Endorsement of Sediment Records Calculation Program (SEDCALC)

In Reply Refer To:                                    August 1, 1995
Mail Stop 415


Subject:  Endorsement of Sediment Records Calculation Program (SEDCALC)

Techniques used by the U.S. Geological Survey to compute daily records 
of suspended-sediment discharge changed little from the 1940's to the 
1970's. In the 1980's, several programs for computing sediment dis
charge were developed independently in district offices of Water 
Resources Division (WRD). Under the direction of John Gray, several of 
these programs were combined to form SEDCALC, an interactive, menu-
driven set of programs developed to facilitate computations of daily 
suspended-sediment discharge records.

The purposes of this memorandum are to announce the publication of 
the documentation for SEDCALC, to endorse its use within WRD to 
calculate daily records of suspended-sediment discharges, and to 
provide guidelines for its use.

The documentation of SEDCALC has been published as:

   "User's Manual for SEDCALC, A Computer Program For Computation of
   Suspended-Sediment Discharge" by G.F. Koltun, John R. Gray, and
   T.J. McElhone, USGS Open-File Report 94-459, Columbus, Ohio, 46

SEDCALC enables one to use an entire streamflow time series to 
compute instantaneous and daily suspended-sediment discharges, 
thereby eliminating computational errors that may result from use of 
a subset of the streamflow time series in manual subdivision for 
periods of rapidly varying conditions. Other advantages of SEDCALC 
over traditional techniques are that it (1) speeds record computation 
by automating some of the labor-intensive tasks associated with 
manual sediment-record computations, (2) provides a uniform set of 
tools for sediment-record computations, (3) facilitates exploratory 
data analyses, and (4) facilitates the archiving and restoration of 
data for later analysis and review.

SEDCALC provides four methods for developing a time series of 
suspended-sediment concentrations on identical intervals as the 
associated discharge time series to calculate suspended-sediment 

 (1) By using a digitilized curve of sediment concentration versus time.

 (2) By log-linear interpolation in time between measured and estimated
     suspended-sediment concentration values,

 (3) As in (2) above, but using a non-linear cubic-spline interpolation
     scheme that approximates the shape of a hand-drawn curve, and 

 (4) By log-log transport relation with potential adjustment for
     retransformation bias.

Any of the first three options are acceptable for calculating loads 
so long as data used for input have been analyzed using methods 
described by Porterfield, TWRI, Book 3, Chapter C3 , (1972) and the 
appropriate checking procedures have been performed. 

It is recommended to have at least a full year of record, and a large 
number (>75) of observations of instantaneous suspended-sediment 
concentration covering the range of flows that are to be estimated 
before using the transport-curve option to compute primary record. It 
seamiest appropriately used for estimating the shape of the 
suspended-sediment trace for periods not exceeding several days at a 
time, and even then only for periods where land-use and hydrologic 
conditions are consistent with those existing when data used to 
define the transport curve were collected. 

Proper use of either log-linear or non-linear interpolation requires 
that a sufficient number of instantaneous suspended-sediment 
concentrations--either measured or estimated--are available to 
adequately approximate the sediment trace. Particular caution should 
be exercised when using the non-linear interpolation scheme, as it 
may yield erroneously large or small suspended-sediment 
concentrations if the sediment trace is inadequately defined.

The log-log transport-relation option should be used for computing 
discrete periods of primary record only when there are insufficient 
data for application of one of the first three options, or until a 
sufficient set of comparative record has been calculated to evaluate 
the performance of this option. 

A bias-correction factor (BCF) must be applied to sediment-
discharges calculated by means of the log-log transport relation 
option if the equation for the transport relation was determined by 
means of regression on log-transformed variates. Either of two 
methods -- Minimum Variance Unbiased Estimator (MVUE), or Duan's 
Smearing Estimator -- may be appropriate (OSW Memorandum 93.08). They 
are used when the distribution of errors are normal and non-normal, 
respectively. A statistical routine based on the statistics package, 
Statithas, has been developed to assist in determining the parameters 
required to perform the bias correction. It is critical that results 
of BCF calculations be reviewed by a surface-water specialist or 
other person knowledgable in hydrology and statistics before using 
the results to calculate suspended-sediment discharges.

Greg Koltun, Ohio District, is responsible for supporting SEDCALC. 
Please direct any questions about obtaining or using SEDCALC to him 
by E-mail to gfkoltun@usgs.gov, or by telephone to 614/469-5553. A 
limited number of users manuals are available fromthe Ohio District.

                                Ernest F. Hubbard
                                Acting Chief, Office of Surface Water

WRD Distribution: A, B, FO, PO