WSP11764

• File Number: 8273.100
• Description: Colorado River Basin Salinity Control - Federal Agencies - Bureau of Reclamation
• Basin: Colorado Mainstem
• Water Division: 5
• Date: 1/1/1984
• Author: BOR
• Title: Water Quality of the Colorado River System - Historical Trends in Concentration/ Load/ and Mass Fraction of Inorganic Solutes
• Water Supply Pro - Doc Type: Report/Study

,....
o
o
o

For streamflow interpolation equation 1 is linear
with respect to time.

Conductance and concentration are transformed
by the weighting factors into parameters(ak andbk
in equations 2 and 3) that are expected to be
relatively constant over short periods. For inter-
polation of conductance, equation 1 becomes a
linear interpolation, with reSpect to time, of the
parameter ak in the equation:

B
Lk=akQk

Lane [9] found this power relationship to be
generally applicable to natural streams in the
Western United States. Conductance interpolation
was not attempted when the missing value period
included days of no flow.

The interpolation of solute concentration is a
linear interpolation of the parameter bk in the
equation:

Ck = bk Lk

Over small ranges, this linear relationship provides
a good estimate of concentration. After inter-
polation, solute concentrations were adjusted to
account for differences between the instantaneous
specific conductance (Lsl at the time of sample
collection and the mean daily conductance (Lo), if
the latter value was available.

Lo
Ck=CkT;

This adjustment was made only for dates with
continuously recorded specific conductance. It
was made to correct potentially nonrepresentative
instantaneous sample concentrations to average
conditions on the sample date.

Interpolations were limited to intervals of 20 days
for streamflow and conductance, and 60 days for
solute concentrations. The effect of the length of
the missing value period on the accuracy of
specific conductance estimates was tested using
data from the Colorado River at Cisco, Utah.
Artificial gaps of varying length were created in the
conductance record, and generated estimates were
compared with measured values. Standard errors
were 6.8 and 10.2 percent of the observed mean
for 7-day and 20-day intervals, respectively.
Although increasing the interpolation interval
decreased the estimation accuracy, ala percent
standard error was not considered unreasonable.

(2)

The effect ofthe length ofthe missing value period
on the accuracy of solute concentration estimates
was tested using data from the discrete sampling
period for the Colorado River at Lees Ferry, Arizona.
Increasing the maximum period from 45 to 60
days resulted in change of only 3 percent in the
average monthly concentrations for the sum of 6
ions. In addition, the number of months with
inadequate data for computing monthly means, as
defined below, was reduced by 21 percent. To
maintain representative data on the maximum
number of months, the interpolation interval was
set at 60 days, which allows interpolation between
any two samples collected in consecutive calendar
months. Testing the accuracy of streamflow inter-
polation was considered unnecessary because of
the infrequent occurrence of gaps in these data.

(3)

Following interpolation, monthly mean values of
streamflow and flow-weighted concentrations
were computed from daily values. When measured
or estimated values were unavailable for more
than 25 percent ofthe days in a month, analysis of
that month was discontinued. Therefore, the
resultant monthly values are intended to preserve
the maximum amount of information from existing
data without masking long-term trends or including
unreasonable estimates. Monthlyvalues are listed
in table 3 of appendixes B through P.

Analysis of Seasonality

(4)

Prior to statistical analysis, the generated monthly
data were tested for seasonal variability. Normal-
ized deviations (Ll.) from the annual mean were
computed for each month using the formula:

Xy.m-Xy
Ll.= -
Xy

(5)

where: X = streamflow, solute concentration,
or load
X = the annual mean of each
parameter
y = year
m = month

The average monthly values for Ll. were then
plottedtodeterminetheseasonal pattern. Figure 3
shows the seasonal variation in streamflowforthe
Colorado River at Cisco, Utah. A peak runoff period
in May and June is clearly indicated. April andJuly
are transition months, and the remaining months
appear to constitute a baseflow period. However,
chemical variation, shown in figure 4, indicates a

7