Response to Comments on “Problems Associated with Using Filtration To Define Dissolved Trace Element Concentrations in Natural Water Samples”

• Published in: Environmental science & technology 1996-10, Vol.30 (11), p.3398-3400
• Main Authors: Horowitz, Arthur J, Lum, Ken R, Garbarino, John R, Hall, Gwendy E. M, Lemieux, Claire, Demas, Charles R
• Format: Article
• Language: English
• Subjects: Freshwater
• ISSN: 0013-936X; 1520-5851
• DOI: 10.1021/es962006n

Shiller and Taylor do not appear to take issue so much with the scientific content of Horowitz et al. (1) as with one conclusion and particularly the new U.S. Geological Survey (USGS) protocol for inorganic constituents in filtered water (2). This is really a policy rather than a scientific issue. The USGS engages in two types of studies: (1) routine monitoring and specific projects carried out in cooperation/collaboration with local, State, and Federal agencies or (2) basic/applied research. The current regulatory definition for dissolved constituents also is the almost universally accepted operational one-materials passing a 0.45- mu m membrane filter. This protocol is now standard operating procedure for all USGS studies addressing regulatory issues. In non-regulatory-related studies, USGS personnel may select any procedure deemed appropriate, such as that used by Taylor and Shiller (3). The current operational/regulatory definition of 'dissolved' probably has little physicochemical/thermodynamic meaning. However, as long as the environmental/regulatory community opts to define dissolved constituents on the basis of a physical separation (e.g., filtration), this problem will remain (see ref 1). Evaluations of operationally defined dissolved concentrations against toxicity do not appear to exist; however, current regulatory limits are surrogates for inferred toxicity because they are supposed to be based on the statistical treatment of toxicological data. As such, why is one operational definition of dissolved more appropriate for inferring or predicting toxicity than another? Finally, fluxes in fluvial systems for artifact-affected elements tend to be dominated by suspended particulate matter-associated constituents; whereas most dissolved contributions for the same elements are minimal (4,5). Hence, one operational definition of dissolved should be as appropriate as another.