Estimating Contaminant Migration Pathways Using a Time Sequence of Water Level Maps and Particle Tracking

We present a method to investigate historic contaminant migration without using a detailed and intensive numerical model.  Instead, groundwater elevation maps that represent multiple discrete monitoring events are created by interpolating measured water levels in wells using Universal Kriging (kriging with a trend).   To improve the representativeness of the maps in the presence of certain sources, sinks, and boundary conditions, analytic elements are added to the kriging algorithm to represent mounding and drawdown caused by pumped wells, rivers and ponds.  Possible contaminant migration patterns are evaluated using transient particle tracking across the ensemble of calculated maps.

In the example presented, infiltration of large volumes of water occurred during operations at the Hanford site, Washington.  Water levels measured in nearby monitoring wells were combined with estimates of the stage of an adjacent river determined using a measured regression relationship developed between river stage and discharge from an upstream dam.  Multiple water level grids were generated, representing several monitoring events before, during, and after the infiltration period.  An analytic element was added for periods during which infiltration occurred, to help infer the resulting pattern of mounding.  To evaluate groundwater flow directions, particles were released from the area surrounding the infiltration, and transient tracking was used trace their paths across all calculated water elevation grids.

The results suggest that the infiltration caused significant groundwater mounding, which altered groundwater flow directions during and for some time following the period of infiltration; and that the infiltration could have provided a mechanism to transport and disperse contaminants at relatively low concentrations from their point(s) of release.