8276

• UCREFRP Catalog Number: 8276
• Author: Williss, C. C.
• Title: Documentation of Temperature Model Prepared by the Instream Flow and Aquatic Systems Group, in Cooperation with the U.S. Soil Conservation Service and the U.S. Fish and Wildlife Service.
• USFW Year: 1984.
• USFW - Doc Type: Denver, Colorado.
• Copyright Material: NO

The COLHYD program mentioned previously was used to calculate
intervening flows in the network. Some tributary gages had
missing flows, especially in the months of November and March.
The COLHYD program inputs a value for missing data which
represents the arithmetic mean for the time period.
0
0
The final network for the temperature model contains 78 nodes
(Table 2). A model run was made using the historical and
synthetic gage data.
Results
Validation of the temperature model has been completed for other
studies. Theurer and Voos (1982) present the results of their
validation study for the same network covering the 1964 to 1977
period. Their results indicate an overall temperature bias of
-0.12 degrees Celsius with a corresponding probable difference of
0.64 degrees on an annual basis after model calibration. This
means that 50% of all predicted temperatures are between
-0.76 and 0.52 degrees from the published water temperatures.
When the time period was subsequently expanded to cover 1960 to
1980, the bias was -0.26 degrees and the probable difference was
0.96 degrees.
For our analyses, comparisons were made between observed and
predicted mean daily water temperatures (1930-1982) at the Green
River and Jensen gaging stations on the Green River and at the
Cisco and Stateline gaging stations on the Colorado River.
Results from the comparisons appear in Appendices C. D, E and F.
Statistical evaluations of observed versus predicted mean daily
water temperatures have also been completed for the period of the
observed temperature record.
At the Cisco gage, the model under-predicted average water
temperatures (1959-1982) for the months of October, November,
December, January, April, and September and over-predicted the
other months. The average absolute difference between observed
and predicted temperatures at Cisco was 1.05 degrees Celsius
(Appendix C).
At the Stateline gage, the model tended to over-predict average
daily water temperatures (1963-1982) for the months of December,
August, and September. The other months were under-predicted.
The average absolute difference between predicted and observed
average daily water temperatures for the same time period was
0.83 degrees Celsius (Appendix D).
The model tended to over-predict mean daily water temperatures at
the Green River Gage (1959-1982).for the months of January,
February, March, April, and May. The other months tended to be
under-predicted. The average absolute difference for this gage
was 1.27 degrees Celsius (Appendix E).
6