[en] The BIOMOVS II Working Group on Reference Biospheres has focused on the definition and testing of a methodology for developing models to analyse radionuclide behaviour in the biosphere and associated radiological exposure pathways (a Reference Biospheres Methodology). The Working Group limited the scope to the assessment of the long-term implications of solid radioactive waste disposal. Nevertheless, it is considered that many of the basic principles would be equally applicable to other areas of biosphere assessment. The recommended methodology has been chosen to be relevant to different types of radioactive waste and disposal concepts. It includes the justification, arguments and documentation for all the steps in the recommended methodology. The previous experience of members of the Reference Biospheres Working Group was that the underlying premises of a biosphere assessment have often been taken for granted at the early stages of model development, and can therefore fail to be recognized later on when questions of model sufficiency arise, for example, because of changing regulatory requirements. The intention has been to define a generic approach for the formation of an 'audit trail' and hence provide demonstration that a biosphere model is fit for its intended purpose. The starting point for the methodology has three. The Assessment Context sets out what the assessment has to achieve, eg. in terms of assessment purpose and related regulatory criteria, as well as information about the repository system and types of release from the geosphere. The Basic System Description includes the fundamental premises about future climate conditions and human behaviour which, to a significant degree, are beyond prediction. The International FEP List is a generically relevant list of Features, Events and Processes potentially important for biosphere model development. The International FEP List includes FEPs to do with the assessment context. The context examined in detail by the Working Group assumes a requirement to calculate annual individual doses arising from long term release of radionuclides in groundwater at an inland site. Though to a degree limited, this context is broadly relevant to many assessments and so the FEP List may be useful as a generic starting point for new model development or for auditing existing biosphere assessments. An example illustration of how to apply the methodology has been provided, based on the case description of the Complementary Studies Working Group of BIOMOVS II. (The results of the Complementary Studies exercise are provided in a separate BIOMOVS II Technical Report.) The application of the International FEP List to the Complementary Studies case description has been examined. A major component of the methodology is the development of a conceptual model from the available information about processes and the related data. Several approaches are discussed, including 'process influence diagrams', 'event trees', and the Rock Engineering System (RES) Interaction Matrix Methodology. The latter was tested in some depth by the Working Group and found to be effective at helping to distinguish the more and less important FEPs, and in identifying the important interactions between components of the system being modelled. An interesting feature of the interaction matrix methodology is that the FEP List applied to the matrix construction process can be developed independently from the matrix. If the matrix were to be developed by the same people as the FEP List then there could be legitimate criticism that the matrix, and hence the conceptual models, had been designed in a closed loop of assessment modelers. However, any issue can be introduced by any interested party through the FEP List. The auditing step allows these other issues to be introduced in model development and so forces the assessment team to explain how the issues are to be dealt with. At the same time, previous experience within the assessment team in modelling, radioecology etc, can be readily incorporated. Iterative cross-checking of the interaction matrix and FEP List contents is regarded as an important part of the procedure. FEP Lists of the type referred to above can be developed for specific assessments, eg, through applying the interaction matrix methodology. An example of the type of software tool which can be used to maintain and extend a FEP List has been developed within the Group. It is called BIOFEP and is described in an Appendix. Detailed assumptions about migration and accumulation of radionuclides in biosphere media with which humans interact will be strongly dependent upon the assumptions which have to be made about the individuals or population groups for whom radiation doses are being assessed. The Working Group has reviewed these 'critical group' assumptions and found considerable variability in both regulatory specification and in performance assessments designed to meet regulatory objectives. Consistency in approach to regulation and assessment is to be desired. While the Working Group has been broadly successful in setting out an appropriate methodology and providing useful input to model development in terms of FEPs and application experience, further activities are recommended. In summary, these involve further testing and augmentation of the methodology: to consider a wider range of basic system descriptions; to more fully develop conceptual models according to the methodology; to examine other types of release from the geosphere; to develop principles for critical group definition; and to develop principles for applying field and other data in defining parameters used in models to represent processes