Multivariate Statistical Methods for Sediment Contaminant Exposure and Effects

• Activity Number: 424
• Type: Contributed
• Date/Time: Tuesday, August 11, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics and the Environment
• Author(s): Margaret Nikolov* and Virginia Brodie
• Companies: U.S. Naval Academy and USNA, Marine Corps
• Keywords: marine ecosystem ; sediment ; contaminant ; toxicity ; factor analysis

Abstract

Marine ecosystems along the Southern California coast have long been adversely impacted by sediment-associated contaminants. Thus, a major endeavor with regard to environmental monitoring and management is the determination of protective sediment quality guidelines. To this end, researchers seek to quantify both the extent of sediment contamination and the impacts on surrounding benthic life. While there exist vast amounts of data on sediment chemistry and toxicity measures, methods of exposure assessment adept at handling the complexity of the marine ecosystem warrant development. In this paper, we propose the use of multivariate statistical methods, including principal component analysis and factor analysis, to assess aggregate exposure and effects of related contaminants. We demonstrate these methods on data from the 1992-1997 Bay Protection and Toxic Cleanup Program study and find significant reductions in the 10-d amphipod survival of Rhepoxynius abronius associated with exposure to metals (copper, mercury, zinc), pesticides (alpha-Chlordane, trans-Nonachlor), and pcbs (congeners 018, 028).