We present the formulation and validation of an ensemble flow prediction method, which is based on operational hydrologic forecasting. The method is suitable for reservoir management applications, and examples of such applications are discussed for Folsom Lake, California, Lake Lanier, Georgia, and Lake Norris, Tennessee. Ensemble flow forecasting is conditioned on climate model simulations and predictions, and it incorporates hydrologic-model uncertainty. Validation is performed using (a) forecast-reliability measures pertaining to the observed flow-volume distribution, and (b) reservoir management performance measures (e.g., Carpenter and Georgakakos, 2001; Yao and Georgakakos, A., 2001; and Georgakakos et al. 2000). Sensitivity analysis results are presented for various global climate models and for various numbers of ensemble members. The study quantifies benefits of using climate information for predicting and managing reservoir inflows.

REFERENCES

Carpenter, T.M., and K.P. Georgakakos, 2001: Assessment of Folsom Lake response to historical and potential future climate scenarios, 1, Forecasting. J. Hydrology, 249, 148-175. Georgakakos, K.P., Graham, N.E., and A.P. Georgakakos, 2000: Can forecasts accrue benefits for reservoir management? The Climate Report, 1(4), 7-10. Yao, H., and A.P. Georgakakos, 2001: Assessment of Folsom Lake response to historical and potential future climate scenarios, 1, Reservoir Management. J. Hydrology, 249, 176-196.