Validation of a Long-Term Tritium Dynamical Model

Abstract

This study addresses the issue of partitioning between organically bound and tissue free water tritium (OBT and TFWT consequently) and their modelling in plants subject to chronic releases, with major emphasis on this partitioning validation in field experiments. Under chronic releases, tritium level in plants oscillates between background environmental level and the elevated level when the atmospheric plume is present. Typical long term regulatory models are based on the quasi-equilibrium specific activity paradigm, which produces considerable uncertainty, making the model validation difficult. An alternative approach based on dynamical model describing the processes leading to tritium level oscillations is promising, as it allows for validation of process components and parameters while reducing uncertainty. New formulation of bi-directional tritium exchange under chronic releases is deployed for atmosphere-plant interactions modelling. The proposed robust and simple approach is sensitive to the accuracy of calculated evapotranspiration (ET). To minimize the uncertainty involved, the evapotranspiration and ground heat flux were accounted for using the temperature-based Reference Crop Penman-Monteith formulation (PM). The seasonal delay in the development of the summertime latent heat flux accounted via subsurface temperature in PM was found to be adequate for the Canadian Shield (CRL study site). Measurements of OBT (including OBT night-time formation) conducted in 2008 and 2009 at CRL were used for OBT model calibration. Published results dedicated to experimental quantification of OBT/HTO ratios and datasets collected at CRL in different years were used for OBT model validation. It is found, that notwithstanding enhanced predictive skill of the model (close proximity of model results to observations on a QQ plot), substantial difference from observations exist at high OBT/HTO ratios, which occur shortly after the plume departure. This suggests important contribution from fast-forming labile OBT currently ignored in the model.