Potential for high transient doses due to accumulation and chemical zonation of long-lived radionuclides across the geosphere-biosphere interface

Abstract

Abstract.Planning for the disposal of spent nuclear fuel is at an advanced stage in several nations around the world. Licensing of the disposal facility requires correspondingly detailed assessment of the future performance of the facility. With increased site-specific detail available to the assessment, local characteristics play an increasingly important role in determining the potential radiological risk posed by releases to the biosphere. In this paper we go beyond existing reference biosphere models and investigate the potential for specific accumulation mechanisms. The implications for the modelling carried out in long timescale performance assessment are discussed. BACKGROUNDPost-closure safety assessments for nuclear waste repositories (Performance Assessments, PAs) involve radioecological analysis of the release to the surface environment from an underground source(s). Models for PA applications have been under development for over thirty years. At early stages of a waste disposal programme generic modelling is appropriate; the international Reference Biospheres Methodology has been developed to construct stylised biospheres based on present-day biosphere analogues [1]. Traditionally, long-term safety assessments treat the geosphere and biosphere as distinct model domains, coupled only by the nuclide-specific biosphere dose conversion factor.In more advanced disposal programmes, for example in Sweden, an integrated modelling approach has recently been developed [2] in which "landscape models", based on detailed site-specific data, are used to evaluate radionuclide accumulations in the biosphere, together with associated doses. In this approach land uplift induced by climate evolution (glaciation cycles) and a number of connected biosphere objects are considered. However, the coupling of geosphere and biosphere models remains a weak aspect and structures within ecosystem models are greatly simplified [3].A transparent and credible dose assessment must demonstrate that all relevant features, processes and events have been handled appropriately in the assessment model. The focus on "landscape modelling" in contemporary assessments has overshadowed the need for suitably representative models of radionuclide transport in both the traditional biosphere objects and in deeper components of the geosphere-biosphere interface. Independent PA modelling carried out by SSM emphasises the internal dynamics of landscape objects. Once activity has reached the landscape drainage system, contaminants are lost fairly rapidly from the area around the release point. What biosphere accumulation does occur, takes place predominantly in the locality of the release [4] and the degree of accumulation depends on the nature of the ecosystem into which the release occurs [5]. Biosphere processes are inherently dispersive but some promote the concentration of contaminants in relatively small volumes.

Advanced spatio-temporal modelling in long-term radiological assessment models—radionuclides in the soil column

“…This paper shows how important features of the process-based CIEMAT redox model (Pérez-Sánchez et al 2012) can be implemented in an assessment-level model using the GEMA10 framework described by Kłos et al (2011). Although some of the processes were simplified for inclusion in the GEMA10 framework, other processes, such as bioturbation, were added.…”

“…GEMA10, described by Kłos et al (2011), is a similarly constructed ten-layer soil column model, although it was originally configured to consider the upper 1 m of soil. There are further differences both in structure and interpretation that needed to be resolved in order to compare the performance of the two models.…”

“…] for root activity in each of the layers. In the Kłos et al (2011) model, these values were derived from literature describing root profiles, providing a more robust link to field conditions than in the Pérez-Sánchez et al interpretation. To match the CIEMAT model, the GEMA10 approach uses a rate κ ip for the uptake by the roots derived from the biomass and concentration ratio K p :…”

“…Walke et al 2012), there have been recent developments in biosphere PA modelling, including an increase in the spatial resolution with which the soil column is represented (Kłos et al 2011, Pérez-Sánchez et al 2012. Kłos et al (2011) were able to reproduce features seen in experimental measurements of iodine concentrations in a drained agricultural soil profile with a near-surface water table, using a ten-layer compartment model with saturation dependent k d s in each of the ten layers; illustrative modelled and measured profiles for iodine are reproduced in figure 1. Other radionuclides, including 79 Se, 99 Tc, 94 Nb and 126 Sn, have been found to be important in dose assessments (Little et al 2009, SKB 2011 and also exhibit some degree of redox sensitivity.…”

“…The CIEMAT model is one of the more complex biosphere models used in a long-term PA context. By incorporating features from acute release models (Walke et al 2012), it contains a more detailed representation of the dynamic moisture content of each soil layer compared with the ten-soil-layer model described by Kłos et al (2011), which is already more complex than the more 'traditional' models used in the EMRAS II soil-plant system study (IAEA 2012). Differences in the capabilities of these models are further illustrated in table 1.…”