DESCRIPTION OF THE SCIENTIFIC RESULT / CONCLUSION: Geosphere retention of radionuclides can be an important element of the safety case for deep geological disposal of radioactive wastes. Methods to evaluate and model the migration of radionuclides in the geosphere have been developed intensively over many years. The RETROCK project, "Treatment of Geosphere Retention Phenomena in Safety Assessments", has examined how the retention and transport of radionuclides are dealt with in performance assessment (PA) models for deep geological repositories and how this could be further developed and improved. The environment considered in RETROCK is saturated, sparsely fractured rock in the so-called 'far-field' (between the repository and the biosphere), where the interconnected fracture network provides flowpaths for groundwater that could transport radionuclides. The results of the project also are applicable, with some constraints, to other geological environments. The phenomena in the focus of RETROCK are related to matrix diffusion, sorption and the distribution of groundwater flow. These, along with radioactive decay, provide the key terms to the transport equations of PA models. Some other processes are discussed but more cursorily: colloid-facilitated transport, the influence of microbes, gas-mediated transport, precipitation and dissolution, off-diagonal Onsager processes, effects originating in the near-field and the impacts of climate change. The RETROCK participants acknowledge that the currently applied priorities to develop the treatment of the processes under discussion are well reasoned - considering also that the project did not reveal new issues that could call for significant reprioritisation. It looks likely that further work on retention mechanisms and modelling will be at the level of gradual improvements within the existing main directions. The participants believe that sorption, matrix diffusion, the distribution of groundwater flow, and radionuclide decay continue to be the main processes that need to be dealt with in future PAs. A high level of consensus exists on geosphere retention and transport processes, their definition and their generic importance. In recent years, PAs have been able to handle these processes adequately, if not in all cases explicitly, at particular stages of programmes. Basic data may be lacking or site-specific data may be difficult to obtain also for processes for which the level of understanding is good. In such cases, comprehensive utilisation of the available information to substantiate the choice of parameters for each set of PA calculations is essential. Supporting models are expected to handle processes with better coverage of phenomena and with more realism, as more data will become available. They can be utilised in selecting PA parameter values so that the top-level PA models produce conservative results, which is a necessity in the demonstration of compliance. On the whole, the assessments have been considered fit-for-purpose at the time they were produced and have not subsequently been significantly challenged. However, as a repository programme moves towards implementation there will be a greater need for completeness and adequacy of the models and data as well as for supporting arguments. DISSEMINATION OF THE RESULT: The WP3 report was published as the report EUR 21230 EN. USE OF THE RESULT: The report serves as a tool for directing the development of PA modelling effectively.
DESCRIPTION OF THE SCIENTIFIC RESULT / CONCLUSION: The processes emphasised in the report are the influences of the flow field, matrix diffusion, and sorption on radionuclide transport characteristics. These processes, and radioactive decay, provide the key terms to the transport equations of the performance assessment (PA) models. The following processes are handled more cursorily: colloid-facilitated transport, microbial activity, gas-mediated transport, precipitation/coprecipitation, and off-diagonal Onsager processes. The PA models usually take into account retention phenomena using simplified concepts that are justified by their conservatism. They are complemented by a large variety of more detailed and realistic process-specific models that can be used to support the choice of data for PA models, as well as specific arguments made in safety cases. While the fundamental understanding, the conceptualisations of the phenomena, the models and the computing resources develop, the extensive data requirements often become a most restrictive factor to the use of a model. The difficulties in obtaining data tend to hinder the utilisation of the most realistic mechanistic models, but major challenges are also encountered in obtaining relevant and defensible data for the principal simplified PA models. All processes discussed in the report were mentioned in the responses to a questionnaire that was used in WP1 to map the views and the ways to handle the topic in several published safety cases. DISSEMINATION OF THE RESULT: The WP2 report was published as SKB Report R-04-48. In addition, the sections on matrix diffusion were reported as the following public report: Jakob A. (2004), Matrix Diffusion for Performance Assessment - Experimental Evidence, Modelling Assumptions and Open Issues. Nagra Technical Report 04-07 and Paul Scherrer Institut Bericht Nr. 04-08, July 2004. USE OF THE RESULT: The report served as a tool for further judgement of understanding and approaches, and for an assessment of development needs. Thus it was a major input to WP3. In addition, the draft was used to acquire feedback to RETROCK from external reviewers and the other attendees of the Specialist Workshop at the midterm of the project.
DESCRIPTION OF THE SCIENTIFIC RESULT / CONCLUSION: - There is a high degree of confidence in the completeness of the list of potentially relevant transport and retention processes, although this list cannot be considered complete in a rigorous sense and new facts and new interpretations of results can never be excluded. - The general approach is to model flow with a high degree of detail, aiming at a realistic representation of the flow field, and to adopt a more simplified approach to the modelling of transport. - Transport is generally modelled along one or more transport paths or streamtubes identified by flow modelling, with transport parameters spatially averaged along these features. - There are very few differences between PA transport models in the transport and retention processes that are included, the principal difference being the inclusion or exclusion of colloid-related processes. - A number of processes potentially contribute to transport and retention, but are currently excluded from transport models, mostly due to lack of the data necessary for modelling and mostly justified on the grounds of conservatism (e.g. sorption on fracture surfaces and infill, diffusion into stagnant water caused by constrictions in fracture openings, diffusion into fracture infill or gouge material, immobilisation processes). - There is generally satisfaction with the adequacy of fundamental understanding of the processes included in transport models (at least for PA modelling purposes), although an improved mechanistic understanding of sorption would be desirable. - The processing or upscaling of data in order to obtain PA model parameter values is an area that warrants more attention in the future. - The lack of, or limited, treatment of time-dependence is generally seen as a weakness in most current PA transport models. DISSEMINATION OF THE RESULT: The WP1 report was produced as Nagra's Internal Report but it is available on request from Nagra. USE OF THE RESULT: The result was fed as an interim result into the WP2.