Development of short-term streamflow forecasts for specific management applications: Case study of flow augmentation requirements for the maintenance of endangered fish habitat.

The effects of reservoir operations and trans-basin diversions in the upper Colorado River basin have significantly reduced the habitat of endangered native fishes. Spring snow melt flows are typically much lower and recede quicker now than before river basin development, and winter base flows are now higher than before. In the alluvial reaches of the Colorado River near Grand Junction, lower peak flows have resulted in preferential lateral accretion and significant in-filling of side channel and backwater habitats. In the Upper Colorado river a coordinated effort among different reservoir operators seeks to release surplus water to augment the natural peak in the annual hydrograph to benefit endangered fish. The goal is to generate flows that are high enough to flush sediment through reaches of the Colorado River where endangered native fish are concentrated, and thus maintain the fish habitat that still exists. For these coordinated reservoir operations to be successful, reservoir releases must augment the natural flow to a level that is above the critical threshold for gravel transport. In a dry year, or if the releases of water do not coincide with the date of natural peak flow, it may not be possible to release enough water to reach the gravel transport threshold, and any release of water would clearly be wasteful.

To improve decisions on when and if to attempt a peak augmentation, a study has been initiated to improve procedures for forecasting the magnitude and timing of the natural peak in the annual hydrograph. Forecasts of precipitation and maximum and minimum temperature from the National Centers for Environmental Prediction (NCEP) Medium Range Forecast model are used as input to a basin-scale hydrologic model configured over the Animas River, an unregulated basin in southwestern Colorado. In this experiment, the U.S. Geological Survey's Precipitation Runoff Modeling System (PRMS) was forced with two sets of precipitation and maximum and minimum temperature: (1) climatological values; and (2) NCEP forecasts. The climatological values provide a baseline for measuring the accuracy of hydrologic forecasts when forcing PRMS with NCEP forecasts. Using climatology, some accuracy in runoff forecasts was achieved at the beginning of the forecast cycle (the first 2-3 days). This can be attributed to the lag time in PRMS and stresses the importance of accurately specifying initial conditions. Significant accuracy was present in the NCEP maximum temperature forecasts in spring and autumn and in the NCEP minimum temperature forecasts during winter. The reliable maximum temperature predictions translated into reliable estimates of snow melt and runoff out to time scales of at least 8 days, with forecast errors much lower than those generated usingclimatology. These improved predictions of the natural peak flow are expected to result in increased ecological benefits of the flow augmentation.