic wildlife then the species should not be considered a representative species. Sections 4.2 – 4.4 really highlight the difficulty of finding the data needed for reliable dose estimation for biota. (Most of the data shown are in grey shading, and the grey values shown are all derived values.) If this lack of data is followed through logically, it seems to call for a more limited number of more generic Reference Plants and Animals with a wider applicability.


10. NRC has numerous licensees who have the species indigenous to their locations that are not captured by the Reference Animals and Plants. Ideally, ICRP should either enlarge the set of reference species by specifying subsets of reference species by ecosystem or geographic location or develop a standard methodology to identify new reference species on a site- or region-specific basis (similar to our previous comments on ICRP 108). However, because currently there is limited data for even the current reference animals and plants, we suggest using a more limited number of generic plants and animals that modelers can apply to a wider range of ecosystems, or geographic locations. The report should address the level of uncertainty that is added to an assessment by using either a generic species or having a noncomparable animal or plant of concern (e.g., a predatory avian).


11. A CR approach may be difficult to utilize for species with large habitats, life cycle changes that result in changes of habitat (e.g., tadpoles to frogs), or other features (such as exoskeletons) that may result in over or underestimation of the exposure of the points of interest for the analysis. We recognize that these are discussed intermittently in the current text. However, if a general section on the use of the CRs is developed (comment 3 above), it should discuss this issue and the uncertainty that results from using CRs in these cases.


Specific Comments:


1. Revise Executive Summary - The executive summary does not appropriately summarize the publication. The first paragraph (70-83) discusses the importance of dose rate in evaluating the effects of radiation on animals and plants. However, the second paragraph (85-93) appropriately discusses how and why ICRP derived the concentration ratio data. Also, the last two sentences of the summary (lines 89-93) seem contradictory. How can there be “a paucity of data” if the ICRP was able to establish “and enormous data base”?


2. ICRP should add as a reference the special issue of the Journal of Environmental

Radioactivity, Volume 100, Issue 9, Pages 671-806 (September 2009) which contains 20 papers devoted to radionuclide transfer in terrestrial and freshwater environments. As an additional reference, ICRP should add NUREG/CR-6941, “Soil-to-Plant Concentration Ratios for Assessing Food-Chain Pathways in Biosphere Models”, August 2007. This report provides soil to plant concentration ratios for 99Tc and 238Pu in onion, alfalfa, corn and potato from three regional geographical locations in the United States.


3. Line 151 – Replace the word “exposure” with “radionuclide”.


4. Line 392 – Given the statement “Modeling the physical aspects of transfer of radionuclides in the environment is beyond the scope of this work,” ICRP could greatly simplify the background section.


5. Figure 1.1 - The figure is hard to read because of the gray “shade” behind the words. Some of the letters are also not clear (e.g., Desorption, resuspension).


6. Figure 1.2 - The figure is hard to read. Possible improvements include:


- Using a color version of the figure or one with line drawings instead of variations in shades of gray

- The use of different color arrows to determine which pathways are from the water column versus the sediment would also be helpful (“via” and “vib” are confusing and hard to read)

- Make the organisms larger in the figure as there is a lot of wasted space

- Increase the size of the figure so that it takes up the entire page


7. Figure 1.3 - The figure is hard to read. Possible improvements include:


- Using a color version of the figure or one with line drawings instead of variations in shades of gray

- Using different color arrows to differentiate which pathways are from the air versus the soil (“via” and “vib” are confusing and hard to see)

- Make the organisms larger in the figure as there is a lot of wasted space

- Develop a different “symbol” for surface water as it is hard to see

- Increase the size of the figure so that it takes up the entire page

- The vib pathway for waterfowl is missing its label


8. Table 1.1 – This table lists Ag-110 (25 second half life). Do you mean Ag-110m (249 day half life)?


9. Line 707 – For clarity, add “as was adopted for reference man” after “a similar approach”.


10. Page 22- The report references two IAEA reports for Kd data for marine and freshwater ecosystems. Another Kd data sets that should also be referenced is Sheppard, M.L. and D.H. Thibault, “Default Soil Solid/Liquid Partition Coefficiends, Kds, for Four Major Soil Types: A Comparison”, Health Physics 59 (4): 471-478 (1990).


11. Paragraphs 68 and 69 - Paragraphs 68-69 on propagation of variance need revision. Is Equation 1 correct? It should probably be something more like:


Vcomb = Vw /  N  + VB


In addition, the symbols associated with Equations 2 and 3 do not match their text descriptors (e.g., m versus M).


12. Lines 1094-1099 – The text should also discuss the temporal and spatial aspects of variability.


13. Lines 1174-1190 – ICRP did not justify well the conclusions drawn between grass and pine trees. Suggest providing additional justification or supporting data.


14. Lines 1231-1235 – ICRP did not justify well the conclusions drawn between ducks and frogs. Suggest providing additional justification or supporting data. Also, the document would be strengthened by more discussion of the impact on uncertainty in using other species or element data to “fill in” the existing data gaps. This is especially true if the majority of the data is another species such as the use of trout data for frogs (water-based CR).


15. Lines 1256-1257 - For the frog, as two sets of CRs are developed, the document should state whether they are independent assessments (an assessment of dose would use one or the other) or are they additive


16. There are no data for 99Tc baseline CR values for the animals and plants listed in Table 1.1 and in Chapter 4, ANNEX A, and ANNEX B of the report.


17. The footnotes to Tables 4.1, 4.2, etc., are not obvious. For instance, the footnote for plutonium simply states “Am”; others state “Pu”. We suspect that a chemical analogy was applied – but not consistently: for instance, Rat is compared to Pu while Deer is compared to Am. Annex B is incomplete and the reasoning may be intended to go there. Each footnote letter should be listed on its own line so that it is easier to follow. Also, the dependent Bq kg-1 should be more clearly identified in each of the Tables 4.1, 4.2, 4.3, and 4.4 as (dry or wet?) soil or water to reduce the chance for confusion of readers.


18. Paragraphs 115, 116, and 117 do not seem excessively useful, in their current form, as they are suggesting a drastically different approach then is presented in the publication. A discussion of how to apply the data in evaluations while accounting for the inherent uncertainties would be more useful. Also, the baseline assumption is to compare concentrations in the reference organism to the concentration in a general environmental media (soil, water, or air). However, comparisons to other media such as concentration in plants for herbivores or plankton for fish may, if data becomes available or is available, be a more valid comparison to estimate resultant concentrations (i.e., similar to how current soil-plant-cowmeat/milk pathways are currently modeled for human exposures). This would be especially important to consider when attempting to develop CRs for species that are more based on a predatory nature (and thus, further from direct ingestion of soil), which make up a lot of the endangered species categories. This also may be important for bees. This comment is twofold, 1) consider on a species basis if such an alternate conceptual model for environmental assessment would be more useful and 2) for the possible data collection in Paragraph 116, intermediate food sources for the Reference Animals and Plants should be directly sampled also to evaluate alternate concentration ratio development.


19. In ANNEX A, page 47, the mixing of Geometric Mean and “best estimate” (i.e., Average) in the same tables is probably a poor presentation approach. If a point estimate is to be used, the Expectation Value (that is, the mean) should be applied to all CR values. A better presentation would give the Geometric Mean, Average, Geometric Standard Deviation, and Standard Deviation of all distributions. If the complete set is unavailable, the user could then understand the limitations of the data and apply the proper uncertainty to each calculation. If the mix of GM and Average is retained, the result will be dose estimates lower for many of the point estimates based on the GM.


20. There appears to be no relationship between the statistical data on concentration ratios for reference animals in plants shown in ANNEX A and the derived concentration ratios listed in ANNEX B. Explain how these two annexes are related. Also, there are no derived concentration ratios shown for terrestrial and freshwater ecosystem plant and animals in ANNEX B.


21. ANNEX A, page 47. There are no data reported for bees, and only little data shown for pine trees, frogs, ducks, rats, and deer