[Comment by ICRP Scientific Secretariat: this was submitted with certain statements boxed but since the boxes are invisible on the web site, we have added the word 'important' where the original statement was boxed] General comments There is very little information on transfer in the environment. This will probably be the largest factor leading to uncertainty within any environmental impact assessment. Surely this should be reflected? At least some indication of knowledge gaps with regards to environmental transfer could be provided. For many of the reference organisms selected, e.g. frog, bee, earthworm, there will be no transfer information for most of the radionuclides that may require consideration within an EIA. Under Section 4.2 Current approaches to dosimetry there is a great bias towards the methods developed by Copplestone et al. (2001) – see paragraphs (57), (58) and (60). Although these methods are undoubtedly of great utility, they are, in essence, good approximations. The state of the art with regards dosimetry is widely considered to be Monte Carlo methods. Several groups are currently working with such methods in relation to non-human biota (e.g. GSF, CIEMAT, ANL)– the text should maybe reflect this balance. Specific comments (14),page 10 – reducing the frequency of effects likely to cause early mortality or reduced reproductive success … to a level where they would have a negligible impact on conservation….maintenance of biodiversity… health and status of habitats and communities. This level is practicably difficult to establish without the use of population dynamic and ecosystem modelling. No details are provided later in the document as to how this might be achieved in practice. (25), Page 14 – “requirements for ‘environmental impact assessments’ – “with respect to impacts on other forms of environmental management” poorly formulated : this can be better described by specifying that the impact of radiation on the environment needs to be considered within broader environmental management practices including those considered within fishery and agricultural management. (26), page 14 – the application of the approach will vary at national or regional level ? (27), Page 15, ….”because ecotoxicological tests are often used in pollution control, it would be important to ensure that the total reference set had a reasonable coverage of the major ecological compartments of terrestrial and aquatic ecosystems” ? This is a non sequitur. (37) P18 – some of the identified ref organisms in Table 1 are not at the family level ! Only “deer”, “trout” and grass are taxonomically defined as true families “flatfish”- order name Pleuronectiformes. Plaice is a family of flatfish. “frog” – order Anura “crab” – members of the order Decapoda “earthworm” - class Oligochaeta “seaweed” – one of the major groups of marine algae - others may be O.K : “rat” – generally applied to several rodent families therefore term probably O.K. “duck” – technically a subfamily (Anatinae) “bee” – technically any member of the superfamily Apoidea “pine tree” – genus pinus This is mitigated later in the document. Under the reference organism description 3.5.1 to 3.5.12, further specifications are provided as to which “family” the authors had in mind. Nonetheless, minor inconsistencies still remain. Reference brown seaweed is not defined at the family level. For earthworm, Lumbricus is classified as a genus as oppose to a family. (38), page 18, “the use of the human animal itself” Slightly unclear – suggest “using humans as the “reference” would suffice for the purpose of demonstrating radiological protection in this managed environment.” (40), page 19. “…medium-size woodland deer..” the deer family vary in size from a height of 30 cm to approximately 2m ! One could specify that the geometry is defined by a typical adult deer falling near the middle of this range. Important: (54), p23. “these expression have been integrated…” citing Brownell et al. (1968). Is this correct ? Brownell et al. (1968) used Monte Carlo calculation to simulate photon diffusion -There was no integration of the aforementioned empirical expressions in this work. (57), p.24. “such data for photons….” ? Important: (61), p. 25. EPIC didn’t use pure Monte Carlo calculations. Dose attenuation and chord distribution functions were used. A random number generator was then used to generate random points and directions for photons and beta particles within user-defined geometries (essentially the Monte Carlo part of the calculations). Note that within the EC-funded project ERICA – DCCs will be generated using Monte Carlo codes (GSF) (73), p28. “Such an evaluation….. for it is the best understood radiation effect” this is unclear. Stochastic effects can encompass a wide range of more specific effects such as cancer induction, hereditary and teratogenic effects etc. Are we saying that all these widely differing stochastic effects are better understood (at a biological-mechanistic level) than deterministic effects ? (78), p.30. “And in order to be generally….. there were no experimental data….” (84), p.31. “…. Relating to the need for and function of…..” Important: (96), p. 34. “ ….within a range of 2 to 67 Gy d-1 for adult benthic crustaceans and molluscs.” • “Broadly similar values have been calculated for European waters generally….” This is confusing as the same reference has been used in the previous sentences to define background dose-rates for different organism groups. This could be changed to read for Arctic boreal water generally, citing Brown et al. (2003). • It may be better to use (1-3 Gy d-1) for macrophytes from Brown et al. (2004) in order to be consistent with the other reported values. There is less clarity with regards to how the EPIC values were derived. (99), p34. “… to try to draw….” P. 35. the text under Section 7.3 creating a reference set. Does this need a paragraph number ? The term “dose per unit concentration” – is this now the accepted terminology. In other groups/ projects (IAEA EMRAS, ERICA), the term Dose conversion coefficient has been generally agreed upon. Important: (105), p.36. This paragraph is identical to Paragraph (103). Important: (108), p.37 Reference to Strand et al. (2003). This paper did not provide the cited conclusion. The reference may be better placed in paragraph (109). The paper did mention extrapolation issues – laboratory to field and individual to population. (110), p.38 “….(e.g. seawater)….” Is this an appropriate example of an environmental situation ? maybe better to define “(e.g. radionuclide release to an aquatic coastal ecosystem). • Might it be appropriate to provide an example table here (e.g from Pentreath, 2002a), to illustrate the use of DCLs ? (112), p.38. the first sentence is awkward. Suggest “ In cases where the animals and plants being considered in the assessment are not similar (taxonomically or in terms of their size and shape) to the twelve types of reference Animals and Plants, then locally derived…..”. (117), p.40. “…..there are sufficient data to begin the process of developing a system…