Flow in the unsaturated zone around a shallow subsurface radioactive waste trench: Interpretation of an infiltration–drainage test at the Chernobyl Pilot Site

This article describes an infiltration–drainage test carried out in the unsaturated zone (UZ) at the Chernobyl Pilot Site during October 2008; this is an international radioecology study site and is the subject of several papers in this special issue. The test has to be seen in the larger context of radionuclide transport from a waste trench. The conducted experiment consisted of infiltrating a layer of 9.5 cm of water in a circular area of 5.51 m2 over 5 h. Its main objective was to create a larger range of water content values (and hence suction pressure values), not only at the top of the soil profile but also at greater depths, in this case up to 1.50 m. Observations of water content and suction pressure were carried out continuously at seven different depths during infiltration, drainage and during the return to natural conditions over a period of several months. This allowed deriving UZ parameter values with greater confidence than those derived from monitoring small natural water content changes over periods of years.The experiment also shows that a 1D model was incapable of correctly reproducing the observed water balance established on the central axis. A 2D axi-symmetric model was needed showing that lateral boundary effects have to be accounted for. This implies that small scale infiltrometer tests should be analysed in more than one dimension. It further became clear from the drainage curves that soil layering played a significant role and that in the present case hysteresis did not play a major role; i.e. the infiltration and drainage event can be reproduced with the same set of parameters and subsequent natural rain events can be reproduced well enough with two numerical models used for test interpretation, one is module oriented for reactive transfer, and the second with an automatic optimisation procedure. The robustness of the estimated parameter values, of the model discretisation and layer identification was tested over a 2-month period with natural infiltration using a 1D model. The results validated the test outcome.

► Hysteresis does not play a major role in this case (sandy deposits). ► 2D axi-symmetric model is needed to reproduce field observations, i.e. boundary effects play a role. ► Geological layering needs to be accounted for. ► Optimisation of parameter values carried out with HYDRUS gave sensible results. ► Numerical model validated by a forward 1D model of the test site under natural circumstances.