Evaluation of Different Source Preparation and Selection Strategies in Chemical Mass Balance Modeling: Multiple Polycyclic Aromatic Hydrocarbon Sources to Urban Lake Sediments

Oral Presentation

Prepared by S. Ahn1, P. Boehm2, K. O'Reilly1
1 - Exponent, Inc., 15375 SE 30th Place, Suite 250, Bellevue, WA, 98007, United States
2 - Exponent, Inc., One Clock Tower Place, Suite 150, Maynard, MA, 01754, United States

Contact Information: sahn@exponent.com; 425-519-8738


EPA’s Chemical Mass Balance (CMB) model is one of several receptor models used to identify the potential source(s) and quantify source contributions to receptor (environmental sample) concentrations. Originally developed by the USEPA for allocation of contaminants in air, CMB is now widely used for chemical apportionment in various media including sediment. CMB model seeks a solution that can best fit the composition of chemicals in a sample via the linear sum of specified groups of sources. Successful CMB application requires identification of the important PAH sources that is relevant to the site under investigation along with the model’s other underlying assumptions: 1) the identified source profiles are known and stable, and 2) there are sufficient differences between sources so that they are linearly independent of one another. Ultimately, the value of a model’s output depends on the quality of the source inputs.

This presentation describes how differently prepared and selected sources impact CMB model results. CMB was used to estimate source contributions to PAH-contaminated lake sediment impacted by multiple potential sources (multiple petroleum product terminals, former MGP sites, chemical manufacturing facilities, and urban background). Strategies for source selection and model input preparation included 1) extensive literature review for candidate source materials and 2) constructing source profiles using field samples collected on and near the candidate source sites. In addition to grouping likely source samples by location, principal component analysis was used to evaluate proposed source profiles. The influence of other factors such as outliers, low concentration source samples, and weathering on source profiles and subsequent CMB results will be also described.