The Impact of Field Data Acquisition Methodology Upon Prediction of Transport in the Made Aquifer

The hydraulic conductivity K of the aquifer of the MADE transport experiment, was measured in the past at a dense three-dimensional network of sampling stations. Two methodologies were employed: initially, the flowmeter (using numerous piezometers) and recently the DPIL (direct push injection logger) method. The results for the statistical moments of the lognormal K distribution were similar for the horizontal and vertical integral scales, but quite different for the geometric mean and the variance. By using our recently developed model of transport in highly heterogeneous formations we were able to predict the observed longitudinal mass distribution spreading and to compare with measurements. We found that prediction based on DPIL parameters led to agreement, while flowmeter data resulted in discrepancy. Using our transport model, we explore here in a systematic manner the impact of the different statistical parameters upon the breakthrough curve at different distances from the injection zone.