Optimizing soil remediation volumes via inverse analytical modeling

Characterizing soil and groundwater pollution is a major environmental challenge. The aim of this work was to develop a better estimation methodology for the total volume of soil impacted by pollution. The proposed methodology employs an inverse modeling approach based on an analytical solution of flow and transport in the unsaturated zone. The model was applied sequentially to triplets of points defined via Delaunay triangulation. By combining the inversion results, a spatially heterogeneous distribution of contaminants in the subsurface was reconstructed.

This approach was evaluated via a heterogeneous synthetic case with different regular and random spatial distributions of the point concentration data to be estimated. Depending on the distribution of observation points, the method developed provides a 40–70 % reduction in the absolute error compared with that of the various spatial interpolation methods. The method is particularly useful for a reduced number of points (less than 10).