Integration of laboratory sorption and diffusivity data in modelling of sorbing tracer tests STT-1 and STT-1b

Tracer migration experiments at the 5 m scale have been performed in a single fracture at the Aespoe Hard Rock Laboratory, Sweden. The tracers used in the well-to-well tests were non-sorbing HTO and sorbing 22Na+, 58Co2+, 85Sr2+, 86Rb+, 133Ba2+ and 137Cs+. of the in situ experiments, sorption and diffusion properties of the tracers have been measured in parallel laboratory experiments. In this work, we use the laboratory sorption and diffusion data to model these tracers in 11 breakthrough curves (BTCs) from two different well pairs. BTCs were evaluated using a modified double porosity model assuming mass transfer into the rock matrix and mass transfer into stagnant zones of water. The majority of parameters in the model were obtained directly from laboratory and field measurements. A total of 4 transport parameters were estimated simultaneously from 4 of the BTC data sets. By doing this, it was possible to estimate a consistent set of transport parameters for all tracers, using laboratory-estimated sorption distribution coefficients. Forward simulations with no estimated parameters (for Ba, Cs and Co) were also carried out. Sorption coefficients measured in the laboratory were sufficient, for most tracers, to adequately model field data. While improved fits were possible using sorption coefficients fitted to field data, a good agreement between laboratory-derived and fitted sorption coefficients was observed. Limited sorption reversibility was needed in the model to fit Cs and Co data. The tails of the breakthrough curves could not be modelled using diffusion into the wall rock only. Diffusion into stagnant zones of water within the fracture plane was used for modeling and also interpreted as the main cause for the observed tailing in the breakthrough curves. The flow path was interpreted as being channelled, with a channel width of a few centimetres