Intercomparison of model predictions of C concentrations in agricultural plants following acute exposures to airborne C.

Carbon-14 (C) is one of the main radionuclides released during normal operation by nuclear power plants, nuclear defense facilities and nuclear fuel reprocessing plants. It is mainly released in the form of carbon dioxide gas denoted CO, which has the specificity of being incorporated into food webs via photosynthesis by primary producing organisms. In order to better assess the environmental and human impacts of CO under normal operating conditions - or after potential accidental releases - from nuclear facilities, it is necessary to improve our understanding and our predictions of the behaviour of this radionuclide along the human food chain. To achieve this goal, the International Atomic Energy Agency (IAEA) Environmental Modelling for Radiation Safety (EMRAS) model evaluation programme included the Tritium and C Working Group (TCWG) which dealt with the intercomparison exercises between several models of environmental transfer in the case of routine and accidental releases of these radionuclides into the environment, and their performance testing. The TOCATTA-χ model developed at IRSN is a dynamic compartment model with high temporal resolution, which simulates the transfer of C (and tritium) in grassland ecosystems exposed to gaseous CO (and HTO) from nuclear facilities under normal or accidental operating conditions. Following this work, IRSN proposed a related project to extend the application of the TOCATTA-χ model to C estimates in leafy vegetables, fruits and roots. This article deals with the application of the TOCATTA-χ model to a specific real-case scenario identified within the framework of the TCWG. The scenario provides experimental data and predicted results from models developed at the international level. Model-model and model-data intercomparison exercises were thus carried out to validate the evaluations of the TOCATTA-χ model. In addition, this paper discusses the parameterization of the TOCATTA-χ model for this scenario and the development of modules for C concentrations in potato tubers, based on the assumption that photosynthetic transfer occurs directly from leaves to tubers and depends mainly on the growth stage of the tubers. It is observed that the predictions of the TOCATTA-χ model for the concentrations of C in leaves and tubers are slightly better than the other models due to the modelling approaches adopted by TOCATTA-χ for the calculation of key ecophysiological processes that govern plant functioning. Overall, the TOCATTA-χ model reduces the Root Mean Square Error (RMSE) by a factor of less than 8 compared to other models. In addition, most of the predicted results of the TOCATTA-χ model better match the measurements and are within the measurement uncertainty limit, while a few are overestimated. This could be due to the high uncertainty associated with the experimentally measured C activities, which reflects the field variability in plant growth rate.