Media such as soil or sediment often contain chemicals that could have resulted from the emissions of differing sources; as such, it is useful to know who or what may have contributed which chemical in whole or in part. Often, even if the contamination occurred 20-100 years ago, the source(s) can still be identified by using various complex mathematical approaches and high speed computers. We have applied chemical fingerprinting approaches to help resolve litigation conflicts involving billions of dollars.Over the years, ChemRisk^® has pioneered the use of such techniques, thus allowing for financial responsibilities to be fairly allocated among various parties. The assignment of responsibility among the various principal responsible parties (PRPs) can be accomplished by using modeling techniques such as multiple components analysis and others.

Our experience has shown that for unmixing/source apportionment, the three most common techniques are positive matrix factorization (PMF), UNMIX (a program that uses self-modeling curve resolution for source apportionment/fingerprinting), chemical mass balance (CMB), and polytopic vector analysis (PVA). To achieve a descriptive understanding of relationships, principal components analysis (PCA), cluster analysis, and multidimensional scaling are commonly used. To classify sample data, cluster analysis and discriminant analysis are used.