Radiological Protection in Geological Disposal of Long-Lived Solid Radioactive Waste
Comments on ICRP’s draft document
Mikael Jensen, email@example.com
1 The value of a document applying new recommendations to the post-closure phase
It is necessary that ICRP applies the general recommendations to geological disposal (of long-lived high level waste) in a timely fashion. After the last general recommendations, ICRP 60, a decade passed before recommendation ICRP 81. During this time, the (then) Swedish Radiation Protection Institute promulgated its regulation, the Swedish high level waste standard (to use an American term). Existing recommendations such as ICRP 46 from 1985, was only of limited use to those who established the Swedish regulations, including the author of these lines.
As I mention below, geological safety and radiation protection are two different worlds, and there is an obvious danger that the present draft may have most of its impact on one of the two worlds (performance assessment), while being discussed in the other (related to areas such as occupational radiation protection).
Some of the comments I offer here are described much more in detail in a recent published book (J. Ahn and M. Apted, editors, Geological repository systems for safe disposal of spent nuclear fuels and radioactive waste, Woodhead Publishing), in which I try to bring the two together in chapter 20 (Radiation protection principles and development of standards for geological repository systems).
2 A mix of many phases
To most stakeholders, the main goal of geological disposal is post-closure safety, not safety (of any kind) in the construction phase. There is too much emphasis on the operating phases in the draft.
I suggest concentrating on the post-closure phase, since that is where most of the question marks are. The other phases, including the operating of nuclear facilities, are covered by other publications. Optimization between the different phases, i.e. including the impact on post-closure safety for a choice between alternatives in the operational phase, is the only justification for treating the routine application of ICRP’s system in the operating phases at all. This important fact is diluted by the description of routine radioactive waste management operations.
3 Radiation protection vs. safety
I note with satisfaction that Claudio Pescatore from the OECD/NEA, Magnus Westerlind from IAEA and Carl-Magnus Larsson are all involved in the process of developing this document. One problem is that the ICRP documents are discussed by radiation protection experts. While this seems natural, it must be underlined that ICRP’s recommendations in many cases come close to being the main safety requirement. There is a risk that the issues involved are not debated deeply enough within the performance assessment community, in charge of safety and with very few radiation protection experts.
(During the first attempt from the European Commission to launch directives regarding safety and waste, the special issue of radiation protection standard for geological repositories was assumed to be a safety issue by some of its members, i.e. belonging in the safety directive, but believed by others to be an issue of waste, i.e. belonging in the waste directive. I have not followed the EC’s present work on directives but I suspect this issue still has potential for generating confusion).
There is a built-in potential for misunderstanding when a central safety issue is treated as a radiation protection issue. The generous treatment by the ICRP of other phases than post-closure can only make the situation worse. It clearly gives the radiation protection community the lead, which they may not fully know how to manage. An example from the Swedish regulatory body may demonstrate this:
In Sweden, radiation physics is a well developed, institutionalized scientific curriculum and the regulator has many experts employed. However, when I left the organization in April this year there was no such discipline represented in the working with for regulation of final disposal, in charge of the regulatory review of a license application for disposal of spent fuel.
I can therefore not underline enough the need for a broad consultation.
4 The concept “Planned exposure situation”
The problem of the term “planned exposure” has been brought up earlier by Claudio Pescatore after a meeting in Paris some 6-8 years ago, in response to a presentation of ICRP’s principles by Annie Sugier. I would like to make a few comments late in that debate.
ICRP’s nomenclature, as ICRP itself, has its origin in radiology, hence the term Planned exposure situations. This term is doubtful in any circumstance, including in medical radiology, since it is seldom exposure itself is the goal (the planned activity), except for the discipline of radiation therapy. “Planned situations, entailing exposure”, would better cover such activities (as radiography) for some areas, but not for geological repositories. My point is that the existing terms, modified or not, is still not accepted in the performance assessment community. The post closure phase does not amount to planned exposure, not even in my modified version. I note the passage in the draft:
(j) ICRP recommends that dose or risk estimates derived from these exposure assessments should not be regarded as direct measures of health effects beyond timescales of around several hundred years into the future. Rather, they represent indicators of the protection afforded by the geological disposal system.
The time scale ICRP prefers for measures of health effects, are periods where nobody expects any releases, even taken “cataclysmic geologic changes such as glaciation and tectonic movements” (from ICRP 81, see below) into account. The potential releases always exist - by definition - but they are connected with very low probabilities.
On the other hand, when time longer scales are used OECD/NEA’s publication describe both that future dose results do not represent direct health effects: “doses and risks calculated on the basis of stylized approaches should be interpreted as be illustrations based on agreed sets of assumptions for particular scenarios and not as actual measures of future health detriments and risks” (Post-closure Safety Case for Geological Repositories, Paris, OECD/NEA, 2004). NEA’s reports both describes better the limitation of the calculated results and the global view of all the assessment activities.
5 ICRP’s view on traditional performance assessment
Understanding the long (about 100 ka) vs. short (0.1-1 ka) time scales
Examples of ambiguity comes from the quotes
1. (864) “be projected to occur at such distant times that traditional concepts such as dose and risk have to be used with caution”,
2. (778) “...the application of dose limits to waste disposal has intrinsic difficulties” (from ICRP77)
3. ICRP81: “To evaluate the performance of waste disposal systems over long time scales, one approach is the consideration of quantitative estimates of dose or risk on the order of 1000 to 10,000 years. This approach focuses on that period when the calculation of doses most directly relates to health detriment and also recognises the possibility that over longer time frames the risks associated with cataclysmic geologic changes such as glaciation and tectonic movements may obscure risks associated with the waste disposal system”
The points above must coexist with ICRP’s suggested limits and constraints depicted in Table 1. How does ICRP itself regard its own recommendations, i.e. for how long is Table 1 valid in the post-closure (called no oversight) phase?
The third bullet seems to give 10 ka as the maximum time period to be considered. The questions that come to mind are that ICRP:
- disregards the need for long term assessment for geological disposal (but is uncharacteristically diplomatic)
- considers the results of long term safety evaluation to be a matter of policy rather than of science.
- as a result of a number of circumstances, as above or other, ICRP is a victim of an existing protection system terminology which is neither intended nor well suited to cover geological disposal.
ICRP seems to include the long term perspective - by not ruling it out - but is vague on both i) the actual time scales to be used and ii) the way one would compare a limit with an assessment of doses which in that process should be “used with caution”.
There is no doubt that areas in need of guidance include the long term (100 ka). To leave this out would be to disregard the main concern expressed by decision-makers, the scientific community and the public.
ICRP’s suggestions amounts to a description of what should not be done (strict comparison with dose limits etc), with no positive guidance. ICRP should instead seek common ground with the performance assessment community and follow such a path even in the face of potential needs for improving its system of protection.
ICRP’s advice regarding components in the normative safety assessment
The idea of a central design-based description of the repository is in need of improvement. It is mentioned in the executive summary using only its negation (non-design-basis conditions), in a way that the reader gets the impression that this is an important and well-explained concept in the main text. This is not the case, however. Rather it is used in a vague form (“696: The multi-barrier, multi-function system that is at the basis of the disposal facility design).
In 696, ICRP introduces confusion by using the “basis of the design” as “the details of the design”, whereas the design bases instead refer to different goals to be achieved by the design.
ICRP’s idea seems still to be to define a special description of a repository and a normative evolution as the special case for comparison with a limit of 0.3 mSv/a. To this is added a number of disruptive events of the undisturbed repository, in consultation with stakeholders.
A better way would be to use the risk represented by a dose limit (using the dose-risk conversion factor as we understand it today) and the event’s associated probability of occurrence. Very small probability events (e.g. meteorite impact) would then be taken out of the process by their low probability of occurrence. A probability-of-occurrence related cut-off could still be appropriate to limit the assessment, i.e. allowing rejection of the need to even address a low-probability tail of events stemming from endless speculation. It is not clear if this is ICRP’s intention or not.
Glaciation in the Nordic countries may be - or not be - part of a central case, rather than looked upon as a disturbance. Limiting the choice to only one set of calculations may be unnecessary restrictive. It may be more natural to have a few possible central cases to compare with ICRP’s suggested (dose or) risk limit. This example again assumes that time periods longer than 10 ka are included.
One possible way for ICRP would be to say that the goal for a “multi-barrier, multi-function system” could be isolation for more that 100 000 years and that efforts should be made to analyze the performance over this time frame, but limiting formal requirements regarding dose or risk to apply to the time frame of 10 000 years (at most).
6 Problems connected with open time scales
One problem related to an open time scale used for quantitative assessment is that as time increases, events with lower probabilities become more likely to occur during the assessment period. There are two issues in this:
1. if no length is given for the assessment period, any unlikely event will take place with an ever-increasing probability, and
2. a choice must be made about whose risk or dose should be referred to in the comparison of dose or risk with a limit or constraint.
The second point refers to what is called risk dilution. It is described in the Swedish guidance document (SSI FS 2005:5 The Swedish Radiation Protection Authority’s guidelines on the application of the regulations (SSI FS 1998:1) concerning protection of human health and the environment in connection with the final management of spent nuclear fuel and nuclear waste) but nor in ICRP’s draft.
Another philosophical issue is called the clairvoyant test, which a dose or risk limit will fail. If a clairvoyant, able to look into the future as far as the assessment period, tells us that no doses occurred, it does not tell of about the risk; we could just have been lucky. Referring to dose instead of risk does not solve this problem: a dose must be an expectation value of dose over different probabilities (unless the design base case also includes design base hydrology parameters, design base chemical reactions etc, limiting the assessment calculations to a single line of multiplication).
Prediction with formal probability paraphernalia can be avoided by using a more vague term “expected dose (or risk)” as in the US standard for Yucca Mountain.
The problem involved by the use of dose or risk at all has been pointed out by the German Reactor Safety and Radiation Protection Safety Committees: ‘The difficult communicability of the risk approach and difficulties in quantifying the entrance probability still limits the use of the risk criterion’. (The German Radiation Commission 182nd Meeting, 4–6 December 2002). (In German).
8 More on time frames
“(670) Time of no oversight: when oversight is no longer exercised because memory is lost. This timeframe coincides with the post-closure period in the distant future. “
This is an absolute statement, better to be replaced by a more humble admittance of uncertainty, such as:
“Time where oversight cannot be assumed: when oversight can no longer be taken for granted because memory of the repository may be lost”. Improved information conservation cannot be ruled out completely, even for extremely long time scales. It is just beyond our capability to assess the likelihood of such future scenarios.
This overconfidence - that some things with certainty cannot exist in the distant future - also has the potential to block valuable attempts of optimization by improving information conservation. On the contrary, better measures to safeguard records may well – as other possible measures – accomplish a limitation of “the likelihood of incurring the number of people exposed, and the magnitude of their individual doses”, thereby following the Commission’s own definition of optimization.
9 A few comments on human intrusion
As above, one can only agree with the statement:
(994) Therefore the dose or risk constraints recommended by the Commission for the application of the optimization of protection in geological disposal do not apply to inadvertent human intrusion.
The process of optimization in connection with human intrusion cannot realistically be bounded with dose limits. However, optimization can and should be carried out using several measures, such as avoiding known mineral sources in the siting process, as pointed out. This is also true for record preservation. I believe this issue should be mentioned explicitly.
Optimization in the post closure period include the maintenance of records, including the establishment of a well structured archive (throughout the operating phases by establishing a strategy leading up to a final archive and its emplacement in international bodies - this important component is referred to only as post closure “oversight”).
The same goes - and is perhaps even more important - for a marker system. There are various formal expectations of how an archive may look like, but much less consensus on marker systems, such as are required for the US WIPP repository.
10 Further on optimization:
The quotes used above: `...the application of dose limits to waste disposal has intrinsic difficulties' (ICRP, Publication 77, 1997b, paragraph 19, also makes constrained optimization prohibitively difficult also in the general case of safety analysis. However, it is worth noting that limits may sometimes be used in a negative way, to point to a (hypothetical) scenario where the dose limit requirement would be violated, thus identifying a (hypothetical) unacceptable alternative.
11 The goal of final disposal
The draft defines the goal of final disposal. It is possible to define an alternative goal of disposal along the following lines. The principle of sustainable development is defined by the Brundtland commission: "Sustainable development is development that meets the needs of the present without compromising the ability of future generations to meet their own needs".
Following this principle we must allow future societies to use and change the biosphere so as to fulfill their needs, and hence we must require that a number of different biospheres must all be able to coexist with the repository. This view brings one (important) uncertainty out of the safety assessment. The biosphere would no longer be in need of prediction. The need is instead one of decision, (by the regulator in a dialogue with stakeholders).
I have promoted this view in a number of years (e.g. in chapter 20 in the book mentioned above) but with little international resonance so far. I suggest ICRP takes a look at this alternative goal formulation.
12 BAT and optimization
It seems the ICRP has a view similar to the former SSI, but it may still be worth while to mention the Swedish guidance document’s definition of optimization and BAT in connection with geological disposal:
Measures for optimization of a repository should be evaluated on the basis of calculated
Application of best available technique in connection with final disposal means that the siting, design, construction, operation and closure of the repository and appurtenant system components should be carried out so as to prevent, limit and delay releases from both engineered and geological barriers as far as is reasonably possible. (The documents go on to explain that “When striking balances between different measures, an overall assessment should be made of their impact on the protective capability of the repository”.)
Application of best available technique is thus used to describe the need for good choices between alternatives which may not be easily described in terms of risk estimates, such the use of effective QA requirements throughout the program. The term covers rudimentary but important approaches.
Strictly spoken, the formulation of optimization above from ICRP103: “the likelihood of incurring the number of people exposed, and the magnitude of their individual doses” covers both optimization and BAT. There is nothing BAT could reasonably accomplish outside the present ICRP definition.
13 Protection of the environment
I agree that protection of the environment offers “an additional line of argument and reasoning in building a safety case, using endpoints that are different from, but complementary to, protection of human health”, but I also see complications.
The ICRP approach, as well as other approaches, is in the research phase, rather than representing a long established consensus. Compared to performance assessment, the area is less mature. Its use today will therefore no doubt be viewed by some to represent, rather than proof of safety, a collection of good intentions. Perhaps this is reflected in the passage “at least similar to the challenges of demonstrating compliance with dose/risk standards”, but these challenges from less developed area has the potential of influencing the results from the more mature areas of performance assessment.
14 General comments
The emphases on stakeholder dialogue and vagueness on time scales places the central problems of acceptance nearer the realm of politics rather than science-guided actors such as the operators and regulators.
The acceptance is judged in what resembles a traditional political process, rather than in a technical-scientific case-testing. This was also a view taken by individuals from ICRP in the 1980ties (the issue mainly being one of political acceptance).
While one must agree that performance assessment is burdened with a number of philosophical issues, much can be done lessen the distance between the two worlds, of ICRP and the performance assessment community.
Rather than offering specific advice on how performance assessment experiences should be incorporated I suggest that the performance assessment community is involved in an active way.