Reliability of transport models calibrated on field tracer experiments: Breakthrough curve sensitivity to tracer injection conditions

The reliability of groundwater transport models strongly depends on the accurate identification of hydrodispersive processes and the quantification of corresponding parameters. Calibration of these models is often performed using results from field tracer experiments. However, little attention is usually paid to the influence of injection conditions and well-aquifer interactions on the resulting breakthrough curves at the sampling well. The tracer input function in the aquifer is often assimilated to a Dirac impulse and the breakthrough curve is supposed to only reflect the hydrodispersive processes affecting the tracer behaviour in groundwater. In fact, flow conditions, aquifer heterogeneity and well-aquifer interactions can have a significant influence on tracer injection. Using a new mathematical approach developed to model well-aquifer interactions and injection conditions in a more physical way, numerical experiments were conducted in order to highlight the potential impact of injection conditions on the results of radially converging flow tracer tests. This analysis clearly shows that neglecting the influence of actual injection conditions can lead to: (a) errors on fitted parameters, (b) misleading identification of the active transport processes.