1. Test for patterns of association between distance from the site in a given direction and soil concentrations of site-associated contaminants. An entire map of the concentration field of contaminants around a site can be estimated efficiently, and boundaries and uncertainties can be calculated and displayed. The technique also supports sequential sampling strategies that make the most efficient and economical use of sample data. This step establishes statistical boundaries between background locations and locations potentially affected by the site. It also allows the discovery of probable locations of unknown sources and it automatically generates confidence bounds on the location of the boundary between background and site-affected areas.
2. Use a worst first approach to setting remediation priorities among non-background locations. This principle is both health-protective and cost-effective. It compares favorably on all dimensions to standard approaches such as cleaning up entire clusters or sectors based on average contaminant concentrations.
3. Stop remediation activities when exposures or risks have distributions that are acceptable compared to background distributions. While definitions of acceptable must typically be negotiated between PRPs and regional environmental regulators, the distribution matching philosophy provides a principled method for achieving fair, effective allocations of remediation effort.
This approach has been applied in practice with promising results in a residential neighborhood in Chicago, IL.