Learned Discourses: Timely Scientific Opinions.

Abstract

Listen

Translate article

The fate of a chemical in the environment is primarily controlled by the physico-chemical properties of the substance, prevailing environmental conditions, and patterns of use (Mackay et al. 1997). Thus, accurate determination of physico-chemical properties is critical to the formulation of valid environmental models and assessments (Bennett et al. 2001). As litigation becomes an increasingly common aspect of environmental science, such studies may be challenged and even refuted. Recent legislation in the United States mandates that federal agencies must ensure the quality of data upon which regulatory action and policy decisions are founded (Reichhardt 2002). Some observers have suggested that because this legislation provides for unprecedented scrutiny of agency data quality, regulatory agencies might face a marked increase in the number of legal challenges. Two of the most important physico-chemical properties relating to the environmental behavior of hydrophobic organic compounds are aqueous solubility (S w ) and the octanol-water partition coefficient (K ow ). Much of the interest in these properties comes from the fact that they can be used to estimate or predict other properties of more immediate environmental and ecotoxicological interest. For this reason, a primary focus of research in the last 20 years has been development of linear regression models in which S w and K ow (or more accurately, their logarithms) are correlated with parameters such as the organic carbon-normalized partition coefficient, bioconcentration and bioaccumulation factors, and indices of biodegradability or toxicity. These, in turn, can be incorporated into models that attempt to characterize the equilibrium distribution and transport rates of organic contaminants among environmental media or to predict impacts on biota. The predictions of such multimedia models have taken on greater immediacy with the recognition that certain persistent organic pollutants, such as DDT and DDE, are globally distributed and may be exerting adverse biological effects far from their original source areas (Wania and Mackay 1996). Our interest in the K ow and S w values of DDT and DDE arose in the context of research we were conducting on the geochemical fate of these compounds in sediments of the Palos Verdes Shelf, California. We expected that studies reporting S w and K ow values for DDT and DDE would be well documented because DDT has a very long history, and its widespread application and biological effects are well known. Results of our initial literature survey, however, were unsatisfactory due to a large spread in data values. We decided to undertake an exhaustive review to identify all measured and estimated values of S w and K ow for these compounds. It soon became apparent that a number of serious problems severely limited use of the database and that our initial expectations had been naive. A complete description of the study discussed here can be found in a USGS report (Pontolillo and Eganhouse 2001) available in PDF format on the web (http://pubs.water.usgs.gov/ wri01-4201/). We examined more than 700 publications including databases, handbooks, review articles, and bibliographies for data, methodological details, and pertinent literature citations. The initial survey was augmented by computerized searches of the scientific literature using the Chemical Abstracts Search Service Index (CASSI) database and the World Wide Web using a variety of Internet searchengines. Several hundred potential articles were generated in this fashion. Each article was retrieved and scrutinized, and additional relevant leads were identified from the reference lists. The search progressed until no further leads could be identified (approximately 2 years). Two problems are evident with the available database: egregious errors in reporting data and references, and poor data quality and/or inadequately documented procedures. The published literature is characterized by a preponderance of unnecessary data duplication. Numerous data and citation errors are also present. The percentage of original S w and K ow data for DDT