The composite method: an improved method for stream‐water solute load estimation

The composite method is an alternative method for estimating stream-water solute loads, combining aspects of two commonly used methods: the regression-model method (which is used by the composite method to predict variations in concentrations between collected samples) and a period-weighted approach (which is used by the composite method to apply the residual concentrations from the regression model over time). The extensive dataset collected at the outlet of the Panola Mountain Research Watershed (PMRW) near Atlanta, Georgia, USA, was used in data analyses for illustrative purposes. A bootstrap (subsampling) experiment (using the composite method and the PMRW dataset along with various fixed-interval and large storm sampling schemes) obtained load estimates for the 8-year study period with a magnitude of the bias of less than 1%, even for estimates that included the fewest number of samples. Precisions were always <2% on a study period and annual basis, and <2% precisions were obtained for quarterly and monthly time intervals for estimates that had better sampling. The bias and precision of composite-method load estimates varies depending on the variability in the regression-model residuals, how residuals systematically deviated from the regression model over time, sampling design, and the time interval of the load estimate. The regression-model method did not estimate loads precisely during shorter time intervals, from annually to monthly, because the model could not explain short-term patterns in the observed concentrations. Load estimates using the period-weighted approach typically are biased as a result of sampling distribution and are accurate only with extensive sampling. The formulation of the composite method facilitates exploration of patterns (trends) contained in the unmodelled portion of the load. Published in 2006 by John Wiley & Sons, Ltd.

[1]  V. J. Bierman,et al.  An evaluation of methods for the estimation of tributary mass loads , 1989 .

[2]  Gene E. Likens,et al.  Biogeochemistry of a Forested Ecosystem , 1977, Springer US.

[3]  Dale M. Robertson,et al.  Influence of various water quality sampling strategies on load estimates for small streams , 1999 .

[4]  Arthur H. Johnson,et al.  Estimating solute transport in streams from grab samples , 1979 .

[5]  E. S. Corbett,et al.  Comparison of Methods for Estimating Sulfate Export from a Forested Watershed , 1986 .

[6]  Robert B. Thomas.  Estimating Total Suspended Sediment Yield With Probability Sampling , 1985 .

[7]  T. Huntington,et al.  Hydrologic processes controlling sulfate mobility in a small forested watershed , 1994 .

[8]  J. E. Glynn,et al.  Numerical Recipes: The Art of Scientific Computing , 1989 .

[9]  M. Vanni,et al.  Dissolved and particulate nutrient flux from three adjacent agricultural watersheds: A five-year study , 2001 .

[10]  R. Peter Richards,et al.  Monte Carlo studies of sampling strategies for estimating tributary loads , 1987 .

[11]  Robert M. Hirsch,et al.  Mean square error of regression‐based constituent transport estimates , 1990 .

[12]  Dale M. Robertson,et al.  INFLUENCE OF DIFFERENT TEMPORAL SAMPLING STRATEGIES ON ESTIMATING TOTAL PHOSPHORUS AND SUSPENDED SEDIMENT CONCENTRATION AND TRANSPORT IN SMALL STREAMS 1 , 2003 .

[13]  B. Moldan,et al.  Biogeochemistry of small catchments : a tool for environmental research , 1994 .

[14]  Yiping Guo,et al.  Uncertainty of nitrate‐N load computations for agricultural watersheds , 2002 .

[15]  J. F. Walker,et al.  Water and solute mass balance of five small, relatively undisturbed watersheds in the U.S. , 2006, The Science of the total environment.

[16]  Robert M. Hirsch,et al.  Estimating constituent loads , 1989 .

[17]  George B. Shih,et al.  Variance of load estimates derived by piecewise linear interpolation , 1998 .

[18]  Paul D. Capel,et al.  Relations between pesticide use and riverine flux in the Mississippi River basin , 1995 .

[19]  Robert B. Thomas,et al.  A comparison of selection at list time and time‐stratified sampling for estimating suspended sediment loads , 1993 .

[20]  D. Jeffries,et al.  Hydrochemical Methods and Relationships for Study of Stream Output from Small Catchments , 2005 .

[21]  G. Likens,et al.  A Working Model for the Variation in Stream Water Chemistry at the Hubbard Brook Experimental Forest, New Hampshire , 1969 .

[22]  T. Huntington,et al.  Water, energy, and biogeochemical budgets investigation at Panola Mountain research watershed, Stockbridge, Georgia; a research plan , 1993 .

[23]  Robert M. Summers,et al.  The validity of a simple statistical model for estimating fluvial constituent loads: An Empirical study involving nutrient loads entering Chesapeake Bay , 1992 .