8276 - Laserfiche WebLink

Home Browse Search

• Conclusions The existing stream network temperature model appears to be an good tool for the analysis of water temperatures in the Upper Colorado River Basin. Extension of the network needs to be completed to cover areas of biological importance. We need to gather additional data to cover 50 miles above Maybell on the Yampa River, 100 miles above the Watson gage on the White River, and 40 miles on the Gunnison River above Grand Junction. Time scales finer than one month may be desirable to provide results useful for research. However, the monthly time step we used should be adequate for long-term water development scenerio analysis and evaluation of management alternatives. Time frames from other hydrology models dictated the monthly time step. Additional comparisons of results may be made with the results of O'Brien (1984). He used the same stream network temperature model but only for a small segment of the Upper Colorado Basin. O'Brien's application simulated water temperatures on a weekly basis through the Dinosaur National Monument on the Yampa River. Work needs to be done on simulating or predicting reservoir release temperatures. Since ambient air temperature is the overwhelming influence on most water temperatures, management can exercise maximum prerogatives in controlling reservoir release temperatures. The assumption of thorough lateral mixing means that the model alone may not be used for prediction of temperatures in backwater areas isolated from the main channel. Since these backwaters may be important in the maintenance of endangered fish populations, additional refinements in applying the results of the temperature model for these situations are necessary. These cases are currently handled by the stream network habitat model. Theurer (in press) points out that mean or probable differences between published and predicted water temperatures are not exactly the same as errors. Both, however, cannot be correct. Errors exist in all sources: published, regressed and physical process-model derived. None of the published records were an accurate 24-hour average temperature. A single measurement was taken daily at unknown and variable times and thus can only be assumed to represent actual water temperatures between the minimum and maximum on a daily basis. It was for these reasons, and practical cost considerations, that calibration was not attempted on the larger data set. 0 8