Atmospheric ensemble forecasts contain biases that must be removed before they are used as input to hydrological models. Also the spread of the adjusted ensembles underestimates the true uncertainty. Correcting these limitations of weather and climate ensemble forecasts is essential to produce skillfull and reliable ensemble streamflow forecasts. Results of a study of alternative approaches to re-scale NCEP global ensemble forecasts and to compensate for under-estimation of ensemble spread will be presented. These approaches are being used to support the NWS Advanced Hydrologic Prediction Services (AHPS)