Independent Consultants in Environmental and Forensic Chemistry

Volume 4, Issue 1, Winter 2001

Environmental Forensics is a hot topic. There have been numerous conferences with sessions devoted to this topic. However, these sessions have been marketing, rather than scientific, presentations. Most of these talks are associated with age-dating the release of petroleum products or hazardous substances into the environment. The American Academy of Forensic Sciences (AAFS) is the only venue I know that separates the technically defensible science from smoke and mirrors. In their own words, AAAFS is a professional society dedicated to the application of science to the law.@

This month the national meeting of AAFS has an exciting Environmental Track program in the Engineering Sciences section. There are two very hot topics presented in this meeting in Seattle, Washington. This newsletter presents the abstracts for those papers. Come listen and join in a scientific debate about important topics in environmental forensics. Information about AAFS and this annual meeting can be found on the back page of this newsletter, and at www.aafs.org.

How Reliable are AAdvanced Chemical Fingerprinting@ Methods for Oil Spill Source Identification?

Jeffrey W. Short, MS, Auke Bay Laboratory, National Marine Fisheries Service, NOAA, Juneau, Alaska

Determining the extent that a suspected oil source contributes to the hydrocarbons detected in environmental samples is difficult but crucial to natural resource damage assessments of accidental oil spills. The difficulties arise from the compositional complexity of crude and refined oils, the tendency for the composition of spilled oils to change with time (or Aweather@), and the likelihood that the sampled environment already contains hydrocarbons from other biological and anthropogenic sources. However, confident measurement of the amount of oil from an accidental spill that is present in environmental samples is the basis for assessing the extent of injury to natural resources. Despite considerable advances in analytical and statistical methods for coping with these difficulties, protocols for assessing hydrocarbon contributions from multiple potential sources are not so far advanced that other lines of corroborating evidence may be safely ignored.

Three approaches have been used to address the problems introduced by the weathering changes that are typical of spilled oils. First, the statistical analysis may be limited to the most persistent components of the spilled oil, but this usually requires an expensively detailed analysis to identify sufficient components unique to the suspected source oil. Second, relative weathering rates of components can be estimated to account for weathering losses and thereby increase the number of unique components considered, but the validty of the weathering model used must be verified independently for the case considered. Finally, ratios of components that weather at approximately the same rates may be considered, but this approach can be very misleading when multiple sources are present, especially if there are large differences in matrix effects among the contributing sources.

Because of these inherent limitations, results from multivariate pattern-recognition methods for assessing contributions of prospective source oils require independent corroboration with circumstantial facts, an evaluation of completeness, and an evaluation of the underlying assumptions used. Relevant circumstantial facts include consistence with mass-balance constraints and with plausible transport mechanisms from the site of spillage to the sampled environment. Completeness requires that all prospective sources are evaluated as potential contributors of oil to an environmental sample, and that all the analytical results produced are consistently accounted for. Finally, any underlying assumptions should be explicitly justified. These concerns amount to satisfaction of the three elements of scientific proof, viz., consistent association of the cause with effect, stipulation of a mechanism consistant with established physical laws, and definitive refutation of all possible alternative explanations.

The case of the 1989 Exxon Valdez oil spill furnishes an example of the limitations of chemical fingerprinting methods that ensue when these fundamental scientific issues are not fully addressed. The deeper subtidal sediments of the spill area contain a characteristic hydrocarbon fingerprint that included polycyclic aromatic hydrocarbons (PAH). This fingerprint indicated the hydrocarbon source was clearly unrelated to the spilled oil, and the geographic extent of the fingerprint ranged far to the east of the spill area. Scientists working for Exxon corporation concluded that the source of this regional hydrocarbon background was terrestrial oil seeps along the coastline of the Gulf of Alaska (GOA) east of the spill area. They calculated that the equivalent of 3% of the oil spilled from the T/V Exxon Valdez enters the spill area annually from these seeps. If true, this alternative source of bioavailable hydrocarbons would confound assessments of biological injury, because biota in the region would be continually exposed to this Anatural oil pollution.@ However, the Exxon scientists failed to demonstrate that oil from the seeps actually reaches the GOA, or how the seep oil could weather enough to sink while at the same time conforming to the composition fingerprint characteristic of the background. Also, they did not consider all the natural hydrocarbon sources in the region, nor all of the analyte information available. Attempts by government scientists to corroborate the Exxon-sponsored work led to the identification of natural coal eroded from terrestrial deposits as a major hydrocarbon source in this region, and further work by Exxon scientists revealed hydrocarbon source-rock as the other likely major source. Hydrocarbons in coal or in source rock are strongly sequestered by the matrix, and hence are not biologically available so they do not constitute a confounding hydrocarbon source for assessments of biological injury following PAH exposure.

The GOA east of the spill area is almost entirely pristine, with only three natural hydrocarbon source classes that might contribute to the majority of the background hydrocarbon fingerprint: seep oil, coal, and source rock. Despite the relative simplicity of this system, the initial source identification that was based almost entirely on analyte ratios presumed diagnostic and derived from a subset of the available analytical data turned out to be substantially incorrect. This underscores the need for close scrutiny of hydrocarbon source identification claims in damage assessment cases. In addition, the need for standardized quality assurance criteria for evaluating multivariate source identification schemes is urgent, particularly blind performance tests. At present, the reliability of these schemes is largely unknown.

Using the Chemical Fingerprint of Pharmaceutical Compounds to Evaluate the Timing and Origin of Releases to the Environment.

Remy J-C. Hennet and Laura Chapp, S.S. Papadopulos & Associates, Inc.

Until very recently the release of pharmaceutical compounds to the environment has been largely unrestricted and ignored by environmental professionals and regulators. The release of pharmaceuticals has been on-going for several decades and has gradually increased in parallel with economic development. As a result of these long established releases, pharmaceutical compounds and their metabolites [pharmaceuticals] are present in surface water, groundwater, soils and sediments. The sources of pharmaceuticals to the environment include excretion of unreacted compounds and metabolites by human or animal users, the direct disposal of leftovers, and the effluents at manufacturing facilities. The pathways of release include cesspool and septic systems, public and private sewer systems, the outfalls of waste water treatment plants, landfills, as well as the direct release of industrial, human and animal wastes to surface and subsurface water bodies.

Numerous pharmaceutical compounds are produced and consumed in high volumes. A significant portion of this production is released to the environment. Pharmaceuticals can be persistent and are often mobile with water. The analysis of pharmaceuticals has not yet been developed as routine procedures available at commercial laboratories that service the environmental industry. Conventional analytical methods for the detection of trace chemicals in the environment are mostly inadequate for the detection of pharmaceuticals. This situation is changing because of a growing research interest to characterize the potential impacts and toxicity of pharmaceutical residuals in the environment. This new interest coupled with advances in analytical chemistry will soon make the acquisition of data and information for fingerprinting interpretation possible and cost effective.

Pharmaceutical fingerprints can provide the reliable data that are necessary to better estimate or quantify the timing and history of releases to the environment. Pharmaceuticals are regulated substances and quantitative information is available for the date of introduction (or withdrawl) of individual compounds to the consumer market. This information allows for the use of pharmaceutical fingerprints to constrain the timing and history of releases with a degree of certainty that can greatly improve upon conventional approaches. Pharmaceutical fingerprints can also provide reliable data to determine contaminant pathways and sources in complex environmental settings. Pharmaceutical compounds and their metabolites can be viewed as tracers of migration pathways that carry a specific signature from a source area. As such, these compounds join other environmental tracers and increase the tool baggage for this type of environmental forensics.

The combination of tracing and timing capabilities for pharmaceutical fingerprints is conceptually a powerful tool that is in the process of being developed. Conventional approaches to investigate the origin, migration, and history of environmental releases include calculations and simulation models, the review of site-specific information, and the opinion of professionals. These methodologies are most often limited in their accuracy and reliability and would benefit from the information that can be derived from pharmaceutical fingerprints. A quantitative understanding of the source, migration, timing and history of chemical releases is of great importance. To characterize and remedy environmental contamination it is primordial to understand the origin, fate, and transport of contaminants. Absent such understanding remedial actions cannot be properly designed and optimized. In the litigation arena, it is common to encounter situations where the history of events is crucial to determine and allocate responsibilities and costs. The monitoring of landfills and new industrial developments on previously contaminated properties could greatly benefit from an approach that can differentiate between on-going and pre-existing releases. The details and limitations of the approach will be presented.