At the National Centers for Environmental Prediction (NCEP) and using Global Ensemble Forecast System (GEFS) coupled with the Noah Land Surface Model (Noah LSM), experimental ensemble streamflow prognosis was generated for the CONUS domain at 1/8 degree resolution. The runoff forecast from the coupled NSF-Noah system is used to force a river routing model. Comparison between the output with analyzed streamflow suggested potential forecast values of this coupled approach (Hou et al., 2006). In this paper, the experiment is extended to the global domain with a lower resolution (0.5 degrees) and a preliminary analysis is carried out to the output. A streamflow “analysis” is generated following the methodology of the Global Land Surface Data Assimilation (GLDAS) project and by forcing the Land-River system with observed precipitation. This analysis is used as the initial condition for the river routing model in the coupled air-land-river forecast system, and as a proxy for truth in the verification of the experimental ensemble river flow forecast. In addition, uncertainty associated with the hydrological forecast system (land and river) is avoided. With these limitations in mind, the current study concentrates on natural flows and evaluates the quality of the output of the coupled air-land ensemble system as external forcing for river routing ensemble forecasting.