We present a retrospective assessment of an end-to-end hydrologic forecasting approach that uses ensemble climate forecasts to drive the Variable Infiltration Capacity (VIC) macroscale hydrologic model for large river basins. The method was previously tested in real time over the East Coast U.S. during the summer 2000 drought, and subsequently over the Columbia River Basin, during the summer 2001 Pacific Northwest drought. The forecasting domain has since been expanded to include the Colorado, Rio Grande and Sacramento-San Joaquin Rivers and the Great Basin. The linkage of the climate and hydrologic models provides a mechanism for exploiting potential skill that results from climate forecasts and/or knowledge of hydrologic initial conditions (soil moisture, snow), with the intent of improving streamflow forecasts at lead times of up to six months. Each month, climate model ensemble forecast output fields (monthly precipitation and temperature) are adjusted to remove climate model bias, downscaled, disaggregated and used to drive the daily 1/8 to 1/4 degree hydrologic models, which in turn produces six-month ensemble hydrologic and streamflow forecasts. Initial hydrologic conditions for the forecasts are estimated by driving the VIC model with observed meteorological data (gridded product derived from NCDC cooperator stations) and updated for the most recent three months using archived real-time Land Data Assimilation System (LDAS) gridded forcings. We evaluate the approach using retrospective ensemble hydrologic forecasts for two climate models, NCEP Global Spectral Model (GSM) and NCAR Community Climate Model (CCM3). In the Columbia River basin, the January forecasts are also compared to the baseline of a forecast starting on January 1 with a perfect estimate of accumulated (initial) snowpack, both with and without future climate forecast information derived from current ENSO/PDO state classification. The skill of the approach in predicting monthly streamflow and basin-average precipitation and temperature, snow water equivalent, soil moisture and runoff, is calculated for both unconditional and ENSO-composite ensembles. Anomalous initial hydrologic states influence hydrologic forecast skill for 1-3 months in many locations, particularly in areas where snowpack plays a dominant hydrologic role. The ENSO-composite analysis show that strongly anomalous SSTs support climate forecasts that add to the skill of the forecasting approach. In other locations, and/or in other years, climate forecasts from each of the models yield results no better than climatology.