On the use of peak concentration arrival times for the inference of hydrogeological parameters

[1] Aquifer characterization is subject to large uncertainty due to spatial variability of the hydrologic parameters on the one hand, coupled with data scarcity and measurement error on the other. Traditional characterization methods rely on direct measurements of the hydrologic parameters, which usually are quite few and with large distances in between. In recent years, attempts have been made to evaluate low-cost, minimally invasive alternative sources of information that can be used to augment the direct measurements. This paper attempts to augment the data available for characterization by looking at tracer experiments. Specifically, we analyze the suitability of peak concentration arrival times to provide information useful for inference of geostatistical models of the conductivity. We found that the peak concentration arrival time is a good proxy for the advection-dominated travel time and that it is particularly useful for inference through comparison with advection-dominated, Lagrangian travel time moments when these models are expressed in terms of the geostatistical models of the conductivity. What makes the peak arrival time suitable for this is its limited sensitivity to pore-scale dispersion at early travel times and its resistance to errors induced by infrequent sampling especially at early and late arrival times. This paper analyzes this concept through analysis of a synthetic case study as well as an analysis of field data collected during a forced-gradient tracer test.