NATURAL ANALOG STUDIES - PRESENT STATUS AND PERFORMANCE ASSESSMENT IMPLICATIONS

Studies of natural geological and archaeological systems as analogues to long-term processes, which are predicted to occur within a radioact ive waste repository environment, have become increasingly popular ove r the last 10 years or so, to the extent that such studies form an int egral part of many national programmes for radioactive waste disposal. There is now a common consensus that the natural analogue approach is a very useful scientific methodology to: (a) identify and understand processes and mechanisms analogous to those which could occur in the v icinity of a repository over realistic timescales, (b) derive input da ta which have been successfully used to test some of the laboratory-ba sed models which form the basis of long-term repository performance as sessment, and (c) to produce data which can be input directly to perfo rmance assessment models. Increasingly, analogues are playing an impor tant role in public awareness, enabling the layman to understand bette r the concept of radioactive disposal and demonstrating the reliabilit y of the disposal system over long periods of geological time. The com plexity of geological systems means that it is very often difficult an d sometimes impossible to quantify precisely the physico-chemical boun dary conditions necessary to model a particular geochemical process or mechanism. Consequently, the availability of quantitative analogue da ta is Limited when repository performance assessments are considered. However, this in no way detracts from their value in building confiden ce by demonstrating that important processes do exist and by showing q ualitatively that they behave in a way predicted by models based on la boratory-derived data. The transfer of natural analogue data from the complexity of field studies to simplistic models which, by necessity, are used in performance assessments, is an area of activity which is p resently being addressed. Field analogue studies are now being planned to interface with laboratory experiments and, ultimately, with in sit u field experiments. This is a promising avenue of research which shou ld provide a more quantitative use of natural analogue data in testing and further developing models used in (or supporting) repository perf ormance assessments. (C) 1997 Elsevier Science B.V.