Trends, Seasonal Components, and Multivariate Relations Among Selected Water Quality Constituents at Clinton River at Mount Clemens, Michigan,
by D.J. Holtschlag and S.K. Haack
This paper presents statistical techniques for assessing trends, seasonal components, and multivariate characteristics of water-quality data. Structural time-series analyses are used to identify stochastic trends and seasonal components, and effects associated with streamflow for 11 constituents including major ions, nutrients, and fecal indicator bacteria. Multivariate relations among constituents are described by use of graphical modeling techniques. Quarterly data obtained for Clinton River at Mount Clemens, Michigan between 1974 and 1994 by the U.S. Geological Survey as part of the National Stream Quality Accounting Network
(NASQAN) program are used in this analysis. Results indicate that concentrations of most constituents analyzed followed a stochastic trend model described as a local level; in contrast, seasonal components were generally deterministic. In addition, streamflow described a significant component of the variability in 9 constituent levels. Finally graphical modeling was used to identify 14 significant correlations among the 55 possible for the 11 selected constituents. Identification of significant correlations may help improve the understanding of interactions among constituents.
Holtschlag, D.J., and Haack, S.K., 1996, Trends, seasonal components, and
multivariate relations among selected water quality constituents at Clinton
River at Mount Clemens, Michigan: 39th Conference of the International
Association for Great Lakes Research, May 26-30, University of Toronto,
Mississauga, Ontario, p. 78.
Back to USGS, WRD Michigan Home Page
U.S. Department of the
Interior, U.S. Geological
Updated On: Thursday, September 21, 2000
Privacy Statement ||