Abstract :
[en] A realistic assessment of the aquifer heterogeneity requires informative reference data and is crucial for robust transport decisions. For collecting reference data, i.e. aquifer characterization, a tracer test is an important hydrogeological tool (Maliva, 2016). Commonly, a solute (salt or dye) of known concentration and volume gets injected in the aquifer and the tracer transfer times, i.e. the spreading of the tracer, will be observed. Thus, a conservative tracer (non-reactive) is suitable for structure-based imaging and to explain fast regional point to point transport by solving the advectiondispersion equation (Singhal & Gupta, 2010). When the matrix porosity is high like in Chalk (e.g. Bodin et al., 2003) or immobile water must take into consideration for alluvial sediments (e.g. van Genuchten & Wierenga, 1976), then less diffusive solute tracer may biases badly the interpretation of the late-time tailing. Thus, for informative local reference data, structure-based imaging must be questioned and extended with process-based imaging, using preferable smart tracer like dissolved gas (e.g. He), heat or cold water (Anderson, 2005; Chatton, 2017; Kurylyk & Irvine, 2019). In multiple forced gradient experiments in fractured rocks (e.g. convergent flow field) an inflatable double packer system isolates a sub-horizontal fracture connecting two adjacent wells for injection of dissolved gas, heat and cold water. A 70 hours continuous injection of hot water (ΔT = + 40 °C) complemented by a 10 minutes uranine tracer in a chalk aquifer in Belgium shows, that the temperature signal arrives in 7.55 m distance with a delay of 12 hours compared to uranine and has a rebound once the injection stops. Thanks to the complementary behavior, the reference data consists now of fracture geometry (uraine advection) and matrix process (heat storage) information. Further to be discussed are dissolved gas (He-Xe-Ar) injections jointly with uranine in chalk or cold water in an Indian granite. Process-based imaging using smart tracer in complex aquifers is promising as it generates more informative reference data assisting on the way to robust transport decisions. A linking to geophysical imaging techniques like the Fullwaveform Inversion can be considered for example for heat. More informative reference data for innovative transport modelling requires data-science orientated prediction tools like Monte-Carlo simulation procedures within a direct predictive framework (e.g. Bayesian Evidential Learning, Hermans, 2017). This is one modelling motivation of the smart tracer reference data (e.g. Prior uncertainty investigation of alluvial sediments, Hoffmann et al., 2019).
