The major drawback of climate mean bias correction is the need for a long training dataset, and since a reforecast works best with a frozen model, the database must be completely rebuilt whenever the model is updated, which requires large computing resources. In this way, routine improvements to the model are incorporated in the reforecast-based products as soon as they are implemented. However, due to the good performance of the climate mean bias correction, the current reforecast ensemble uses an old low-resolution version of the model system and it is worth the effort to generate the reforecast dataset and apply it to the ensemble postprocessing. NCEP has plans with ESRL/PSD to jointly implement a real-time hindcast experiment in the 2011–12 time frame and utilize additional resources to generate a set of historical ensemble reforecasts (20 yr). The operational forecast model will be applied to the reforecast configurations. Our postprocessing study will benefit from this new high-resolution reforecast dataset. Using the reforecast dataset, we will be able to test our postprocessing methodology and compare it with the calibration method developed by ESRL/PSD. New bias-correction methods developed under The Observing System Research and Predictability Experiment (THORPEX) project will also be considered for use in the NAEFS statistical postprocessing system.