Water resources planning and management efficacy is subject to capturing inherent end-to-end uncertainties stemming from climatic and hydrological inputs and models. Accounting for and properly dealing with these propagating uncertainties remains a formidable challenge. In reservoir operation and water allocation decision-making, risk analysis is substantially dependent upon streamflow forecasts. Recent enhancements in climate forecasting skill and hydrological modeling serve as an impetus for further pursuing models and model combinations capable of delivering improved streamflow forecasts. However, little consideration has been given to methodologies that include coupling both multiple climate and multiple hydrological models, increasing the pool of streamflow forecast ensemble members and accounting for cumulative sources of uncertainty.

This work proposes integration and offline coupling of global climate models (GCM), multiple regional climate models, and numerous hydrologic models to improve streamflow forecasting and characterize system uncertainty through generation of ensemble forecasts. Different methodologies may subsequently be employed for optimal forecast combination. For demonstration purposes, the framework is imposed on the Jaguaribe basin in northeastern Brazil for a hindcast of 1971-2000 monthly streamflow. The ECHAM 4.5 GCM, regional models, including dynamical (NCEP regional spectral model, RSM) and statistical (using principal components) models are integrated with the Sacramento Soil Moisture Accounting and Soil Moisture Accounting Procedure hydrological models. Ten precipitation forecasts from the GCM are downscaled via the two regional models, and fed into the two hydrological models, producing forty streamflow forecasts. Super ensemble combination techniques include pooling, least squares, and a kernel density estimator to assess structural uncertainty of climate and hydrologic models.