WSP04333 - Laserfiche WebLink

Home Browse Search

" .... ,. ., j. '., . t. u;:., ,: Day, 1985) provided the capability to make long-term forecasts using the NWSRFS model. Extended Streamflow Prediction Procedure The ESP procedure uses the historical mean areal precipitation and mean areal temperature time-series data to simulate future streamflows under the assump- tion that each year of historical data represents a possi- ble future occurrence of precipitation and temperature. For each year of historical data, a simulation is made with the NWSRFS model by inputting the existing hydrologic conditions up to the present and by input- ting the historical data beyond the present. Thus, a sim- ulated hydrograph is generated for the forecast period for each year of historical data. The ESP procedure then performs a frequency analysis of the simulated future hydrographs; the frequency analysis provides the means to make probabilistic forecasts for the period of interest. Probabilistic forecasts can be made for different types of hydrologic data, such as discharge, reservoir volume or elevation, river stage, or snowpack water- equivalent. Also, different output variables can be con- sidered in the analysis, such as maximum, minimum, or daily average value; maximum instantaneous value; cumulative value (volume); or number of days that some specified value is or is not exceeded (Day, 1985, p. 164-165). In the application of ESP for the PRJUP . study, forecasts are made for daily average discharge, and the output variable is volume. The ESP procedure produces output for a condi- tional simulation, a historical simulation, and the observed (recorded) data. The conditional simulation output is produced by using the existing hydrologic conditions with the historical mean areal precipitation and mean areal temperature data, The historical simu- lation output is produced by continuous simulation of the historical data without resetting the initial condi- tions at the beginning of the forecast period to the exist- ing conditions. ESP output includes frequency curves for the conditional simulation, the historical simula- tion, and the recorded data. Simulation of streamflow always is subject to some error; therefore, there will be some difference between the frequency curves for the historical simula- tion and the recorded data. Based on the difference between these two frequency curves. an adjustment can be made to the conditional simulation frequency curve before a probabilistic forecast is made (Day, 1985, p. 165). The data output by the ESP procedure may be fit- ted to the empirical, normal, or log-normal distribu- tions. Example Forecasts Using the Extended Streamflow Prediction Procedure During development of the operational capabil- ity of the NWSRFS model for the PRJUP study, dis- charge foreCasts made using the ESP procedure were compared with water-supply forecasts prepared annu- ally by the U.S. Soil Conservation Service and the National Weather Service by using regression tech- niques. The forecasts are made for the April- September period on the first day of each month from January through May; only the April I forecast was used in the comparisons. The water-supply forecasts (U.S. Soil Conservation Service and National Weather Service, 1991, 1992) for the at-Granite and at-Salida stations (fig. I) were used for comparison to the ESP forecasts. [These stations were used because the ESP procedure was not implemented downstream to Pueblo Reservoir until after April 1992, even though the NWSRFS model had been calibrated downstream to Pueblo Reservoir during 1991; therefore, ESP forecasts of inflow volume to the reservoir could not be made in 1991 and 1992.] Comparison of the water-supply forecasts and the ESP forecasts is shown in figures 22 and 23. Gen- erally, the differences between the forecasts made by each method are not substantial. Use of the ESP proce- dure, though, has the additional benefit of the capability to provide forecasts for any time period, such as April 15 through May 14, whereas the published water- supply forecasts are only for the April-September period. . SUMMARY Part of the storage space of Pueblo Reservoir, which is located on the Arkansas River in southeastern Colorado, consists of a 65,950 acre-ft joint-use pool (JUP). The JUP can be used to provide additional con- servation capacity from November I to April 14, The operating procedures for Pueblo Reservoir, however, require that the JUP be completely evacuated by April 15 and used only for flood-control capacity until November 1. During winter, the JUP primarily is used to store water for agricultural uses, but because April IS is before the crop-growing season in south- eastern Colorado, little beneficial use can be made of the water released prior to April IS, A study was com- pleted during 1992 by the U.S. Geological Survey, in cooperation with the Southeastern Colorado Water Conservancy District to determine if the April 15 evac- uation date could be extended for any number of days from April 15 through May 14 under certain hydrologic conditions. 42 Use of Frequency Anslysls and the Extended Streamflow Prediction Procedure to Estimate Evscuatlon Dates for the JoInt-Use Pool of Pueblo Reservoir, Colorado